Category Archivedarpa uc 2007
computers and technology & darpa uc 2007 & geek & interesting & news & personal & programming & youtube 14 Jul 2008 05:31 pm
Robocars on Discovery Science Channel to feature autonomous vehicles from DARPA Urban Challenge, including Team CajunBot.
The series “Robocars” will be premiering tonight at 9pm (Central time) on the Discovery Science Channel. The show follows ten teams of top engineers from around the U.S. compete for a $2 million grand prize, struggling to build the first vehicle to drive itself through an urban environment and features Team CajunBot.
Here is the schedule and episode descriptions:
July 14th 9-10pm - Episode 1 - follows Stanford Racing, Tartan Racing, Team Jefferson, Team Gray and The Golem Group as they prepare for the Urban Challenge and pass through the DARPA site visits.
July 21st 9-10pm - Episode 2 - follows Highlander Racing, Team Oshkosh, Team Cajunbot, Team MIT, and Team Case as they prepare for the Urban Challenge and pass through the DARPA site visits. The show ends with DARPA announcing the teams who made it to the semi-finals.
July 28th 9-10pm - Episode 3 - covers the semi-finals. Stanford Racing, Tartan Racing, Team Jefferson, Team Gray, The Golem Group, Team Oshkosh, Team Cajunbot, Team MIT and Team Case are all included. The show ends with DARPA announcing the teams who made it to the finals.
August 4th 9-10pm - Epiosde 4 - covers the finals. Stanford Racing, Tartan Racing, Team MIT, and Team Oshkosh are all included.
August 11th 8-10pm - Episodes 5 and 6 - The first hour is a summary of the last four episodes and the second hour focuses on futuristic car technology and contains excerpt from the DARPA Urban Challenge.
Read more: Welcome to the CajunBot Lab website.
computers and technology & darpa uc 2007 & geek & interesting & news & personal 11 Jul 2008 11:35 am
New Radiohead Video is Shot with Lasers, Not Cameras.
I’m sure y’all remember the expensive ice cream buckets on top of several DARPA Urban Challenge vehicles…
Radiohead, never ones to shy away from trying new things, has shot its new video for “House of Cards” without using cameras at all. Whaa? Yes, they’ve used two fancy new technologies called Geometric Informatics and Velodyne Lidar.
Read more: New Radiohead Video is Shot with Lasers, Not Cameras.
darpa uc 2007 15 Nov 2007 09:55 am
Adventures in Vegasland: Standing on the Corner, Watching All the Bots Go By
MEGAN EDWARDS: Adventures in Vegasland: Standing on the Corner, Watching All the Bots Go By
Although I was of course not privy to all the data and discussion that went into deciding who won the DARPA Urban Challenge, I couldn’t help getting an Animal Farm feel about how the teams were treated. Over the course of the weekend, it seemed obvious that while all the teams were nominally equal, some were considerably more equal than others. Stanford, for example, was not assigned to a pit area right next to a giant pile of dirt. The dust storm generated by every slight breeze can’t have been good for UCF’s computers, and they had no tent. Also, during the race, I observed MIT and Cornell both get the opportunity to restart their ‘bots when they stopped performing, but when Team AnnieWay’s car stopped for a minute or two, they were immediately cut from the competition. And then there was the issue of Carnegie-Mellon’s delayed start. Not only was the team allowed to keep on trying, they were allowed to replace hardware on their car. At the awards ceremony, Tony Tether placed blame for the ‘bot’s troubles on radio frequencies generated by the JumboTron nearby. While I was not the eye of God observing from on high, from my lowly spot on the sidelines it looked as though the teams from select major American research institutions were given numerous special indulgences.
Which of course is just a reminder that the DARPA Urban Challenge was not really a car race. In actuality, it was a clever way to spend tax money to get difficult and expensive research done as quickly and cheaply as possible. If, as I’m sure is true, Carnegie-Mellon’s research capabilities and likely results exceed those of the University of Central Florida, well, I guess it really does make more sense to throw cash and acclaim their way. As for the German teams, I find it easy to believe that DARPA found it in their own best interest to avoid giving any of them the purse, even though they had the required American sponsorship. Otherwise, they might have to explain why $2 million of U.S. taxpayer money earmarked for research here in the homeland was winging its way to Germany. As for the Stanford/Carnegie-Mellon rivalry, Stanford won the big check last time, and anyway, isn’t it just better to give the biggest eagle to a good old all-American Chevy rather than a Volkswagen? I couldn’t help wondering about all this once again when I learned that the officially published difference between the finish times of Stanford and Carnegie-Mellon was one second. Given all the times the ‘bots had to pause and all the judgment calls about whose fault any mishaps were, a lot of subjectivity went into coming up with that one second.
darpa uc 2007 & news & personal 02 Nov 2007 09:53 am
CajunBot Blog: Updates from NQE.
We’re Heeeeere
Hello Team Cajunbot fans. This is Joshua coming to you through the interwebs from sunny California. We arrived in California yesterday and, after refusing to order cajun shrimp at a TGI Friday’s in California, hit the road and drove about 45 minutes to Hesperia. I knew we were staying at a Days Inn here, but I wasn’t really worried about it. As long as I have a bed I’m happy right. That was before I got here. The place isn’t so bad, except for being in a seedy area. There is a Circle K on the one side with customers throwing up gang signs and scaring the clerk (seriously). Oh and did I mention last night a police helicopter flew over our parking light shining a spotlight like they were looking for an escaped criminal. But no worries. I’ve lived on the hard streets of Lafayette for a few years so I can handle it.
True to Cajunbot form, we were up until 3am this morning (5am Lafayette time - I hate jetlag) calibrating sensors and making sure the bot was in working order. Since Chandan and I were up late (or early I guess) getting the bot ready to go, I got to wake up at the crack of 11:30 while the other team members were running Raginbot and doing silly things like eating breakfast and lunch. I prefer to sleep through both, but I have a feeling that will be the last time I’ll get to sleep like that until I get back home. We’re not sure where we are going to test the bot yet, but we might go visit slashx, the area of the first Grand Challenge, and test there for a while. We have a lot do do still, and only a few days to do it. Wish us luck, because we’re not going to sleep until Raginbot is ready to go.
Oh yeah, I’m going to try to post a new blog every day about what is going on with the team, so stay tuned for updates, and thanks for all your support!
Joshua.
(with edits by Doc A using Disney style guide)
Deja Vu — Joshua building new sensor mounts in hotel parking lot
VICTORVILLE, CA - Well, I was up late last night capturing the excitement. I was saving the writeup as I went along. Finally, I added some pictures and saved, and what came back was a blank page. All my creative writing was gone in the blackhole of cyber space.
So here I go again…
A few days ago I exchanged note with a fan who used the phrase — change is the only constant. Its one I have heard before, but for some reason it stuck in my mind. I thought it was too cliched, and had to respond with some thing else. So I came up with another tautology — some things stay constant through change.
That’s what I was experiencing last night. A new bot, all new sensors, a new challenge, a different (not new) hotel. But one thing stayed constant through all the changes. Joshua waa out in the parking lot cutting and grinding metal. We were there before in the first challenge, except there were a lot more hands besides Joshua. The first version of the ME team, big man Jonathan Raush using his hands as vice and Patrick doing the cutting and welding. We were there the second time with Adrian, Joshua, and Murali making sensor mounts in a church parking lot. And here we are again, Joshua breaking the quiet of the night with the machine tools, helped by John Herpin.
So why this last minute metal work.
We currently have one very significant weakness. CajunBot-II (henceforth, I’d refer to as CajunBot) cannot see very far to the left and right. This particularly happens when the road has a hump, such as to drain water. The vehcile has a slight ‘roll’, which makes its lidar beam shoot over a vehicle over 30 meter to the left and shoot on the ground 30m to the right. Which means CajunBot cannot see any vehicle beyond 30m to the left and right. This is very, very limiting when she is trying to merge into traffic. It takes the vehicle about 8 seconds to complete a turn. If the traffic is flowing at 20 mile/hour (about 10 meter/sec), it would take 3 seconds for a vehicle (with right of way) to plough into CajunBot.
Why are the sensors so limited? Well, we finally found out yesterday that is a feature of the sensor, not a bug. We are using Ibeo ALASCA Scanner. We have been having trouble with the sensors seeing objects at 150m sometime, and just about 30m other times. We’ve been stumped. Their sales engineers narrowed down the issue — “The sensors are faulty.” Which is a very unsual and candid statement from a “sales” person. That’s probably why he is also called an “engineer.” The company was extremely kind and sent us two new sensors, which we received the day CajunBot was leaving for Victorville. Our hopes of fixing the problems were squashed once we tested the vehicle with the new sensors. We just did not find any improvement.
But there is a silver lining there. We now have two more Ibeo scanners. Which is like Christmas coming early since their price tag is about 1/4th of my entire budget. Mark McKelvy, John Herpin, and Joshua figured out how they can put ‘our’ (until we return them) sensors to good use.
Joshua was busy cutting metal to build mounts for ‘our’ additional sensors.
– Arun Lakhotia
Team Overbot and Team Cyberrider come to rescue with extra radar
VICTORVILLE, CA - Time and again strangers have rushed to our rescue some very expensive property, just on the basis on the phone call. Last year it was “2 the Max ATV” (Gina — we remember you), who had sent a brand new Max IV ATV to use as spare. The project itself was started by the generosity of Brother Ray Majors, at that time a stranger. He sent his ATV after just 15 minutes on the phone.
This time we have previous Grand Challenge competitors sending us their Eaton Vorad Radar’s.
Last night once John H, Mark, and Dallas finally diagnosed why Ibeo Scanners were not seeing as far in certain situations. The short version of the diagnosis — there is a significant blind spot in our current sensor configuration. The solution — fill the blind spot using additional sensors.
We are now running two strategies in parallel. The first one, as mentioned earlier, is to use ‘our’ (until we return them) new Ibeo Scanners. We can use the extra scanners to cover area left blind by the ones we have.
The second strategy, the subject of this note, is to use additional radars. We currently have two Eaton Vorad EVT-300 Radars. We bought three of these units for the first challenge. At that time they ended up being used like jewelry on the bot, they were not even power on. Three weeks ago, when we were having trouble with the Ibeo Scanners, we installed two of these units on the bot, one on each side of the vehicle’s hood. Suresh G, our lead software engineer, finally finished integrating the radars with the rest of the system on the 45 minutes flight from Lafayette to Houston. The integrated system was tested on Sunday and found to be very good, so long as the radars could see other vehicles on the road.
The good thing about these radars is that their range — almost 150 meters. The bad thing is their field of view (FOV) — about 12 degrees. This FOV is sufficient to see along a lane if the radar is centered on the lane. But we want to use the radars at intersections, before the vehicle turns into a lane. This reduces the effective FOV.
Our second solution is to install two radars on each side of the hood. The 24 degrees FOV, and crossed fingers, would cover the blind spot well enough.
There is one catch though. Eaton Vorad does not manufacture these units any more. About a month ago when I started my contingency purchases, I had coincidentally received a note from a researcher in Singapore asking if I could spare my VBOX (the electronic box that goes with the radar). It was not much of an issue at that time because our plans to use the radars were not firm. However, with the radars taking center stage now I had a big problem — we have only three units. The company does not make any more of these. Where do I get the fourth from?
Google came to rescue. It turns out the only people who have publicly stated that they use Eaton Vorad Radars are four teams from 2004 DARPA Grand Challenge. The other three are Team AION, Team Cyberrider, and Team Overbot.
After half hour on Google I had the phone numbers of two teams. Twenty minutes later I had commitment from Ivar Schoenmeyr of Team Cyberrider and John Nagle of Team Overbot to loan their radars. No official letter, not even a hand shake, just a simple “hello, please, and thank you” later I have two radar units on their way to the hotel by Fedex.
Isn’t that amazing. These radars, mind you are not inexpensive. Besides now that they are out of production, they become more valuable for people who need them. They work really well for what they do. And these two guys are sending it to a stranger. Well, its likely we met during the past challenges, but we were too busy to remember the encounters. So really we are strangers for all practical purposes.
My guess is that the reason each of the person I have mentioned above stepped up to help a stranger (my team) is because of the magnitude of the endeavor we have undertaken. Great life lesson, albeit pretty late in the career.
– Arun Lakhotia
Team Cajunbot comes together
Today was an interesting day. I woke up at 8AM after going to bed at 4am because my roommate still has his cell phone alarm set. The room wasn’t cold but I was freezing. For the next 2 hours I lay in the bed, covers over my head and too cold to get up and turn the heater on. I thought “I know this feeling. I’m sick”. I guess its the change of weather messing with me, but all day I’ve had a fever and have been feeling just generally bad. The 4 hours of sleep probably didn’t help either. But Team Cajunbot is versatile, and so we made a list at lunch of what tasks had to be done and my teammates each took a task and worked on it. Drilling, running wires, messing with the computers - they stepped up and did jobs they don’t usually do so I could get some rest. It’s about 11:30 in California now, and we have 2 radar mounts finished (Dallas and Big John), the bottom laser sensor back on the bot and calibrated (Little John, Amit and Chris), and the disk from the computer backed up and replaced (Mark). Thanks to everyone who pitched in today and did my many jobs. Hopefully I will get a good night’s sleep tonight and be back to my multitasking self tomorrow. I have some pictures of our day to day work that I want to put in the picture gallery, but I haven’t had time to do that yet. Hopefully I will be able to do this tomorrow so you all can see what we are getting done here. By the way, it has occurred to me that it takes an entire team and a sick me to do my job. You know what - I’m feeling better already 
Joshua.
Santa Ana winds and forest fires
VICTORVILLE, CA - If you’ve been watching the news, you’ve learned about the forest fires, Santa Ana winds, and closed highways. You may be wondering if we are safe.
From our hotel we can see the huge bellows of smoke, flowing from the north to the south along the mountains, we can also see the fire itself, yet we are not in the area (currently) threatened by the winds and the fire. The wind is so strong that it has toppled several big-rig trucks. One can only imagine what it may do with cars, may be turn them into projectiles.
Though we are not in the line of fire (literally and figuratively), we are affected. Earlier today we had planned to send Danny to Irvine to pick up a radar from Team Cyberrider. But Danny had to turn around very soon after getting on the highway; the highway was closed. Thanks to the highway closure I have also had to re-route Amit Puntambekar via Las Vegas , instead of Ontario, CA. Amit, who now works for Vmware in the valley, was scheduled to fly through Ontario. We cannot get to the airport due to the natural disaster. But he has been adamant that he needs to get to work. So we had to change his itinerary.
Other than limiting our ability to go West to Los Angeles region, Santa Ana winds are not really affecting us much.
- Arun Lakhotia
New sensors mounted and operational
October 24, 2007. VICTORVILLE, CA - In case you are holding your breath, you can let go. We have finished mounting ‘our’ two new Ibeo Scanners and two radars (thanks again to Team Cyberrider and Team Overbot). The sensors have been completely integrated into the system. We still need to run some tests to assess the improvements. It sure cannot be worse than what we had.
This is a miracle, considering Joshua was down sick. Or may be it is because of that. We had five people doing things that he does. Now that his territory is encroached, it would be hard for him to keep others off.
It was John Loftin (christened Big John) to shine. He is a gentle giant. A big frame with a very polite demeanor. He rented a gas welder and got to work. With specs from Joshua, he knocked out the mounts for the two sensor suites in no time.
Dalls Griffith finally got to do what he has so often offered, but has previously been shooed away. He got to drill a hole in the Jeep. (No don’t choke.) Its a hole to route wires from the under the hood into the cabin. Dallas did a meticulous job laying the cables. Goes to show you cannot judge people by their clothes.
The big surprise to me was seeing Chandan Uddaraju helping with routing cables in the Jeep. I didn’t expect to ever catch him looking in the engine block of a vehicle. But there he was his hands deep under the hood. Goes to say you can not judge people by their degree either.
After the hardware was integrated to the vehicle, Christopher Mire and Chandan took charge of integrating the sensors to the software system. Since we have other similar sensors, integration was simple enough.
But nothing is straightforward. More sensors, means more data, means more cpu cycles, means our one ‘main’ computer was choaked. In normal situations this would have been an issue. But Joshua has already loaded up spare machines on the computer rack. In no time Christopher had distributed the computation to the ‘overflow’ machine.
With all these additional sensors, I think we have significantly addressed one of our weakness. CajunBot is still limited in its sensing capability in comparison to other challengers. But the sensors we have are likely to be sufficient for our needs.
- Arun Lakhotia
Pit and hospitality area setup
October 25, 2007. VICTORVILLE, CA — The drill starts tomorrow. Teams have setup their camps. We have been assigned areas in a parking lot. Each team’s area is about eight parking spots long. Danny and Big John setup our camp last night. Its the nice 42′ enclosed trailer (courtesy Quality Transport, Inc.) that CajunBot was transported in and a rented Motor Home 35′ long. The motorhome will serve as the lab, and the trailer will serve as workshop and storage for the bot.
Danny has also setup a hospitality tent. Its the most attractive and pleasant setup of all teams. While all others have the generic white canopy tent. Ours tent is straight out of Mediterranean Nights. Its white canopy with flowing Red and White shades (or curtains, not sure what they are called).
We have also received the prime location for the hospitality area. Its right overlooking the DARPA tent, where all the traffic is.
How so ever nice the hospitality tent is, there is a problem. I do not have enough hands to man it. PR folks — its an opportunity, you will surely regret missing.
But I am looking forward to the Cajun Cookout to be hosted by Brother Ray Majors on November 3rd. The hospitality tent will be put to good use that day.
- Arun Lakhotia
Process of selecting the finalists
October 25, 2007, VICTORVILLE, CA. — There are 35 teams in the semifinals. On November 1st there will be AT MOST 20 remaining. Notice the emphasis on AT MOST. For a team to qualify as a finalist it MUST meet some minimum criteria, and that minimum criteria is a lot higher than the Grand Challenges.
Here is a quick overview of the selection process.
Lets first get the terminology straight.
The semifinal ‘tests’ will be held from October 26 to November 31. Each day is split into two ‘time’ blocks–morning and afternoon–except the first day and last days have only one block. Thus, in the entire duration there are 10 blocks.
There are three test areas, creatively named Area A, Area B, and Area C. While their names are not informative, their locations are not secret. About a month ago we were given a Google Map marking the locations. We were also given the directive to stay our of those areas, or else risk being disqualified.
The 35 teams are split into five groups. DARPA does not appear to have given a name to each group, which seems like a procedural oversight. The other teams in our group are: Stanford Racing Team, Team Cybernet, Team Jefferson, Gator Nation, Honeywell/IVS Team, The Golem Group.
Now let us put it all together.
DARPA has announced the schedule for the first five test blocks. What will happen after the first five block is still a secret. I think its a secret not so much out of habit (after all its a DoD organization), but it gives DARPA some flexibility.
During each of the first five test blocks, there will be some tests happening in each test area. Since there are only three test areas and five groups, it follows that not every group will be tested in each block.
We have tests scheduled in Block 1, 3, and 5. This is extremely nice. These odd numbered blocks happened to be ‘afternoon blocks’, from 1220 to 1700 (military time). Compare that with the morning block time 0730 to 1200, and you know why I think it is extremely nice. My team is just not a morning team.
Now let me say the same thing in simple words. Our first three tests (which is all that are scheduled) are in the afternoons of Friday, Saturday, and Sunday. (I had to write the other details just to show off my professorial skills.)
Now that we got our testing times straight, lets get on to the specific tests.
Our first test (Friday) is in Area B, next (Saturday) is in Area C, and third (Sunday) in Area A.
According to DARPA “Each team is scheduled for ONE attempt at each of the three test areas” (emphasis added). The crucial part is we get one shot. And then we know on Monday where we stand.
The specifics of the tests in each area is not disclosed. But we have received the ‘RNDF’ file, which is essentially a map of each area. We can guess from the map what the test questions will be. Here is what I expect.
Area A - Test ability to drive with sparse waypoints (or obstacle field).
Area B - Test ability to merge in traffic and park.
Area C - Test ability to negotiate intersections (with other traffic)
As we stand today, we are pretty good in tests for Area B. We have covered most issues expected in Area C, except for the ability to negotiate jammed intersections. We are not as strong in the tests for Area A, but we have time on our side.
Now its time for the team to head off to the pit area. We are scheduled for a photoshoot with Dr. Tony Tether, the head man of DARPA.
- Arun Lakhotia
No NQE test on Friday, tests start from Saturday
October 25, 2007. VICTORVILLE, CA — Quick news. DARPA has changed testing schedule. There will be no NQE tests on Friday. Our new testing schedule is — Area C on Saturday PM, Area A on Sunday PM, and Area B on Monday AM.
Tomorrow is left as a preparation day, both for the team and DARPA. They need to work out the kinks in their testing protocol.
- Arun Lakhotia
Broadpoint satellite brings Internet to Team CajunBot pit
October 25, 2007, VICTORVILLE, CA. — If you are eager to follow the happenings in the challenge, you got to thank Mark LaGrone and Danny Burgess of Broadpoint. Thanks to them I am sitting comfortably in a motorhome in our pit and writing this note. I hope to bring you daily events as soon as they happen.
Broadpoint has provided us with a satellite dish and subscription to the Internet service. The day before the bot left Lafayette I called Mark, my former student and neighbor, requesting a satellite hookup. Next morning 8:00am it was all ready to be picked up, with tech support instructions and all.
I was surprised how nice and compact the whole package was. Just three pieces. A satellite dish, a big suitcase with all the electronics, and a cardboard box with stand for the antenna. All I had to acquire on this side was sandbags to hold the base down.
Setting up the satellite dish was easy too. At least Dallas made it look easy. He is another multifaceted person. A computer science major, who is part of a rock band, does house renovation, great with logistics, and can climb up the RV to setup satellite dish.
While getting the dish setup on the roof of a RV is physically challenging task, for someone like me, the big work in setting up the dish is pointing it to the right satellite and setting up the correct orientations. Thanks to Broadpoints technology this was a breeze. Mark has given us an instrument called Birddog. With this instrument, you just punch in a few numbers, and it does all the work needed to talk to the satellite.
Isn’t that great. Another Ragin’Cajuns supporter stepping up to the plate to give a little helping hand to Team CajunBot.
Judice Middle students - I have read your comments on the blog. You can EARN T-Shirts, CajunBot visit to the school, and even an invitation to CajunBot’s homecoming celebration. All you need to do is find some creative way to thank my sponsors (individuals and companies that have helped my team).
Lets start with Broadpoint. Show me what you can do to make them proud of supporting us. I have other sponsors too. So get started fast. I’d be reeling out other names once I see your spirit. You help me with appreciating the sponsors, and you can earn a visit to Cajunbot’s homecoming on November 8th.
In case you need it, Broadpoint’s address is 113 N. Pat St., Scott, LA 70583.
The above offer is open to students of other Lafayette area schools as well. Please don’t mind my bias towards Judice Middle. My wife works there, and I couldn’t have done what I am doing without her support.
- Arun Lakhotia
Started off on the wrong foot
October 26, 2007, VICTORVILLE, CA. Through these reports, I want you (the reader) to experience what we are experiencing. This means sharing with you the excitements of our successes and pain of our failures. Today, I have to share a pain.
Today was supposed to be an uninteresting day. Start with opening ceremony at 0700. Then go for ’safety testing’ at 1030. We have gone through safety testing so many times that it was expected to be a mechanical step.
Turns out that things did not go as smoothly as expected. We exhausted our alloted 30 minutes, and would need to do the test again.
Lets start with the basics of the capability being tested. The robots are autonomous, they drive by themselves. But you’d not want a car to go beserk. So DARPA has provided us with an ‘E-stop’, a wireless equipment. The receiver is setup on the bot. The transmitter is held by the DARPA operator. The E-stop has two buttons: DISABLE and PAUSE/RUN.
Before starting the bot, the operator presses the PAUSE button. Then the bot is given command to launch. The operator than releases PAUSE (and changes to RUN). The bot can now move. The operator may press PAUSE at anytime, and the bot must come stop. When the PAUSE is released, the bot may move. However, if the operator deems the bot to be getting dangerous, s/he may press DISABLE. At which point the bot must turn off the engine and never start again even if its put in RUN mode.
The E-stop safety test is intended to test this capability. If the bot behave as described above, it cannot proceed to any other tests.
The E-stop safety test proceeds as follows. We are given a ‘E-stop Test RNDF’, which essentially has a straightline path with two waypoints. The bot is placed on the first waypoint, the operator presses PAUSE, we are asked to give bot the command to go to the second waypoint, the operator releases PAUSE (changing to RUN mode), the bot begins to move, the operator presses DISABLE.
Nothing to it, we went through these steps for the DARPA site visit in June.
But, sometimes its the simple things that can bite you. In the above sequence when the operator pressed released PAUSE, the bot would not move. When the bot goes autonomous she is expected to turn on its emergency (flashing) lights and the siren. This time she was completely quiet. Not a beep and not lights.
The launch team, consisting of Joshua, Suresh, and Little John, were in the spotlight. Crew from Discovery Channel howering over goading them to verbalize what they were thinking. DARPA radio crackling, reporting the issue to higher up the chain. The teams in line for e-stop testing were queuing up. Just a frenzy of activity and excitement. Top it all the Sun was showing no mercy either. The high desert has no clouds. If anything is blocking the heat, its the smoke from the fires on the west.
Most people would crack under such pressure. But Suresh and Joshua have been hardened, having gone through such things and worse in the past two challenges. This is first experience for Little John, and he was holding his nerves well. The trio, who were the only people around the bot, were thinking on their feet as fast as they could. They came up with a slew of possibilities. Its hard to believe all of the possibilities they came up with turned out to be correct. Very soon they ran out of clock, and had to give way to the next team.
Just to be fair to the team, we are not the only team in this boat. There are other teams that have experienced problems as well. That’s what makes this challenge so much fun.
We are now schedule for a re-test at 1700. But DARPA is running behind schedule, and its possible we may have a retest in the morning.
- Arun Lakhotia
E-stop retest went ok, good to go tomorrow
October 26, 2007, VICTORVILLE, CA. — Quick note. We went for the E-stop retest. Everything went fine this time. The issues were all corrected. Now the main event starts tomorrow. The first test is in Area C.
- Arun Lakhotia
Get to know the NQE test area
October 27, 2007, VICTORVILLE, CA. — The NQE start today. Since you are following the challenge, you may be interested in the knowing about the physical location.
As a starter, locate us on Google Maps. Click on View Map or search for ‘Starfighter Lane, Victorville, CA.’ The ‘arrow’ on the map points to the ’start chute’ location for the finals of the challenge.
This area is now called ‘Southern California Logistics Agency.’ It used to be George Air Force Base. All that remains in this base are some empty buildings and an elementary school. Since no one really lives here, I wonder how this school gets kids.
Here is an aerial image of the Test Area A. (Click on the image to see a bigger picture.)
The test area is created in an abadoned parking lot. The yellow lines on the map show the DARPA map given to us. This track will be used for testing the ability of the bot to merge in traffic. The area consists of a rectangle with two way traffic, and a path cutting through the rectangle. During the tests, vehicles operated by DARPA drivers will be doing laps on the rectangle. The bot is expected to arrive at the intersections, and wait for other vehicle before merging in the traffic.
The following is an aerial image of Test Area B.
Area B looks the most complicated. It will be used to test the bot’s ability to drive even if the map given by DARPA does not have enough data points, the ability to park and unpark, the ability to drive through open space while avoiding obstacles.
Finally, the aerial image of Test Area C.
This area will be used for the bot’s ability to re-route it planned path if it comes across a blocked road.
Today we test in Area C, then in Area A on Sunday, and in Area C on Monday.
- Arun Lakhotia
Day 1 - Acitivities picking up
October 27, 2007 (0738 hrs), VICTORVILLE, CA.
Day 1 activities are picking up. Most members of the team have been working through the night. Yesterdays licking in E-stop test served a purpose. The first timers have gotten a taste of what to expect in the next few days. The overall energy level is high.
Pablo Mejia, our Chief Architect, arrived last night. His presence changes the dynamics of the team. He has earned respect of everyone by his sharp mind, shooting from the hip simple solutions to difficult problems, and staying calm like a rock even in turbulence.
Pablo and crew have worked through the night. We have had a to close a lot of loose ends, issues we have known, but have not had time to attend.
At 0850 we have a practice run scheduled. We get a 40 minute slot in a wide open, dirt area. But for the bot, nothing else is allowed in this area. Which means we cannot really test some of the advanced capabilities in this area. However, our plan is to test the system’s ability to run for the whole duration. This will be a test of the electronics of the system. We experienced an electrical interference yesterday during E-stop test. Need to test whether similar issue may arise after the machine has been running (with all systems on) for a long duration.
The NQE test starts at 1220. We are the first in the group, followed by Stanford.
My role today is to wait on the team members. Joshua, Chandan, and Mark slept at 0430. I am not sure if Pablo slept at all. Suresh slept early, but woke up at 0100. They will be tired when they wake up. My role is to drive them to pit, power them up with coffee, and cater breakfast.
Quite coincidentally this hotel is owned by an immigrant of Indian descent. This guy has been great, serving me ‘chai’ every morning. So, my first stop out of my room is to go to his office and get my chai.
Time to shutdown the machine and roll.
- Arun Lakhotia
Day 1 didn’t go as well….
October 28, 2007 (0933 hrs), VICTORVILLE, CA. — You’ve not heard from me in 24 hrs, and I am sure you guessed that cannot be good news. Our performance in Test Area C was sub par. CajunBot goofed up on things we have tested over and over again, and didn’t complete one task we knew she may have difficulty.
The moment we finished (rather, did not finish) the test, we went into a huddle. Analyzed the reasons, and then started focusing on the next test today in Area A. This one is going to be a killer. Almost everyone has been working through the night. That makes it two nights barely with any sleep for several members — Pablo, Suresh, Mark, John, Chandan, and Dallas.
Here is a quick summary of what happened in Test Area C. The area is a ring with two 4-way stops (intersections) on opposite end. The bot is expected to go around the ring, and at each 4-way stop she is met with other traffic vehicles. She has to maintain the rules of a ’stop’ sign, give precedence to other vehicles that come before it.
This capability was tested few months ago in the site visit, and we test it routinely.
CajunBot did well for the first two interactions at the stops. In the third interaction, there were three vehicles already at the intersection. She let two vehicles go, but then took before the third vehicle moved. This was a complete surprise to us.
After that the behavior kept getting worse. A similar scenario was created in the next stop. This time CajunBot almost kissed the bumper of the car ahead. In the next loop, she was riding the curb and didn’t care of any other vehicle at the stop sign. You kind of get a feeling she was irritated to be given such stupid test.
As the loop progressed she was to encounter a blockage in the road and expected to turn around. This is also something she has done many times. But this time around, she just went stopped in front of the blockage and decided not to move.
That was the end of the test, for the bot would not progress any further.
What caused this misbehavior? Our analysis indicates two things: first, it looks like after two loops the bot began experiencing a GPS drift, which essentially means she was shifting the middle of the road about two feet to the right (in some situations). That made her ride the curb. Second, we have never tested in a situation like this before. All our tests are in wide open spaces. This is the first time we are testing in a real neighborhood. The main difference now is that our sensors pick up a lot, lot more data in this environment. In the process something has tripped.
More on this later… I need to rush out to the test area.
- Arun Lakhotia
Day 2 - A fantastic turn around
October 28, 2007 (1533 hrs), VICTORVILLE, CA. — CajunBot went through the test of fire, and came out looking real good. Its been a great 36 hrs day for many, in particular, Little John, and now he is too excited to fall asleep.
The test in Area A is plain and simple scary. The track consists of a two lane (one each way) rectangular road, about 120m long and 40m wide. Across the middle of the long segment, top to bottom, runs a one lane road.
The test setup consists of 10 (may be more) cars making laps around the course, five in the inside loop and five in the outside loop.
CajunBot (as are other vehicles) is given the mission to loop around as well, but she does not travel the full rectangle. Instead she is supposed cut across the rectangle, along the one lane road.
The vehicles in the outer loop have the right of way.
The test requirement is that when CajunBot turns left from the outerloop to the one lane segment, she has to yield to the traffic in the oncoming lane. When she makes the next left turn from the one lane segment to the outer loop, she has to yield to the traffic coming from the left and the right.
The traffic flows around 8 to 12 miles/hour.
The traffic vehicles is manned by race car drivers. They are driving specially built with amazing amount of reinforcement. The drivers too are in full race-car regalia, with helmets and suites designed for race car driving.
The evaluation is based on number of loops completed in the 40 minutes alloted, the number of loops made without any safety error. There is penalty for making the other drivers honk. There is also a penalty for taking too long to merge into traffic.
I forget the number of loops CajunBot did, but it was in double digits (around 15, or more). Over six of those loops were flawless. (I didn’t keep track of the count. The experience was too gut wrenching to worry about counting loops.)
The grapevine has it that our performance is amongst the top quartile.
The team is very upbeat, and feeling really good about the accomplishments. If only you knew how hard this team has worked every single day since arriving in California you’d not believe it.
I am so glad that Adrian “Fatty” Aucoin is here. After retiring CajunBot (the mama), we had thought we’d never have to haul a bot in a trailer. We sure were wrong. We really do not want to be driving a $2MM car in normal traffic. One fender bender, and we’d be history. So we have rented a flat bed car hauler. Fatty is a pro in hauling the trailer. He also serves as one of the drivers during traffic tests.
Besides renting a car hauler, we have rented another RV, a small one. This will serve as a ‘nap mobile’ for the team, especially useful when the team is working in the field. It came in real handy last night. I got to take a nap in it at from 0400 to 0600 (hrs).
Tomorrow we test in Area B. This is the longest test, and has some really interesting situations. Tonight will be another long night.
After today’s performance, we have a chance to be back in the running for the final. Tomorrow’s test is very crucial.
I need to run to a team meeting. After the meeting several members will go to bed, and others will head to work.
- Arun Lakhotia
Someone — send my little daughter a card
October 28, 2007 (2009 hrs), VICTORVILLE, CA. Took a three hour sleep. I am really glad Scott Wilson, my “lets go, lets go, lets go” guy is here. He has taken charge of the drum beat for the team. And he beats really fast.
I woke up and realized I had not talked to my little daughter Manjistha in three days. She is six years old. Last I talked she said, “Daddy you shouldn’t have gone on this stupid trip.” She had wanted to come with me, and I knew there is no way I could be attending to her as well.
Every day she likes going to the mail box, the real physical mailbox, looking for mail. Most of the mails she gets are Disney magazine and brochures. Every now and then she gets a birthday invitation. Any mail with her name on it causes a big excitement for her.
I have been meaning to send her cards. But the eight days I have been here are a blur. I have not yet sent her a card.
I’d like to appeal to send her a card. You could send it to my office address, and it will get delivered to my home.
My office address is on the following website: www.cacs.louisiana.edu/~arun
(Rochelle: Should you receive any mail, please pass them on to Enam. Given the open nature of this blog, I may be putting my daughter at risk if I put my home address on the website.)
- Arun Lakhotia
Angels driving the bot
October 29, 2007 (0010 hrs), VICTORVILLE, CA. If you do not believe in angels, this should make you a convert. After looking at the data from Area A test we are completely stumped why CajunBot did not ram into the traffic vehicles. She came quite close to running over some of the cars. But she always stopped, just short.
Here is a video of the test run on YouTube http://www.youtube.com/watch?v=iOfNGJw09Ng
Looking from outside, she looked like just an aggressive driver with little respect for other drivers. But cautious enough to brake at the right time.
But having looked through the data, it comes as a surprise that she avoided any vehicle at all. While she did not have a human driver, she sure looks she was bring driven by an angel.
Christopher Mire, the unassuming and quiet graduate student on the team, patiently trudged through the loads and loads of data collected by CajunBot, looking for an explanation for her aggressive behavior. His conclusion the vehicle was running WITHOUT its primary sensors, the Ibeo and SICK lidars. The physical sensors themselves were working, but our program that analyzes the sensor data to identify objects in the world had crashed. And by crashed it means, the program was not running. The sensors were generating data, but there was no program using the data. For most practical purposes, CajunBot was running blind.
CajunBot was running on her secondary sensors, the radars, two of which were acquired and mounted after we came to Victorville, CA.
It just cannot be a happenstance that we acquired two radars after arriving in Victorville, CA and those lidars save the bot from total destruction. There must be an angel watching over CajunBot.
Christopher has diagnosed the reason for the crash. When we run the bot in Cajun Field, a wide open parking lot, CajunBot sees a very few (between 50 and 100) objects. Our obstacle detection and tracking system was configured to track 100 obstacles. But in the test area, an abandoned air base, there is a lot of clutter, The area has buildings, trees, bushes, and other traffic. As a result the sensors see about 500 obstacles. This exceeds the max limit of 100 programmed into the system, and our obstacle detection program simply crashes.
The fix for this problem is rather simple, and has now been taken care of.
- Arun Lakhotia
Day 3 performance, thumbs down
October 29, 2007 (1322 hrs), VICTORVILLE, CA - Our Day 3 performance was not really good, but we knew the issues going in. We still need to take care of a few things, before we’d be ready for passing this test.
In this test, held in Area B, the bot is expected to get out of a start chute, navigate an open space, go through a narrow winding lane bounded by massive K-rails, drive around a traffic circle, and then drive into a very convoluted neighborhood. We do not know what all goes in the neighborhood, but the general guess is that it has stalled vehicles parked on the side of the roads. The bot needs to safely travel through that area, avoiding rear ending any vehicle. In this area is a parking lot. The bot has to find a designated parking spot, not run into vehicles parked into adjacent lanes, and then get out of the parking lot. Finally, in the area is an ‘obstacle zone.’ This is an open space, with barricades forming a gate. The vehicle has to find the gate and go through it.
CajunBot had trouble right out of the start chute. It was stumped by all the K-rails along the border of the open area, and decided not to budge. After two unsuccessful attempts, Suresh did some quick thinking and made some changes. That made CajunBot, go past the initial hesitation, travel through the winding lane, and onto the traffic circle.
Just short of completing the traffic circle, CajunBot came to a stop. For a few minutes you could see her attempting to take off, but stopping again. The trouble shooting team, consisting of Mark and Suresh, were rushed to the site. They attempted a few things, but nothing helped. By then we were out of our alloted 40 minutes.
The good news is DARPA announced a second round of tests in the remaining 2.5 days (Monday PM to Wednesday PM). Earlier they had announced the schedule only through Monday AM. We expected that after the first round of tests in the three areas, some teams will be selected as having passed the tests, and others will be given another go. It turns out all the teams are scheduled to run in the second round of tests. I can infer that to mean that there is clearly no perfect performer.
Inspite of all the long hours, the team is holding up well. Everyone is stretched to the limits, but no one is cracking. There are no outbursts, or name callings, or “I want my mommy” calls.
The clock is still on, and we intend to keep hustling.
- Arun Lakhotia
Alternator busted; replaced in an hour
October 29, 2007 (2330 hrs), VICTORVILLE, CA. I thought I’d never face this again after retiring CajunBot (I). I thought all the emergencies due to mechanical failures were history. Turns out I was wrong.
Around 1800 hrs I got a call from Little J. that the Jeep was not cranking. We got Joshua out of the bed and rushed him. He diagnosed that one of the alternators was busted. CajunBot has two alternators, one added after market by a alternator-holic in Alexandria, LA. One of the alternator is bought from Autozone and the other from High-Output Alternators (www.highoutputalternator.com). The latter is a speciality product.
By 1830 hrs Joshua, Adrian, Dallas, and Scott were headed to the pit, which is where our trailer is parked. The trailer, thanks to our contingency planning, held a spare alternator, the high-output one. By 1900 hrs, the broken alternator, which happens to be the high-output one, was replaced. Joshua has messed with these alternators for so long, he says he can change them in his sleep. Which is good, for the next three days I expect him to be walking in his sleep anyway.
Well then again, may be I was not wrong. The mechanical failure did not create any sense of emergency in the old-timers. I never stepped out of the hotel. The software guys stayed focused on their work. We were pretty confident that this shall also pass.
I have placed order for another HO alternator, just in case there is another failure in the next five days.
- Arun Lakhotia
Day 4 - Successful run in Area C
October 30, 2007 (1200 hrs), VICTORVILLE, CA. If you’ve been staying up biting nails, rest easy. CajunBot had a very successful run in Area C. This is the same area where we started our Day 1. Today’s performance was not picture perfect, but close enough.
CajunBot completed the mission, which involved making eight laps around a loop. The loop had two four way stops. At each four way stop, CajunBot was to come across other vehicles. She was expected to follow the precedence rules for stop signs. Vehicles that arrive before you at the intersection have the right of way, before you.
Unlike our first run, CajunBot did perfect this time, in all except one intersection. The good news is that she did not go out of precedence. Which means none of the drivers had to hit pedal to the metal to escape her kiss, as they had to for the first run. The only exception was when she was a bit too gracious, and would not take her turn. The situation was a bit complex. Three vehicles were already at the intersection before she arrived. One vehicle left, and another (a fourth) vehicle drove up and stopped at the same place. The other two vehicles left. Now CajunBot should have moved. But she did not. Finally, the fourth vehicle moved, at which point CajunBot moved on.
For all practical purposes this appears to be a minor flaw, say in comparison to chasing a driver out his wits.
Talking about drivers, I chatted up with one, to learn how he could put himself in front of these reckless bots. He, and other drivers in Area C, are moving stunt men. They get thrill in making close calls. But I thought movie stunts are choreographed, so the risks are a lot controlled. But in these tests, there is no way to guess when a bot may go beserk. I sure do not want to be in their position. And I am sure they think they same about me.
Other than the intersections, there was one more test in this area. The bot had to detect that a road was obstructed, turn around, and find another way to get to the checkpoint. The checkpoint was on the other side of the obstruction, so she had to get to the other side, make another turn and come back.
CajunBot was perfect in replanning.
In essence we had overcome two of three issues that stumped us in the first go. We had fixed our algorithm to use sensor data to understand the world. We had fixed our replanning algorithm.
The one issue that is still not fixed is our ability to deal with imperfect GPS data. In the last test, the bot was driving on the curb but, based on GPS data, thought was on the lane. We did diagnose the reason for the GPS ‘drift’. Before the bot was launched, she is brought in a staging area. One end of the staging area has large trees with thick canopy. This was blocking a significant part of the sky, thus limiting the satellites that were in direct view.
We addressed the GPS drift at least for this run. Joshua was monitoring the GPS data in the staging area. When we started receiving error messages, we asked DARPA officials to move us to a different location. This led to a lot of radio chatter, getting approval from the higher chain of command. The request was granted. So when the bot was finally launched, she had already recovered from the temporary loss of GPS.
Several of you have sent compliments on the video from Area A. The compliments go to Scott. He is a videographer, director, and editor all rolled into one. He is working on a video from today’s run. By the evening we should have an upload.
Tomorrow we have two runs. In Area A at 0730 hrs and in Area B at 1220hrs.
Pre-run sunrise (by Adrian):
Approaching front intersection (by Adrian):
- Arun Lakhotia
Day 4 - DARPA speeds up elimination; six out; called in for Area A
October 30, 2007 (1430 hrs), VICTORVILLE, CA. DARPA has turned up the heat, and sped up the elimination process. Six teams have been eliminated. The schedule for other teams is being compressed. Our Area A test was due tomorrow morning. We are now scheduled for 1700hrs today. I was asked to report at 1430 hrs (now). I said I couldn’t, my launch team is sleeping. I got an extension of a few hours.
Grapevine has it that Area A is the killer test. If you fail it, you are out. This test requires merging in traffic. CajunBot did well last time, so we are very hopeful. But then there are too many uncertainties when dealing with real-time systems. So cannot take anything for granted.
I have the team pulled out of bed. Actually, some of them had not slept anyway. Their mind was too occupied with the remaining tests.
Barring any unforeseen situation, I see CajunBot perform well in Area A. Our remaining weakness is Area B. We are hoping not to be called to Area B in the morning. We still have some work cut out.
- Arun Lakhotia
Area A/B tests both in the morning; thirteen
October 30, 2007 (1900 hrs), VICTORVILLE, CA. Rumor has it that thirteen teams have been cut. Area A is serving as the chopping block. We were called to repeat Area A at 1630hrs today, but there were two other robots ahead of us when we got there. Around 1730 hrs DARPA decided that the daylight was not good enough to run the tests. We are now scheduled for Area A test in the morning.
Our Area B test has been moved up as well. With all the teams eliminated, a lot of time has opened up. We are expected to run the Area B test back-to-back with Area A test.
We feel pretty good about Area A. Little John has been doing some very aggressive tests. He has shaken the system well enough. I feel if its survived his testing, it should survive DARPAs.
We have been discussing whether we be cautious or aggressive in Area A. Cautious means running limiting the speed to 10 mph, whereas an aggreesive stance will be to run at the track limit of 15mph.
At the higher speed there is a greater chance of making a dangerous turn, which is a no-no. At lower speed, CajunBot will be safe, but may be eliminated for not satisfying the mission.
We are a bit weak in the capabilities needed for Area B. That may change by the morning. Suresh and Pablo are focused on addressing our limitations.
For old time cajunbot members, DARPA has sure upped the ante. The tests have been really very rigorous. And if we look at the activity in progress to build the track for the final, the last two challenges look like kindergarten.
As I write this Danny Majors and Big John have gone to Subway to get dinners. I have a typed document of everyone’s Subway order. So the sandwiches will be personalized. People are tired of friend chicken, pizza, and cold cuts.
The team members have gone to take a nap until the food arrives. After that we start humming again. This could be the last night that pushes us in the final.
- Arun Lakhotia
Day 4 - Video of Area C test run
October 31, 2007 (0400 hrs), VICTORVILLE, CA,
Scott has placed video from yesterday’s run on YouTube:
http://www.youtube.com/watch?v=Xy5WnkfNnYM
You’d find videos from other teams there as well.
Chris Vaughn called last night. He was at the airport in Ontario. I was not expecting him until Friday, but am glad he is here. Scott put him to work right away, to join the testing crew.
I was woken up about half hour ago. Time to get to work. The software team is stuck on a bug related to parking/unparking. I think they are too exhausted to function. Scott woke me up to join Pablo.
We plan to send a minimal launch crew to our test in Area A, and keep the software people huddled up working on issues for Area B. We are literally going down to the wire.
- Arun Lakhotia
Day 4 - Got all the pieces together, but not in enough time
October 31, 2007 (1900 hrs), VICTORVILLE, CA. In case you have not heard it yet, CajunBot did not make it to the finalists. We did come very close, having finally gotten all the pieces of the puzzle. But we ran out time to do the fine tuning.
CajunBot did exit the challenge with a bang, literally. She knocked a car parked on the side of the road (as part of the test environment) in Area B, shooting the car across the road, over the curb, into a house.
In this test area, DARPA had created a neighborhood scenario, with 15-20 cars parked on both sides of the road. The robots were expected to weave through these vehicles. The DARPA Urban Challenge Requirements document did mention the need for having the capability to avoid a vehicle partly protruding in the lane. But the requirement was almost mentioned as an afterthought, like something good to have. The emphasis in the requirements document, as also the site visit evaluation, was on avoiding stalled vehicles on two lane roads.
Over the last few months W\we had focused on our energies on passing vehicles stalled on a road, essentially blocking a whole lane. It turned out that capability was not even relevant for the NQE.
After having finished Area A and C, we finally began focusing on the requirement for Area B. Yesterday evening we did not have a clue on how to handle driving through neighborhood. Since the start of NQE we had come up with various solutions, but none were satisfactory. However, last night things started falling in place. Around 2100 hrs (9:00pm) Pablo, I think, suggested using ‘particle field’ approach. In this approach the path of the vehicle is treated as particles. Objects (like cars) on the road act as forces that push the particles. This approach sounded very promising, and I encouraged him to code it up.
Around midnight he had a solution, but it did not behave as expected. It resulted in paths that would make the vehicle go way out of the road to avoid something.
In the discussion, Daro came up with the brilliant idea of treating the path not as independent particles, but as an elastic string. The particles of the string pull each other.
Pablo coded the solution, and at some point fell asleep. Daro kept ticking, tuning the system. By about 7:00am in the morning we could see that the method worked in a simulated environment.
By 10:00am we were ready to run in Area B. The run did not go off well. The bot did not even make the progress she had made yesterday.
We got back to the motorhome, and continued fine tuning the system. We were offered to run again about 2:00pm.
This time around, CajunBot was a completely different beast. She cruised through the initial segments, areas she was earlier hesitating to go through. She was clipping at 25mph, and looked real good.
But once in the neighborhood with all the parked cars, she clipped a car while trying to avoid it.
Data from the run indicates that the method Pablo and Daro came up with worked really good. The only limitation was it was not tuned well enough. The forces exerted by the objects were a bit too weak. So while CajunBot attempted to swerve, she did not swerve hard enough. In the process she could not avoid the parked car.
The fender bender knocked CajunBot out of the challenge.
While I am disheartened, I am not disappointed. I think this small band of boys from Louisiana has shown that they can take on a challenge, and have fun with it as well.
Dr. Tether’s parting words were, “will see you next time.” We may be down, but not out, if no one wins the challenge. We have gained considerable experience, and will come back stronger next time.
- Arun Lakhotia
Finalists announced, could have made it with less aggression
November 1, 2007 (1600 hrs), VICTORVILLE, CA. Earlier in the day DARPA announced the finalists. As expected we were not in the list of 11 that made the cut.
I have not kept up with the performance of other teams. My focus has been on moving our team forward.
But based on what we hear, it looks like we went a bit too aggressive on Area B. Our goal was to finish that area, just as we had finished the other two. We knew we could not complete one segment of Area B, and worked on developing the capabilities needed. Had we left that capabilitiy alone, CajunBot would have just reached that troublesome area and stopped. That could just have us qualified for the finals. The most overriding requirement for any vehicle to be in the challenge is SAFETY. By going aggressive we ended up violating that requirement, and getting knocked out all together.
Almost a year ago when I was getting the team together, the intent was to shoot for a win. We had already been in two challenges. In the last two challenges, we knew going in that we did not have the chance to win. We were quite happy being in the challenge. It was not the same this time. The excitement of being in the challenge had worn out. Besides, based on past experience we knew we had the capability to win. In the past we just did not shoot high enough.
We took the neighborhood scenario of Area B as a capability needed to win. Besides, since we completed the other two areas, it made sense that we complete this scenario as well.
It turns out that some of the teams that qualified for the final did not even reach the neighborhood scenario in Area B. To their credit, they were safe. If their vehicle had a chance of hitting something it would simply stop. That’s how CajunBot would have been had we not taken the neighborhood scenario seriously.
I just hope that the tests DARPA created truly reflect the actual challenge. It would be a shame if they mellow down the challenge so that someone can win.
- Arun Lakhotia
Majors family arrives for a Cajun cookout
November 2, 2007 (0730hrs), VICTORVILLE, CA. In keeping with the tradition established in the first Grand Challenge, the Majors family arrived yesterday for the Cajun Cookout. The entourage is led by Brother Ray Majors, the person who gave this project the first break. He gave us the 6-wheeled ATV that got us rolling. He is joined with his brother John Majors who, through his company Quality Transport, Inc, has been kindly supporting us by providing vehicles and trailers to haul the bots back and forth from California. Mark Majors, of MedExpress Ambulance Service, another sponsor, and his youngest son close the pack along with Danny Majors, who is part of our team.
In the backdrop of the several ‘Did Not Qualifiers’ (DNQs) that have packed and left, the Majors family continuing the cookout is an anathema. But it further reinforces the Cajun spirit already very well known in the DARPA circle. We play hard, lick our wounds, and enjoy life nonetheless.
The veterans of the challenges, which includes DARPA officials, media, and other team members, look forward to the event. We were asked at checkin whether we were having a cookout. The same question was popped up by many when it was announced Team CajunBot DNQed.
The Majors woke up early and took off shopping. They plan to serve about 200 people. That calls for buying and cutting a load of onions, celery, bell pepper, garlic, and sausages. While all of these ingredients will be bought locally, they have carried with them two pounds of crawfish. In keeping with the tradition the crawfish has been sponsored by Andre Leger of Chez Francois.
The Majors will spend the next four hours cutting all the ingredients and cooking. They will cook right at our hospitality tent. Cooking with open fire outside in the hubub of activity is a very big deal. Its a big deal when you know that millions were displaced a few weeks ago by forest fires. This area is a tinder box. One spark could get the whole place aflame. But, in keeping with another tradition, DARPA went out of their way to get permit for the open fire cooking. The fire marshall is invited.
We started cranking again. Today we will work on getting CajunBot opertional again. The hit has misaligned her sensors. After realigning the sensors, we plan to do some test runs, and then bring her offroad (where a Jeep really belongs). Our goal is to collect data for conditions we cannot find in Louisiana.
- Arun Lakhotia
darpa uc 2007 & news 02 Nov 2007 12:08 am
DARPA selects 11 robotic grunts to take driver’s license test
DARPA selects 11 robotic grunts to take driver’s license test | The Register
Shudder. Have you seen R2D2 parallel park?
By Ashlee Vance in Mountain View → More by this author
Published Thursday 1st November 2007 20:44 GMT
DARPA applied a firm hand when narrowing down the list of robot vehicles that can take place in the weekend’s $3.5m Urban Challenge. The organization - an arm of the US Defense Department - has deemed only 11 out of 35 competing teams as worthy to take place in the final event.
The teams have spent the past week showing off their so-called autonomous vehicles at a retired Air Force Base in Victorville, California - a small city near Los Angeles. And DARPA has not been joking around with the test runs. The vehicles have been forced to perform tasks such as parking, passing slower vehicles and even merging with oncoming traffic.
These skills prove far more demanding than those required of vehicles during DARPA’s past robot races. In 2004 and 2005 events, the unmanned machines had to make their way at top speed across desert courses where the major hurdles included avoiding boulders and ditches. The vehicles were and are guided by GPS coordinates, along with a host of sending gear such as radar, lasers and fancy software.
Stanford University won the 2005 event and has again made it to the finals of the Urban Challenge. Carnegie-Mellon University - another strong competitor - has made the finals as well. On the corporate front, Oshkosh Truck - a maker of massive vehicles - has garnered a spot in the finals. Team Oshkosh was one of five to finish the 2005 race. No vehicle made it more than a few miles in the 2004 race.
Other finalists include teams from Cornell Univeristy, MIT and Honeywell/Intelligent Vehicle Solutions.
TeamUCF made the event and described the selection showdown.
The first announcement that they made was that there would be only 11 teams in the finals. That was a far cry from the 20 teams they expected … and worse, 6 of the 11 were already known. The crowd was very quiet when this announcement was made. That left 5 slots for about 15 good teams, all of which were in the audience.
These slots went (in order) to MIT, TeamUCF, AnnieWay, Honeywell/IVS and Oskosh which were added to CMU, Stanford, Ben Franklin (UPenn), Carolo, Cornell, and Virginia Tech. Of the 11 finalists, 8 had received the “Track A” award of $1M to build their car. Only three of us, AnnieWay, Corolo, and TeamUCF did not and of those only 1 is from the US.
DARPA seems to have this robot race stuff down and is looking to keep the field tight in order to avoid past disappointments.
From what we’ve seen so far, these are some very impressive vehicles and Saturday’s race looks to be a show stopper.
The cars will have six hours to go around a 60-mile course, preforming a variety of tasks. All told, they will have suffered through the equivalent of a California driver’s license test. [Please hold your jokes. Oh, what the hell, fire away - Ed]
As DARPA tells it,
From the time each robotic vehicle leaves its starting chute and begins the course, it is entirely under control of its onboard mission computer - human observers may intervene only for purposes of safety. The entire field of robotic vehicles will be on the course at the same time, interacting with one another as vehicles in urban areas across America do each day. The vehicles will face driving challenges that include traffic circles, merges, four-way intersections, blocked roads, parking, passing slower moving vehicles, and merging safely with traffic on two- and four-lane roads.
We’ll have live coverage on Saturday.
darpa uc 2007 01 Nov 2007 11:59 pm
Stanford, CMU among robot race finalists
Stanford, CMU among robot race finalists | Tech news blog - CNET News.com
A robotic truck, Prius, and family sedan will be among the 11 autonomous vehicles trying to pass their driving test this weekend in Victorville, Calif.
The Defense Advanced Research Projects Agency (DARPA) said Thursday that it has selected the finalists for its Urban Grand Challenge, a robotic race at the former George Air Force Base with $3.5 million in total prize money. The 11 teams, whittled down from 35 semifinalists in qualifying rounds this week, will compete Saturday. They must drive a 60-mile urban course with live traffic in less than six hours.
Among those in the race are: Stanford Racing Team, the 2005 DARPA Grand Challenge winner; the Ben Franklin Racing Team, from Philadelphia; MIT; Carnegie Mellon University’s Tartan Racing team; Team Cornell from Ithaca, N.Y.; and Team Oshkosh Truck.
Those eliminated in the qualifying rounds included: Team CalTech; Princeton University’s team; Team CajunBot from Louisiana; and Team Gray from Metairie, La., which was one of the robots that performed best in summer trials.
“The (qualifying event) tested the vehicles capability to merge into traffic, navigate four-way intersections, respond to blocked roads, pass on-coming cars on narrow roads, and keeping up with traffic on two- and four-lane roads. In fact, the only major difference between the (qualifiers) and the final event is that other robotic vehicles will be part of the traffic in the final event,” DARPA Director Tony Tether said in a statement.
darpa uc 2007 & news & personal 01 Nov 2007 11:04 am
CajunBot continues in qualifying
The Daily Advertiser - www.theadvertiser.com - Lafayette, LA
VICTORVILLE, Calif. - After a few roadblocks during Saturday’s qualifying round of the Department of Defense’s 2007 Urban Challenge, Team CajunBot spent the night in the field. The sleepless night paid off Sunday.
The 12:30 p.m. qualifying round Sunday had CajunBot II finishing 13 laps without swapping paint with the five other moving vehicles on the track - a feat not many bots were able to perform.
“These guys are amazing,” Dr. Arun Lakhotia, team leader, said in a news release. “Some of them are working on no sleep and others only got about two hours of sleep last night. CajunBot II’s great run today is a result of their continuous dedication to this project.”
On Saturday, the bot struggled with a course of intersection right-of-ways and eventually paused.
The team packed into a makeshift lab - inside the trailer that transported CajunBot II - to analyze video from the bot. Adjustments were made and on Saturday, the bot performed almost flawlessly.
Bright and early this morning, CajunBot II will have to park and perform other driving maneuvers in its last qualifying event.
Finalists are expected to be announced Thursday. The top 20 teams from the 35 semifinalists will compete against one another on Nov. 3.
Bots will have to travel in a city setting - obeying all California traffic laws - at the former George Air Force Base.
The first team to finish the race will receive $2 million with second receiving $1 million and third $500,000.
This is the third Grand Challenge put forth by the DOD’s Defense Advanced Research Projects Agency. Team CajunBot has participated in all three races.
darpa uc 2007 01 Nov 2007 10:45 am
DARPA Grand Challenge: Million-Dollar Babies
DARPA Grand Challenge: Million-Dollar Babies
More than a year ago, DARPA awarded 11 teams $1 million each in seed money for vehicle R&D. Of the 60 applicants, Stanford, MIT, Carnegie Mellon, Autonomous Solutions, Virginia Polytechnic, CalTech, Cornell, The Golem Group, Honeywell, Raytheon and OshKosh were the only ones with proposals deemed worthy of the prize. So, what’s a million dollars get ya these days? Anything from a mostly stock Toyota Highlander Hybrid to a gigantanormous robotic monster truck that belches Terminator sounds when in autonomous mode. Every team has a unique approach, and while similar hardware shows up on many of the vehicles, nobody seems to have the exact combination to match anyone else. Case in point. Last year’s winner thinks it can get the job done without any optical cameras; Autonomous Solutions figures Windows XP is up to the task of commanding an SUV. Things are bound to get interesting on the A Track, and we’re not even to the privately funded B Track entrants.
darpa uc 2007 01 Nov 2007 10:33 am
Video: Autonomous Prius rebels against its masters, burns out and crashes
TG Daily - Video: Autonomous Prius rebels against its masters, burns out and crashes
Victorville (CA) – This could be world’s first autonomous burnout and crash of a Toyota Prius. At the DARPA Urban Challenge qualifications in Victorville California today, the Golem Group’s “Golem 3” bot ran most of its course smoothly, but after a computer glitch and code readjustment, it turned sharply left, burned out its tires and crashed into the curb. The crash hurtled the car over and flattened one tire.
darpa uc 2007 & news & personal 01 Nov 2007 10:21 am
CajunBot II sidelined in $2 million race
The Daily Advertiser - www.theadvertiser.com - Lafayette, LA
Vehicle stumped on navigating neighborhood street
Marsha Sills
msills@theadvertiser.com
CajunBot II did what it could, but it wasn’t enough to please DARPA officials Wednesday.
The autonomous vehicle designed by UL students and researchers was sidelined late Wednesday, ending the team’s hopes of competing in its third race of the country’s driverless vehicles.
The race is sponsored by the Department of Defense’s Defense Advanced Research Projects Agency and was created about three years ago to spur private research in the area of autonomous vehicles.
On Saturday, teams from across the country will race in Victorville, Calif., for a $2 million prize.
For the past three days, 35 teams have been testing round after round on a former military base in Victorville for one of the 20 spots in the final race.
But CajunBot II wasn’t prepared for one of the obstacles - how to navigate smoothly and at the appropriate speed down a neighborhood street lined with parked vehicles on both sides of the lane.
The team scrambled to write a program to give the bot the capability, but they needed more time to test it, said Arun Lakhotia, Team CajunBot leader.
At 1 a.m. Wednesday morning, a code was written to help the vehicle negotiate the new obstacles and by 6 a.m., the new programming was in place.
“We had a comeback, but we didn’t have enough time to tune the parameters,” Lakhotia said. “It was a matter of tuning everything right. The key thing was we pulled it all together in time for this, but for the tuning, we needed more time.”
On Saturday, the Department of Defense’s Defense Advanced Research Projects Agency’s third race of autonomous vehicles will be held in Victorville, Calif. While in the past two years, the race has been held across desert terrain, this year’s challenge tests the vehicles’ ability to navigate through city streets.
Earlier Wednesday, CajunBot II showed that it could merge in and out of traffic circles. In that test it shared the roadway with five vehicles traveling in an inner loop and five in an outer loop. The team’s Jeep Wrangler was able to merge in and out without any glitches.
Later in the day, while testing its ability to navigate through the neighborhood the bot went off course.
“It’s been a pretty amazing run,” Lakhotia said. “It’s been a difficult run. I knew it would be hard, I didn’t know it would be this hard.”
The team worked through the past few days on too few hours of sleep and had last minute hiccups like the need for a new alternator.
Because of the degree of difficulty in this year’s competition, Lakhotia suspected that there may not be a full roster of 20 teams competing Saturday. The finalists will be named today.
The team will stick around to see the race play out on Saturday.
They’ll also continue their tradition of serving their competitors spoonfuls of crawfish étouffée - courtesy of the Majors family - on Friday.
“We want to see how the others are doing,” Lakhotia said. “We’ll need to know what they did right and what they did wrong and we’ll learn from them. We came here for the challenge and we want to see how it goes.”
darpa uc 2007 01 Nov 2007 10:08 am
DARPA to Robot Car Builders: Don’t Hit Anyone!!!
DARPA to Robot Car Builders: “Don’t Hit Anyone!!!” | Danger Room from Wired.com
The qualifiers for DARPA Urban Challenge robot car race got underway this morning here at the Southern California Logistics Airport in Victorville, California.
DARPA director Tony Tether outlined the basic rules for competing teams in a briefing last night, concluding with the most important rule of all: “Don’t hit anyone!!!”
That’s with three exclamation points, in about a 10,000-point font size, just so no one would miss it.
Thirty-five teams from around the country are here to compete for $3.5 million in prize money. The winning vehicle will have to be the first to complete a 60-mile urban course in less than six hours, with no human input at all.
Oh, and it’ll have to obey all California traffic laws–signal turns, obey speed limits, negotiate four-way stops with other vehicles…think of it as the kinder, gentler auto race.
The main event is set for November 3. Meantime, judges are busily putting the cars through their paces to narrow the field of competitors to just twenty.
First test: successfully demo emergency stop capabilities. Each car has been outfitted with big red stop buttons on the outside as well as a wireless kill switch so that judges can remotely shut down any vehicle that decides to go charging off the course.
Not that a rogue car would get far. We humans aren’t allowed on the course, and waist-high concrete barriers will keep us safe. We hope.
Starting tomorrow, the vehicles will move to more advanced testing to demo left turns, traffic merges, navigation, and…parking. Yes, parking. Like I said, definitely not your usual auto race.
darpa uc 2007 30 Oct 2007 11:41 pm
DARPA eliminates six more teams from Urban Challenge race
TG Daily
Victorville (CA) – DARPA officials have just eliminated six more teams from any further qualification runs and the Urban Challenge race itself. The eliminated teams are 1. Caltech 2. Mojavaton 3. Team Jefferson 4. Team Urbanator 5. Gator Nation 6. Team Juggernaut.
Four teams are sitting pretty because they have already qualified for the finals and won’t have to do anymore qualifying runs. Those teams are Cornell, Stanford, Tartan Racing (Cargenie Mellon University) and Virginia Tech.
Earlier in the day, six other teams were eliminated. It’s going to be a nail-biter night and day tomorrow as the 23 remaining teams try to impress DARPA judges for the final 20 race spots.
darpa uc 2007 30 Oct 2007 11:38 pm
Seeing the World Through Robot Eyes on Danger Room
Seeing the World Through Robot Eyes on Danger Room
Those of us who here at the DARPA Urban Challenge who haven’t built a robot car lately got a nice little seminar on the subject. The teacher: Stanford Racing Team software lead Mike Montemerlo.
You can see a view of the way the team’s Volkswagen Passat TDI, called Junior, sees the world through its laser eyes over Montemerlo’s shoulder.
The team likes lasers rather than camera-vision because they generate their own light sources rather than relying on fickle sunlight, with its constantly shifting shadows. They also work just as well at night as during the day.
To build up a continually-updated picture of the world, Junior’s LIDAR units use the scattered light from their lasers to build up what Mike calls a 3D data cloud of rings at varying distances. As the rings shift positions with the cars movements, Junior can separate moving objects from stationary ones.
Junior_estop
To drive, Junior relies on rules about the relative cost of various maneuvers. How costly is a lane-change? Depends on if there’s another vehicle in that other lane, where it is, and how fast it’s moving.
For the Urban Challenge, team engineers will have Junior drive cautiously, though in after-hours simulations, they enjoy torquing it into the aggressive driving style they call Rambo Mode.
Stanford is considered by many the team to beat, since it won the Grand Challenge in 2005.
But victory is by no means a given. Even in E-Stop testing, the first of the trials the robot cars have been put through here at the qualifiers leading up to the Urban Challenge race itself, the Junior displayed the same kind of glitches experienced by other teams.
In this case, Junior’s computers inexplicably stopped collecting Controller Area Network, or CAN, data that shows things like wheel position and other essential features of the car’s state. Without the data, the car refused to budge after planned pause.
The minutes ticked buy while the engineers worked the problem. Finally, with only two minutes to spare before their time for testing was up, the rebooted the computer charged with collection the data, and Junior came back to life.
darpa uc 2007 & personal 30 Oct 2007 11:34 pm
#24 Team Cajunbot - Cajunbot II on Flickr - Photo Sharing!
#24 Team Cajunbot - Cajunbot II on Flickr - Photo Sharing!
Taken in (See more photos here)
LIDAR overload from a quartet of Ibeo units. It’s kind of hard to tell what advantage, if any, this will give.
darpa uc 2007 & news 30 Oct 2007 11:22 pm
Typical Bot Workday - AIMing for Automated Vehicles
Typical Bot Workday - AIMing for Automated Vehicles - Blog on Control Engineering - 1180000318
Typical Bot Workday
Each of the thirty-five DARPA Urban Challenge bots are being put through their paces, here is CajunBot II’s day, which is typical of what each of the other teams and vehicles are doing.
CajunBot II’s ability to park and obey traffic signals was tested at 7:30 a.m. today. The bot made it out of the starting gate and around a curve to a stop sign. It put its blinker on, turned right then made a left at the next stop sign. It traveled a bit more then paused. A partition in the road which reduced from two lanes to one caused the bot to stop right as the 40-minute time trial was ending.
“We’ve gotten a lot further this morning than a lot of other bots,” said Scott Wilson, Team CajunBot advisor. “It’s not about pass or fail in qualifying rounds. Its how you solve problems you encounter.”
The team was heading back to a makeshift lab in the pit area for a debriefing on this morning’s run. Cameras inside the bot record its every move. A simulation program developed by Team CajunBot is used to analyze data collected from computers in the bot along with the video.
http://www.youtube.com/watch?v=iOfNGJw09Ng note: once at this location there are many driverless vehicle videos to watch so plan to be there a while and check some of them out.
“We’re going to go back and look at the data we have. We’ll make the necessary adjustments and then begin planning for tomorrow’s qualifying round,” said Wilson.
Team members were also hoping to catch some sleep today also. Many have only had a total of four hours of sleep since qualifying began.
“We’re working on this day and night,” said Wilson. “The talent and teamwork we have on this team is unmatched.”
Late yesterday, officials with the Defense Advanced Research Projects Agency announced additional qualifying rounds. Each team will once again run through the three testing areas. Tomorrow morning, Team CajunBot will head to Area C where the bot has to react to oncoming traffic in intersections.
Testing will begin once again at 7:30 a.m. Tuesday morning.
The top 20 teams selected as Finalists for the 2007 DARPA Urban Challenge Race will be announced Nov. 1, 2007 and will compete against each other on Saturday Nov. 3, 2007. Prizes in Saturday’s race are $2 million for first place, $1 million for second place and $500,000 for third place. All thirty five teams were given a participation award and bragging rights the first day of the elimination trials.
GO ROBOTS !
darpa uc 2007 & news & personal 30 Oct 2007 10:59 pm
DARPA set to give $2M to the greatest robot car racer
DARPA set to give $2M to the greatest robot car racer | NetworkWorld.com Community
Submitted by Layer 8 on Fri, 10/26/2007 - 12:07pm.
Thirty-five driverless vehicles will race over hill and dale as well as faux city intersections next weekend in the Defense Advanced Research Projects Agency (DARPA) Urban Challenge held on a former airbase in California. The National Qualification Event will take place at the same location this weekend October 26-31, 2007.
DARPA says its third-annual Urban Challenge program has the lofty goal of developing technology that will keep soldiers off the battlefield and out of harm’s way. The Urban Challenge features autonomous ground vehicles maneuvering in a mock city environment, executing simulated military supply missions while merging into moving traffic, navigating traffic circles, negotiating busy intersections, and avoiding obstacles. The machines must also handle parallel parking and intersections with two- and fou