Dear Supporter of the UC Bearcat DARPA Urban Challenge Team for 2007:

            I want to thank you for supporting our team for the DARPA Urban Challenge. We succeeded in making a cooperative research and development agreement with the National Automotive Center of TACOM and getting a vehicle. Many thanks to Jerry Lane, Kevin Mills, Bernie Theisen and others for supporting our effort.

            We also modified the steering, speed and brakes to computer control and submitted a video to DARPA for judging in May: http://www.youtube.com/watch?v=aPsKsbLdBRc However, DARPA said it did not demonstrate autonomous operation and rules us out. See my Blog at: http://madeinexpress3.spaces.live.com/blog/ for details.

            We also had vision sensors, a SICK laser scanner and a Novatel GPS.

However, we also had some problems that slowed down our progress.

            The vehicle we got was great and is powered by two 7.5 hp electric motors. However, it was not a production vehicle. So our students had to add head and tail lights, an emergency brake, a safety glass windshield and get it inspected by the state patrol to get a title then apply for a license and get insurance coverage. It is now street legal and on the UC vehicle property list.

            One technical problem we encountered was the motors on the jeep would cut off when we went up a hill or went very fast. This summer we took the jeep to OKI systems and they helped us discover that the speed controls for the front wheels and rear wheels were set with different control codes which made the wheels go at different speeds which caused the cutoff. Many thanks to Ed Curran and Charlie for helping solve this problem and giving us a teach pendant controller for the speed controller.

            Another technical problem was we used our old Galil DMC 1000 speed control which is a plug in board using an ISA slot and we couldn’t find a modern computer with these. So our students linked two computers and actually got this to work. However, it wasn’t reliable. Kovid and Scott are still working on this problem and may have to buy a new PCI bus Galil controller.

            Another technical problem was with our order for a Novatel GPS. By the time we got the funds to order it, it was too late. It didn’t come in until June. Scott now has it working and it will still help us on the Cub for the IGVC.

            Our software team did great but was not able to show it with the hardware not working. It was a great experience and we all learned a lot.

            Tomorrow is the DARPA Urban Challenge. They will have a webcast starting at 10:30 Eastern time which out team will avidly watch. http://www.darpa.mil/grandchallenge/

            If there is another such challenge in 18-24 months, should we try again? I would appreciate your opinions and advice. Thank you again for your support. The educational value of this effort was well worth it to all of us and I think we are better engineers for the experience.

 

    I finally got an electronic copy of our Robotics A User Friendly Introduction with Chinese translation posted on my web site at: www.ernesthall.com

If you click on Robot Book tab on the right side of the page you can download a 560 page, 123 Mbyte pdf file that has both the English and Chinese versions of this book. The Chinese translation was done by Professors Zuoliang Cao of Tainjin Institute of Technology, Professor Yu Yu Huang of the Beijin Institute of Astronautics and Aeronautics  and Professor Liu of Beijin University. ( Please check the Chinese translation for correct affiliations).  Professor Cao has authorized the release of the translation.

     Bettie and I wrote this book on our first home computer, a TRS 80 Model III which I bought with a $1,000 teaching award. The word processor was Scripsit and a far cry from Word 2007. Our editor was John Beck and he found the wonderful cover photo that was done by Larry Williams and Alvy Ray Smith of Star Wars fame. He scheduled a visit to Cincinnati as our deadline drew near and Bettie and I worked in shifts 24 hours per day to finish the manuscript on time. We got it done and had a nice dinner visit with John and delivered the manuscript. I am still teaching essentially the same content of this book in my course, Introduction to Robotics, at the University of Cincinnati, with updates of course. For example, we have just gone over economic justification and I saw that in 1985 our electricity cost $.04/kW hour. Now it is $.12. We also now use an Excel template for computing the internal rate of return rather than the hand iteration we used to do. The U.S robot market passed one billion dollars a few years ago and I shopped worrying about it. I believe it is proven technology but we stall have a way to go to make intelligent robots.

    

   Well we are off to a fine start for the 2007-2008 school year.

As the photos show 10 of our UC Robot Team took throo of our robots in the UC Homecoming parade. This was really an exciting year not because of our robots but because our footbal team had won all of their starting games and were ranked nationally. Plus, the new footbal stadium and other campus facilities just put everyone in a very good mood. I told one new graduate student that there was a new exercise facility that even had a hot tup. He said "what is a hot tub.? See that is why we still need teachers. I am posting photos taken by Saurabh.If anyone wants to add more just send them to me.

   I and our team are getting over not being in the DARPA Urban Challenge contest set for Nov. 3 in California. But in a way I am relieved. It is a lot easier being a spectator than a player. I don't know how we would have gotten the funds to go to California for three weeks. There is a way since our Solar House team have taken their three trailer solar powered house to the Mall in DC for three weeks. I understand they raised $240K.

   Three of our robot team, Saurabh, Scott and Kovid, and I attended the 25th Intelligent Robots and Computer Vision Conference in Boston Sep. 8-11. We presented three technical papers about our research- one on eclectic control, one on our IGVC algorithm and one on creative control. We will post these on our robotics web page at www.robotics.uc.edu when we get time.

    I was also able to visit with some old friends including Professor Zuoliang Cao from Tianjin Institute of Technology who now has a third generation omnidirectional vision guided robot; Tony McPherson of the MIT Draper Lab who presented an excellent paper on a Mars Rover computer vision system and Lakshmi Srinivasa a 1997 graduate who introduced me to his wife and new daughter.

    SPIE has decided not to have a conference in Boston in 2008 so Professors Dave Casasent, Juha Roening and I have to come up with a plan for the future. Maybe intelligent robots and computer vision will be combined with Defense Applications in Orlando or Electronic Imaging in Palo Alto or just fade away.

    When we started this conference in 1983, it was clear to me that there were many potential applications of intelligent nachines that combined sensors and controls with mechanisms. One clear example that I had been involved with earlier was the computed tomography medical imaging system. Another was the cruise missile systems. I still believe that we are just starting on the really important applications for industry, defense and medicine. The general idea is good but we need to forsee the specific breakthrough technology developments. Will it be in mechanisms or sensors or computers? I believe the government is emphasing sensors and certainly these have a major impact. Small cameras that could be inserted in the human body, sensors to make cars safer and sensor information to help us live easier would be great.

    I recently noticed that Picasa is permitting a map location for each photo I upload to google. Here are Lakshmi's photos from Boston:

You are invited to view LakshmiNarasimhan's photo album: An Evening with Dr. Ernie Hall

    Now anyone who has taken my robot vision class knows that a picture is worth about 108,000 words. And if we stored all of the images we see we would fill up our brain in about 4 months. So the problem is to extract relevant information from a background of irrelavant noisy data.

    Since we didn't qualify for the 2007 DARPA Urban Challenge on Nov. 3, 2007, the best that I hope to do is attend as a spectator. I am still trying to figure out what happened to our jeep. One possibility is that our speed controller had a thermal cutoff that made the jeep stop when it had too much of a load. This is being checked out now. Our team got discouraged when the jeep had a problem we couldn't fix. We are still working on it.

   In the meantime DARPA has put out some more challenges:

These are more mathematical and less hardware oriented.  Maybe we will have better luck on theory than application.  

    As I get ready to start classes this Wednesday, I guess I will get the Blog started again. Send me any comments or questions you have. Let's keep the technology developing.

 

We just returned from Detroit and the IGVC. We were there from last Friday, June 8 through Monday June 11. It was quite an event and the results will be posted on the offical web site at www.igvc.org as soon as everyone's sun burn heals. Our team did great and won one of the JAUS awards and placed 6th or 7th in the two main events. I believe there were about 38 entries and 22 of these qualifieds to run.

We conducted a Lessons Learned by individual questions and a survey form and here are some of the results.

The Survey

A survey was sent out to all the team members to capture their viewpoint. This survey contains questions in the areas of project planning, project execution and delivery, human factors and overall. Here are some summarized results from the survey.

A large majority of the members felt that the project was planned well but was not executed as professionally. To be more specific, a significant number of the team members felt that the objectives of the project were specific and attainable. They also realized that the project is result-oriented and time-oriented and its outcome is measurable. A majority of the team agreed that the plan and schedule initially drawn up were well documented with good structure and detail. The schedule also included all aspects related to the project and the tasks in the schedule were defined adequately. The specifications and requirements were clearly defined and documented.

But when it came to execution, most of the members blame constant wandering off the objectives as costly blunders. The changes that were introduced were not manageable within the scope of the project. Moreover, these changes were not introduced formally through a change control process thereby leaving the rest of the team in the dark. There was no quality control process in place. Absence of a visible management hierarchy was also pointed out by several members as a possible point of concern. Regarding human factors, many members thought that there were conflicting goals between the different departments, which caused some problems.

1.1.2.   Positives

The positives to carry out of this endeavor are given below:

·       The team has enough experience now to put together a working vehicle for future competitions.

·       The art of project planning is already in place. If the execution phase can follow as per milestones defined in the plan and changes to be made are arbitrated through known change control processes, overall efficiency would improve manifold.

·        Proper delegation of authority and reporting structure is required for the team to accomplish its goals. A clearly organized management hierarchy with well-defined tasks for each team and delegation of work would give the team members a sense of belonging and accomplishment.

·       The team has enough expertise to accomplish the goals for the project. It also comprises of hard-working members who are result-oriented.

   Future Scope

Though the DARPA Urban Challenge team did not succeed in accomplishing its goals, the feedbacks obtained shows that there is an increased will to get back on its wheels for future competitions. It would benefit the team to improve on the transparency of its overall operation, provide better communication tools, be more focused on its goals, have checks and bounds to keep the project on its schedule and take milestones seriously. As Mahatma Gandhi once said, “The future depends on what we do in the present”.

 

 Conclusions:

For a project as complex as developing an autonomous ground vehicle, well defined project management methods like Gantt chart and PERT can make all the difference between success and failure. In addition to completing the project successfully, proven planning and agreement between the different groups enables the manager to achieve the agreed objectives. Also this provides clarity towards the goals and the control measures, saves both time and cost, helps co-ordinate resources and identify and manage risks.

As is shown in the case study, an absence of defined management hierarchy, change control processes, and adherence to documented milestones and associated timelines resulted in failure of the project. However, on the up side, there were many positive aspects of the project which could be the foundation for success in the future. As is once said by Confucius, “Our greatest glory is not in never falling but in rising every time we fall”.

 

Kovid and I took the jeep to a show and tell meeting in downtown Cincinnati and met some interesting visitors.

Scott and Saurabh tested the Cub on campus. Kovid made a video and put it on YouTube with an Elvis song in the background. Neat!

 http://www.youtube.com/watch?v=yfuYZBMwOTQ

 

Our new challenge is the Intelligent Ground Vehicle Contest, www.igvc.org, that will be held at Oakland University near Detroit on June 8-11, 2007. I am attaching a video from the test track at Selfridge last year for the Autonomous Challenge ( line following with obstacle avoidance) part of the contest. We are now designing a test track at UC for our team. I have also attached a video of the GPS navigation which is like the Navigation Challenge part of the IGVC. This clip is from our DARPA site visit in 2005.