bfd schedule

October 2, 2009 - 8:32 pm Comments Off

So you are a BFD fan, and you want a chance to see it in action before it goes to compete in the competition, what do you do? You go to one of our remaining beach test days of course!

Due to a bust by the california state parks and beach law officials, we are no longer able to test at dockwieler beach state beach. That’s okay though, because we scoped out some new spots in long beach, which are city beaches. We strong armed our inside guy in the city of long beach, and got an informal go-ahead to test down there. Okay, I’m kidding about that last part. It’s easy to ask for forgiveness than permission right?

Remaining test days:

  • 10/3/2009  – hard sand test
  • 10/4/2009 – nominal runs
  • 10/10/2009 – last test day / pre challenge party

On the 10th, we plan to spend the whole day at the beach (usually we just got for a couple of hours). We figure that it will be the last time the vehicles are operating before the competition, and we will go out in a bang. Family and friends who can’t make it up to the challenge are invited to come and join in on the BBQ.

Anyone not directly affiliated with a team member should send an email to contact@terraengineering.org for more details on the beach days.

t minus 2 weeks

October 2, 2009 - 3:50 pm 2 Comments

Well, we have about 2 weeks until the competition, and things are definitely heating up.

We continue to do 30 minute test runs with a configuration as close as possible to the competition. We continue to get excellent performance, and we are very confident going into the competition. The team has really come together in putting things into the final configuration. It’s not easy to work a fulltime job and compete in this competition. Everyday from now till the competition has been planned out. The schedule is very reasonable and we have some built in margin for any surprises.

The vehicle is mostly the same, with the exception of new steering motors. We had noticed that the current draw on the motors (especially the rear ones when fully loaded) was exceededing 5amps. For various reasons, this is a problem. Todd machined adapters and we will do formal testing with the new motors this weekend. Also, one of the  gear boxes was disassembled and we noticed a decent amount of sand had sneaked inside. All of the gear boxes have since been opened, cleaned, and sealed.

On another note, we are very excited about other teams posting videos and pictures of their vehicles. Teams like The Invading Huns, Pauls Robotics, Team Waldbaum, GreenCheese, and CMU. We particularly like GreenCheese’s solution as it has so many mechanisms, and engineers love mechanisms. Our only wish is that more teams would post more videos and pictures. We know you check out our site, how about sharing some of your cool work? Of particular note is CMU, they visit our website almost daily, but haven’t posted a new video in a month and a half.

Spirit of the Competition

September 22, 2009 - 3:44 pm Comments Off

While I was sitting and waiting for batteries to charge, I started thinking about the “Spirit of the Competition”. It is a phrase that I have heard quoted in virtually every engineering competition I have been in. I am approaching twenty engineering contests (Robot Wars, Battlebots, Robotica, FIRST, DARPA Grand Challenge, Regolith Excavation) and there always seems to be three types of contestants in each. The first is the purist who attempts to create an entry that satisfies not only all of the written rules, but the stated intent of the sponsoring organization. The second is a builder that has a design in mind and only modifies it to satisfy expressly written rules. The final type is the participant that starts by pouring over the rules looking for unintended holes to exploit. Unfortunately, the more inexperienced the sponsoring organization, the more type three participants there are.

In Battlebots, for example, the purist builds robots that are not only effective, but satisfy the unstated goal of being highly entertaining. These robots rarely win but they get lots of attention. An example of this robot type is Nightmare. This robot is a crowd favorite but tends to destroy itself in matches. As an example of the second type of builder, Biohazard is the most decorated and successful Battlebot. It is mostly defensive with just enough offense (lifting arm) to flip other robots. Many bouts involving Biohazard only last a few seconds and result in an inverted robot that cannot move. Highly effective, but not very entertaining to the general public.

Sadly, Battlebots has seen it share of the third type of builder. I was heavily involved in the creation of the Battlebots rule set and we struggled between the need to keep people safe and the desire to allow builders as much flexibility as possible. The most challenged rule we had was the rule that allowed walking robots to have a 50% weight advantage over wheeled robots. Many people tried to create a mechanism that looked like it was walking but really was effectively a wheel. In the early days of Battlebots, prospective walking robot designs were sent to me for review. I rejected virtually all of them. Somewhere along the third year, Battlebots stopped asking me to review designs and somebody showed up with a 330 lb. heavyweight that he claimed was a walker but was really a kinematically segmented wheel. With the extra 110 lbs, the vehicle had a spinning weapon with enormous destructive power. He breezed through the heavyweight class (the toughest of all in the competition) until the finals and then had an epic battle with Biohazard (which he won). After that event, Battlebots was forced to change the rules about walkers.

Theses three classes of builders carry over even into pure academic competitions. In talking with other FIRST mentors, I have discovered an almost a universal trait on teams. There are some students (and advisors!) who spend enormous effort trying to find holes in the game rules. There have been robots that show up at a regional competition and discover their borderline legal method of scoring is outlawed on the spot. It makes for some very unhappy students.

That brings us back to the Regolith Excavation Challenge. This year, we all have a pretty good idea what NASA wants. A truly mobile excavator that could collect enough regolith to help construct a lunar base and keep that base supplied for years. Unfortunately, it is not practical for the contest form to mimic a realistic mining field. If they increased the size of the regolith box to 10m by 10m to better reflect more available area, they would need 6 times more regolith and 6 times more time between competition attempts to compact it (90 minutes). So we end up with a 4m by 4m box and rules are put in place to try and produce entries that take the form NASA wants.

Those of you who are active on the forum saw that I challenged many of the rules (and lack of rules) early on. I (like others) was trying to get the more obvious holes closed. I knew that $500,000 is a lot of motivation for somebody to skirt the edges of the rules. Looking at some of the questions asked revealed a lot of the builder’s intent. Things like:
• “If my vehicle is underweight, can I start with regolith in it?”
• “Can I satisfy my traversal requirements by putting my reference point on the end of a rod and having that rod fall down?”
• “I only have to cross the traversal line once, right?”
• “My stationary powered conveyer system that moves regolith to the box is really a navigational aide, and therefore exempt from the traversal rules”
• “I have a two part robot, a scout (RC rar) and a stationary digger. The car is my mobile part.”
It will be interesting to see what designs passed the initial inspection. I am sure that there will be some lively discussion at the event for those designs near the edge.

Despite the paragraph above, I am looking forward to seeing all of the innovative designs. I always learn things at engineering contests and usually find a way to use that information in the future. I am hoping CSEWI has things under control and that we all have a pleasant experience.

another beach day

August 31, 2009 - 11:48 am 1 Comment

It seems like all we do is post about beach testing and beach testing results. Well, I guess that is better than last year when we didn’t have anytime to make posts! Anyways, we are headed back to the beach for another fun-filled day of BFD testing this friday (9/4/2009). We will be meeting up at the shop around 10am, and plan to be at the Dockweiler State Beach around 11am. See a previous post for directions.

On the agenda for this friday this:

  • Save the complete video/sensor data from every run using Aaron’s new tool.
  • Continue to train backup drivers (They almost have a handle on it).
  • Testing with WANEM in a virtual machine as described by the Competition FAQ.
  • Testing with Rob’s filter of the IR sensors (and subsequent GUI updates).
  • Testing with more realistic obstacles.
  • At least one 30minute competition dry-run.

ir sensor processing

August 30, 2009 - 7:02 pm 1 Comment

Rob’s IR sensor filter has been completed and integrated into the vehicle. Lab tests show excellent results, and we are very excited to try it next time we go to the beach. Below is an example of data collected from our last beach test. As you can see there is a lot of noise attributed to both the cutters passing under the sensors and just general noise of the sensors.  Hopefully next time we excavate we will have a very clear idea of how much material is contained in the vehicle’s collection box.

ir_sensor_filter

results from last beach test

August 22, 2009 - 5:33 pm Comments Off

We had another successful testing day yesterday (8/21/2009). There were much less people observing us, so we were able to take it easy. However, we did manage to do some fancy tests. Denny and Aaron did the first couple of runs using WANEM and avoiding obstacles. Rob and Ken also did some runs with the WANEM. All the packet loss problems associated with previous runs with WANEM have been solved.

We added a new capability that made a huge impact to our performance. One thing we have struggled with as a team is how much information and technology to share with the competitors. We like to report our progress and challenges, but not in extreme detail. In this case, the change was simple technologically, but implementing it correctly was the result of extensive vehicle testing. If we were to share it, we would be giving away the fruits of an enormous amount of work.

The auto-cut algorithm was tested and still needs to be tuned. It is difficult to tune this algorithm for both sand and regolith. Our approach needs to be able to adapt on the fly.

The logging of telemetry/video worked very well. Aaron’s new recording program works like a charm and will be used to save data from the perspective of the vehicle for all future tests. You will notice the triplet videos with the 3 cameras in a single frame. Expect more of those. As I am writing this post, Rob is busy crunching numbers from the box sensors to improve the determination of the amount of collected regolith. Hopefully our next run will include his new filter(s). I will try to post some graphs showing vehicle performance metrics in the near future.

beach testing 8-21-2009

August 19, 2009 - 12:07 pm 2 Comments

We are headed to the beach for a day of testing this friday (8/21/2009). We will be meeting up at the shop around 11am, and plan to be at Dockweiler State beach around 12pm. See previous posts for directions.

On the testing agenda is:

  • drive with delay and bandwidth limitations
  • drive with auto-cut algorithm
  • driving with realistic obstacles
  • store video/sensor data for later post processing
  • continue to train backup drivers

new webcam

August 14, 2009 - 11:24 pm Comments Off

We have been testing a different cameras for the vehicle. Some of them just didnt work like we expected/needed. As a result of all this testing, we have taken one of the higher qualities ones and made it the new webcam for the bfd. The quality is now much better and it places a timestamp on the camera. Go check it out!

beach testing results from 8-7-2009

August 12, 2009 - 8:53 pm 1 Comment

Our beach testing results from last friday went well. We had a fairly large crowd consisting mostly of friends and family. Baywatch decided to come over and see what kind of shenanigans we were doing. I asked if C.J. was working today, they said no. Oh well, next time.

The new sensors and algorithm for determining the level of regolith worked so-so. There is still room for improvement, and we will work on a solution to it. The vehicle now does not leak any material. (Which you may have noticed from previous videos). Greg’s ingenious folding teflon cloth worked beautiful. As a result, we were able to dump about 70kg as opossed to about 64.

We did have some problems with WANEM and packet loss, and aborted testing with the delay. For some reason the commands from the controllers only occassionally got through. We will have to investigate. The dropping of packets is curious because we are only sending very little data in that direction and we are using UDP.

Ken and Rob also got a chance to practice driving the vehicle. They are currently the backup drivers incase something catsotropic happens or someone wins the lottery. They picked up the gui interface surprizingly fast, and were doing circuits around the arena with relative ease. Except for that one time that Ken thought the vehicle was a crab and moved it across the arena drawing a large “W”.

With such a large crowd comes Murphy. When Murphy shows up, something is always bound to go wrong.  We had a slight mishap where Denny turned the wheels when they were touching the side of the arena. This unfortunately loosened the wheel, and it eventually failed to follow commands. No big deal, just a quick tightening of a nut and the BFD was ready for more action.

The following is a video showing the perspective of the drivers while operating the vehicle. What might not have been obvious from previous videos, is that the front two cameras are on servo’s and can rotate. You will also notice a point where the cameras seem very wavey. This is when the shaker is turned on to move the collected regolith.

beach testing 8/7/2009

August 3, 2009 - 11:06 am 3 Comments

We are headed to the beach for a day of testing this friday (8/7/2009).  We will be meeting up at the shop around 12pm, and plan to be at Dockweiler State beach around 1pm. To get there, simply take the 105 west all the way till the end. We will be in the payed parking area. However, you can park on the street for free and walk down.

google map:

http://maps.google.com/maps?q=33.930708,-118.434595&sll=33.934014,-118.433883&sspn=0.01264,0.021265&hl=en&gl=us&ie=UTF8&ll=33.925842,-118.408756&spn=0.110959,0.185051&z=13

map of testing location:

beach testing