Saturday, January 15, 2011

New Week, new to do list

Hi Folks,

The progress reports look good. Keep the video coming. I am happy to see that the platform is driving around.

For this next week, I see our overall goal as being able to drive the platform around while logging the GPS values and having the platform update its position on the map in real time. Ideally we would see a video by the end of the week of the platform driving through the field with an overlay of the gps track. It occurs to me that we will probably want to shift to google earth for the real time mapping. This should make the real time gps tracking or replay a cinch. (I remember we did this on our near earth satellite project a few years ago.)

Things to do:
  • Secure the netbook to the platform so it will survive. (Maybe bend up some brackets out of vex parts. Feel free to drill holes in the top plastic plate to bolt the backets on. Or zipties?)
Mitch:
  • Get something working with the wheel encoders on the robot even if it is low res. Choose robustness over resolution. I am more concerned about hardware now then software. Feel free to take the platform home if you need to (I think this is our biggest hardware problem now.)
  • Modify the sample arduino software to send out serial messages with the encoder values.
  • Do Calibration of your encoders with the real platform.
  • Work with Brian on getting the compass mounted on the platform. Andy needs it for his mapping work this week. The compass needs to be mounted at least 6 inches above plastic platform and on a NON metallic surface.
  • Take a look at the Maxbotics EZ1 and Sharp IR sensors. We will need to interface these to a different arduino next week for obstacle avoidance.
Andy:
  • We need a map. I like your progress so far. Keep working on implementing the map code with the gps and get the image to move around according to the gps data.
  • Explore google earth as a platform for visualizing the gps data. I don't think this can replace what you are doing, but should allow for nice presentations.
  • Work with Mitch to agree on a messaging protocol from the encoders to the map software. I would suggest that the encoders send out a serial stream at 57600 8N1 which reports the current encoder value of a wheel on each side in a comma delimited format and terminated with a carriage return. Maybe a datarate of 10Hz is a good goal since it matches the GPS.
  • We want to get autonomous navigation started. Using your map, allow the user to specify a target point on the map (perhaps by clicking or entering GPS values.) Calculate how far it is from the current position to the target position and how much the bot needs to turn. Generate a set of turn and speed commands to get the robot pointing at the target. Do this closed loop. (IE turn a little, measure from the compass, decide how to turn, repeat)
  • Test your autonomous navigation on the real platform over short distances using only the compass to make sure you can get the heading to work.
  • Start thinking about obstacle avoidance and how that will figure in to the map.
Brian:
  • Tackle the WAAS and other issues to improve our GPS signal. I dug out an improved GPS ,http://www.sparkfun.com/products/9133 but have to build a USB to TTL interface and test it. Ok, that is done and I will have it for you on Tuesday. Look at http://profmason.com/?p=1468 for details.
  • Run the same tests on the new GPS that you ran on the previous and make any necessary changes to use the new output stream.
  • Read the Venus documentation and the comments on the sparkfun site. Currently it is setup with WAAS enabled, pedestrian mode, 10Hz, 57600 8N1. There may be some other settings to play with.
  • Get the compass on the platform. Andy needs it for his mapping work this week. The compass needs to be mounted at least 6 inches above plastic platform and on a NON metallic surface. Talk to John about how he communicated with his compass in I2C mode and if this is worth doing. (He has the same one we do.)
  • After this stuff is done, focus on the IMU. Make a start at integrating the accelerometers for position.
I searched through my parts bin and couldn't find a sonar sensor, so I have ordered a maxbotics EZ1 (These are great) This will give us front obstacle avoidance with a range of about 5 meters. I also have a pair of SHARP 2YOA02 proximity sensors. (Very narrow beam out to 1.5 meters) I would like this to be the last week we are remote controlling the platform and start moving to autonomous navigation.

No comments:

Post a Comment