DARPA Challenge
AIman
Posts: 531
Andy (Parallax)·I apologize if there was something missing in the email I sent you previously.
For everyone,·hopefully this is helpful·for future projects.
I thought it would be helpful·for people to know a little about some things that work for programming robotic vehicles. I was invovled in the DARPA Challenge· but ended up with drawing because my wife is expecting and my priority order is family before fun.
Here are the basics of how things were set up.
There were originally the Propeller, the SX and a host linked together. The SX was to handle a video feed and the Propeller was to handle other functions. The SX was discarded and replaced with a Propeller because it removed a computer language. Then the propeller was discarded and the video feed sent straight into the host because of being able to remove an additional chip. It was seriously considered that running all sensors into the host directly via USB would be a good thing, however with the resources available from Parallax that was ruled out.
First off the vehical was full sized for several reasons - it gave a much smoother ride, there was plenty of horse power, there was lots of room under the hood and around the car to place sensors and run the needed cables. It was an automatic and was easy find places to add extra power supplies if needed. No additional power supplies were needed because the alternator supplied more then enough to allow the car to function properly and all the tech equipment to function properly. Each function of the vehical (sp?) is broken out by using 2 propellers.
Propellers were set up to function based on cogs NOT on pins.
Propeller 1 -
Cog 1 - Gas
Cog 2 - Brake
Cog 3 - Steering
Cog 4 - Gear Shift
Cog·5 - Sonar - used·10 pins
Cog·6 - IR - used·10 pins
Cog·7 - Dedicated to host computer
Note - IR and Sonar have been put into and taken out of the plans several times because of using a laser. Some sets of plans used both some used only one. In the end they left in as·a back up to the laser.
Propeller 2
Cog 1 - GPS
Cog 2 - Compass
Cog 3 - Required lights and siren
Cog 4 - Hall Sensors
Cog·5 - Required saftey over ride
Cog·6 - Dedicated to host
Hall sensors were used on all 4 tires to measure RPM not only to cross check speed and distance against the GPS but also to make sure tires didn't spin. Example - front tire doesn't move back tire spins, computer cuts power to drive tires.
Video camera - feeds directly to host. It was a standard camera with a RCA hook up.
Laser feeds directly to host.
Propeller chips were hooked in via USB ports.
Main host ran common LISP and sent out commands to cogs 1-4 of Propeller 1
GPS was set to carry math to the farthest decimal upto 1000 places, the GPS also recalibrated every time the car came to a stop or was parked. Stop time was based on 3 seconds for purposes of recalibration and being sure that all traffic laws were obeyed. There was also a manual upload every day for altimeter pressure from the nearest airfield giving more accuracy again.
The compass was only to serve as a cross check for the GPS.
Note - between the hall sensors and the compass RPM and distances could be figured out fairly accuratly and the directional heading could be figured fairly accurately.
Sonar and IR were to be mounted inside the front wheels to make sure tires were steering correctly. Again as a back up feature so that if they didn't turn the brakes could be applied. The reason IR and sonar were used together is because they function well as a team.
Lasers were actually going to be home made. SICK has some very good products which were seriously being considered, however the products located for laser diodes have more then adaqute distance and were much cheaper.·In the end SICK won out because the lasers were of 180 degree 'vision' mounted with one on the roof forward facing, one on the roof rear facing,·One on the drive side·facing up and one on the passanger side facing up. Lasers·do have a draw back of not being able to see within a few feet of the vehicle which is aonther reason·sonar and IR were used.
There were two hard drives giving a total of 120 gigs of memory where all events were recorded - the video was recorded via standard CD in the camera running time code so that events could be linked precisouly.
To cushion the computers they were placed ON the back seat and strapped in via seat belt which gave a smooth enough ride no speacial absorption equipment was needed.
To recap:
Gas, brake, steering, gear shift, GPS, sonar/IR, siren and Hall sensors·were all handled by Propellers with the Video and laser being feed straight into the host.
Sonar and IR were set up side by side around the car - one set for each side and corner and one set for the inside of each front wheel.
Hall sensors were on each wheel
Servos were used to handle steering and selonids used to handle brake, gas and gear shift.
Normal lights on the car came from the factor with an auto on/off which was left in the on setting.
Video and laser were fed straight into the host.
·
For everyone,·hopefully this is helpful·for future projects.
I thought it would be helpful·for people to know a little about some things that work for programming robotic vehicles. I was invovled in the DARPA Challenge· but ended up with drawing because my wife is expecting and my priority order is family before fun.
Here are the basics of how things were set up.
There were originally the Propeller, the SX and a host linked together. The SX was to handle a video feed and the Propeller was to handle other functions. The SX was discarded and replaced with a Propeller because it removed a computer language. Then the propeller was discarded and the video feed sent straight into the host because of being able to remove an additional chip. It was seriously considered that running all sensors into the host directly via USB would be a good thing, however with the resources available from Parallax that was ruled out.
First off the vehical was full sized for several reasons - it gave a much smoother ride, there was plenty of horse power, there was lots of room under the hood and around the car to place sensors and run the needed cables. It was an automatic and was easy find places to add extra power supplies if needed. No additional power supplies were needed because the alternator supplied more then enough to allow the car to function properly and all the tech equipment to function properly. Each function of the vehical (sp?) is broken out by using 2 propellers.
Propellers were set up to function based on cogs NOT on pins.
Propeller 1 -
Cog 1 - Gas
Cog 2 - Brake
Cog 3 - Steering
Cog 4 - Gear Shift
Cog·5 - Sonar - used·10 pins
Cog·6 - IR - used·10 pins
Cog·7 - Dedicated to host computer
Note - IR and Sonar have been put into and taken out of the plans several times because of using a laser. Some sets of plans used both some used only one. In the end they left in as·a back up to the laser.
Propeller 2
Cog 1 - GPS
Cog 2 - Compass
Cog 3 - Required lights and siren
Cog 4 - Hall Sensors
Cog·5 - Required saftey over ride
Cog·6 - Dedicated to host
Hall sensors were used on all 4 tires to measure RPM not only to cross check speed and distance against the GPS but also to make sure tires didn't spin. Example - front tire doesn't move back tire spins, computer cuts power to drive tires.
Video camera - feeds directly to host. It was a standard camera with a RCA hook up.
Laser feeds directly to host.
Propeller chips were hooked in via USB ports.
Main host ran common LISP and sent out commands to cogs 1-4 of Propeller 1
GPS was set to carry math to the farthest decimal upto 1000 places, the GPS also recalibrated every time the car came to a stop or was parked. Stop time was based on 3 seconds for purposes of recalibration and being sure that all traffic laws were obeyed. There was also a manual upload every day for altimeter pressure from the nearest airfield giving more accuracy again.
The compass was only to serve as a cross check for the GPS.
Note - between the hall sensors and the compass RPM and distances could be figured out fairly accuratly and the directional heading could be figured fairly accurately.
Sonar and IR were to be mounted inside the front wheels to make sure tires were steering correctly. Again as a back up feature so that if they didn't turn the brakes could be applied. The reason IR and sonar were used together is because they function well as a team.
Lasers were actually going to be home made. SICK has some very good products which were seriously being considered, however the products located for laser diodes have more then adaqute distance and were much cheaper.·In the end SICK won out because the lasers were of 180 degree 'vision' mounted with one on the roof forward facing, one on the roof rear facing,·One on the drive side·facing up and one on the passanger side facing up. Lasers·do have a draw back of not being able to see within a few feet of the vehicle which is aonther reason·sonar and IR were used.
There were two hard drives giving a total of 120 gigs of memory where all events were recorded - the video was recorded via standard CD in the camera running time code so that events could be linked precisouly.
To cushion the computers they were placed ON the back seat and strapped in via seat belt which gave a smooth enough ride no speacial absorption equipment was needed.
To recap:
Gas, brake, steering, gear shift, GPS, sonar/IR, siren and Hall sensors·were all handled by Propellers with the Video and laser being feed straight into the host.
Sonar and IR were set up side by side around the car - one set for each side and corner and one set for the inside of each front wheel.
Hall sensors were on each wheel
Servos were used to handle steering and selonids used to handle brake, gas and gear shift.
Normal lights on the car came from the factor with an auto on/off which was left in the on setting.
Video and laser were fed straight into the host.
·