Comments

• saba March 7, 2013 at 2:04am

  dear all

  i have a ground vehicle (jeep) named TORQUE HORSE. it has the feature of Forward/Reverse, Right/Left. pl see the below link

  http://detail.en.china.cn/provide/1002172928.html

  i have downloaded the latest alpha software on my ardino board. any one could guide me to configure the ardino board for this vehicle. thanks in advance.

  kind regards

  saba
• Gregory Brill March 5, 2013 at 11:20am

  That sabertooth looks great....I could probably engineer around that.  60A ought to be enough.  Would have made my life a lot simpler...wish I'd known about it earlier!   But glad I know about it now, thank you for pointing that out.  Too late on this drone to make this change...went with Curtis/CAN controllers because I thought of this as more of a vehicle than a robot.  

  Well, should be simple enough to parse the signal myself and convert to CAN equivalent...seems like that would likely be easiest to do.
• 

  Anthony White March 5, 2013 at 11:08am

  Gregory - 

  A slight alternative to option 1 would be using a motor controller that can parse the signal to differential. The Sabertooth series do this (check out the datasheet) - although the biggest they have is 60 amps I think.

  http://www.dimensionengineering.com/products/sabertooth2x60

  And I don't know much about CAN messages  in relation to arduinos in general, but modifying the firmware is going to be tricky. There is no easy place to step into the code and make the kinds of modifications you're talking about, I believe.
• Gregory Brill March 5, 2013 at 10:52am

  Looking for advice on best approach for steering via APM 2.5 a very large, skid-steer ground rover (350lbs). 6x6 tank platform, basically.  Meant to carry 500lbs of gear deep into rugged terrain.

  This class of drone can't use a wildthumper or such, because its motors need industrial controllers (Curtis in my case).  This thing pulls 100amps at times. And these controllers listen to CAN messages.

  The options as I understand them, are:

  1) listen to the outputs of the APM with a separate arduino, and translate these outputs into the appropriate CAN messages and mix/steer my vehicle.

  2) modify the APM firmware itself to send CAN messages instead of the servo commands.

  I prefer the simplicity of #2.  But I don't really know how monumental a task that is.  Is the software relatively easily modifiable? Documented?   And I would need SPI pins to control the CAN module, does the APM expose these (I can't determine it because their online docs just say "io pins" and I am unable to open their eagle file for some reason.

  Appreciate any and all thoughts on this, thanks in advance. 

  Best regards,
  Gregory Brill
• Admin

  Thomas J Coyle III March 4, 2013 at 6:38am

  Hi All,

  I have been seeing some comments concerning the 2013 Sparkfun AVC and it looks like they will hold the AVC offsite at Boulder Reservoir on June 8th.

  My understanding is that the Reservoir area has a parking lot, however if they use the parking lot for the rover competition, where will competition entrants parks? On the other hand, will the rover competition be over a dirt course? This might be a lot tougher competition than around the Sparkfun building. A more open area will make for better GPS receiption, but may be tougher on the rovers.

  Regards,

  TCIII
• Admin

  Thomas J Coyle III March 3, 2013 at 10:12am

  Hi All,

  Continued to work on validating the ArduRover2 sonar function:

  1) Tried the Sharp gp2y0a02yk medium range sensor (same one used by JLN) on my APM 1.4/Oilpan HIL. Got the same results as the Sharp 2D12 IR sensor. At infinity the gp2y has an output of ~0.4 volts and the sonar voltage reading in the CLI sonar test is indicated ~1.45 volts just like the Sharp 2D12.

  2) Tried the Sharp gp2y0a02yk medium range sensor (same one used by JLN) on a different APM 1.4/Oilpan.Got pretty much the same results as on my HIL APM1.4/Oilpan. At infinity the gp2y has an output of ~0.4 volts and the sonar voltage reading in the CLI sonar test is indicated ~1.35 volts similar to the Sharp 2D12.

  3) The hyperbolic function does not seem to work as the range in cm did not change with a change in voltage. It remained constant at 700cm. The sonar_function is set to 2, the sonar_constant is set to 14500, and the sonar_offset is set to 0. The sonar_ constant is the one JLN used, but I believe that he was using the raw sonar voltage counts and not the actual voltage in the divisor.

  4) I feel that I have definitely ruled out the Oilpan ADC as having a conversion problem as both Oilpan ADCs gave very similar voltage readings at infinity for the Sharp gp2y0a02yk medium range IR sensor.

  Regards,

  TCIII
• heliboy March 3, 2013 at 5:08am

  All:

  I have been testing a LEO-6P with active antenna for better positioning accuracy. Strange effects you should be aware of : using ublox ucenter software this setup gives me HDOP 0.9 to1.1 indoors !. However when connected to APM2.5, MP shows hdop 2.1-2.2. Now this makes no sense since the HDOP number is actually transmitted via NMEA from the GPS and it is easy to verify since it is just text format. What is MP doing ? And what is APM doing ?

  - Vijay
• Admin

  Thomas J Coyle III March 2, 2013 at 3:48pm

  Hi Tridge,

  Still working on getting the correct range readings from the Sharp 2D12 IR sensor. Noticed one issue that we will need to work on: when the Sharp sensor is pointed at infinity, the output voltage of the sensor should be ~0.4 volts. For some reason the ADC is reading 1.45 volts. Very strange. I am attaching the Sharp 2D12 data sheet for you to look at.

  Also, the conversion value of 14500, that JLN was using in the Sharp library for the gp2y0a02yk, is based on the "raw voltage" which I believe is in counts.

  Regards,

  TCIII

  Sharp_GP2D12_IR_Sensor_Specs.pdf 

  https://storage.ning.com/topology/rest/1.0/file/get/3692639692?profile=original
• Admin

  Thomas J Coyle III March 2, 2013 at 2:05pm

  @Tridge,

  I will give it a shot.

  Regards,

  TCIII
• Developer

  Andrew Tridgell March 2, 2013 at 1:13pm

  Hi Tom,

  After we finished up our dev call on Friday I did a bunch more work on the sonar code, and added support for both inverted and hyperbolic sonars.

  The key is a new parameter SONAR_FUNCTION, which has this explanation in the code:

      // @Param: FUNCTION
      // @DisplayName: Sonar function
      // @Description: Control over what function is used to calculate distance. For a linear function, the distance is (voltage-offset)*scaling. For a inverted function the distance is ((5.0-voltage)-offset)*scaling. For a hyperbolic function the distance is scaling/(voltage-offset). The functions return the distance in meters.
      // @Values: 0:Linear,1:Inverted,2:Hyperbolic
      AP_GROUPINFO("FUNCTION",  3, AP_RangeFinder_analog, _function, 0),
  
  

  I think that should work for your Sharp 2D12 sensor, but please test it! If needed we can add more SONAR_FUNCTION values for different mappings between voltage and range.

  Cheers, Tridge