Another Sparkfun autopilot--this one with a full IMU!

[Chris Anderson]

It's raining autopilots! Sparkfun has just released a new and much improved version of Bill Premerlani's original IMU-based autopilot board. The previous one was just a 5 DOF board, but this one has a full six degrees of measurement, with three gyros and a three-axis accelerometer. Also has a faster CPU and other goodies (see below).

The code is open source and available in C (a big improvement over the assembly code of the first version).

We'll be hosting the home page for the autopilot board here at DIY Drones, and you can now see it in the tabs above. I've set up a discussion forum here.

If you want to move from a basic entry-level autopilot to a full IMU-based one and stay open source, this is a great way to go. IMUs are a lot more expensive than thermopile-based units like ArduPilot, but if you want full control they're worth it. This one is $299.

Features:

* Connection for a 20-Channel EM-406A SiRF III GPS Receiver (not included)
* PIC dsPIC30F4011 Controller (with onboard 3.3V and 5V glue logic)
* dsPIC runs at 120MHz with 16MHz resonator and PLL
* MMA7260 three axis accelerometer
* 3 ADXRS401 gyros
* 4 Input, 3 output PWM points
* 6-Wire ICSP debug header
* 2 Separate colored status LEDs
* 3 General purpose switches
* On board 3.3V and 5V regulators (150mA max)
* 10m Positional Accuracy / 5m with WAAS
* GPS Outputs NMEA 0183 and SiRF binary protocol
* Spare USART connection for debugging
* 4 Spare digital I/O pins for debugging

[ren]

very nice indeed...
I'm fully aware of the difference in thermopile and gyro+accelerometer combo setups, how they work etc, but what are the intrinsic advantages/disadvantages of a them in an auto-pilot application? (keeping price in perspective)
Capabilities that come to mind that might benefit more from one over the other are controlled rate of turn, bank angle, rate of climb etc
Happy to be pointed to a link/blog/report

[Chris Anderson]

Lots of discussion on that in this thread regarding Attopilot. Basically, there are a few conditions in which thermopiles don't work well (snow, rain). Also, IMUs can give you yaw. (Although you can add a cheap magnetometer to a thermopile unit and get that, too)

[Geetesh Dubey]

A thermopile can never give u a reliable tilt or angle readings for pitch or roll as the temp differences between the sky and earth changes , where the thermopile only looks for the absolute difference and the controller tries to minimize that abolute temp difference...on the other hand an IMU supplemented with an efficient algorithm like that of a KALMAN filter can give u fairly v good idea of exact tilt or angles in terms of ROLL YAW and PITCH.
Note: a kalman filter can give good results for ROLL n PITCH if only IMU is used but if accompanied with something such as a DIGICOMPASS(MAGNETOMETER) then we can get all three Roll Pitch & YAW..