I started a project a couple months ago for a sensors class I'm taking this quarter. It is a two wheeled balancing robot that uses a SX Micro Controller and a IMU Sensor for sensing its angle to gravity. I am still working on the getting power to the MC and sensor but the entire mechanical platform is built.
I was wondering if anyone else had used the SX for this purpose? Does anyone have any tips on this type of robot? Do you think it is possible to program a Kalman Filter with assembly language?
A while ago I bought a SX kit from Parallax but I never used it - they are the main distributors for Scenix. Parallax have a free compiler for the SX although unfortunately it is BASIC. I think there may be some free Ebooks at their site.
There may also be a forum there that could help.
Also there is another site that sells C compilers. One of these can generate code for BOTH PICmicro and SX. It is reasonably priced.
Is it possible to replace the SX with a PIC? That would be a major advantage because of the huge amount of resourcews available to you, although from what I understand the SX is a PIC clone.
Although I am not familiar with the Kalman filter, there would be algorithms available in C, and this could be ported.
One of my teachers at Wollongong TAFE technical college built an inverted pendulum with Lego, but it was controlled with a PC I think. It uses fuzzy logic or PID. I have C the code for it somewhere, and can mail it to you.
geek snipped-for-privacy@hotmail.com
His solution is elegant, and was demonstrated to the neighbouring University of Wollongong who were impressed.
You might be surprised how far a simple control loop gets you (even before Kalman). Apply negative feedback with just enough overshoot to recenter the wheel base...
Depending on how many states you model, the Kalman filter boils down to a handful of additions and multiplications. You might even be able to code this in basic; you just need to get the feedback loop going in the tens of Hertz (an order of magnitude faster than the natural frequency of the pendulum).
Have fun, Daniel
P.S. It's a good idea to have fall bars extended in front of and behind the robot. For when the battery gives out.
I would switch to a PIC, but I dont think I have the time to start over with the controls design. I would still have to implement a PID loop for the actual balancing. The Kalman Filter is to combine the Gyro and Accelerometer Outputs into a good clean, zero drift signal that the PID loop can use. Thought about the fall bars, I hope to add a self charge ability to the robot later on so it never runs out of juice.
I think my teacher just used pots connected to the Lego gear train or wheels... and it works *really* well, nothing fancy, (apart from the fuzzy logic or PID, whatever he used)like I said - elegant.
Yeah I am pretty sure it is tethered to a an old PC and programmed in Borland C++ using mostly C code.
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.