Conclusion on end course project

Through 10 sessions we have been working on making an autonomous car balance on a tube, while a remotely controlled car inside the tube is able to dictate the direction the tube is moving. We were successful at obtaining this goal. However the car balancing on top of the tube is susceptible to gyroscopic drift, despite our best attempts to counteract it. This limits the amount of time the balancing car can stay on top of the tube to a few minutes, at our best runs. We have done extensive work to reduce the drift, and have produced a complementary filter which seems very promising to counteract the gyro drift, but we lack the time to produce a proper implementation of said filter into our balancing car code. We are hopeful, that if pursued, we would be able to enhance the performance of the balancing car by logging and analyzing the data from the tube balancing car, similar to what was done in session 10.

Future work:

• Tests and tweaks on the implemented version of the complementary.
• // Alternative do further research on the EMAOFFSET from session 9.
• Fine-tuning of PID parameters, maybe through Ziegler Nichols method.
• Make more tests investigating how different power values, using setPower(), impacts the gyro sensor readings.
• Mount caterpillar tracks on the car inside the tube, this should make the car stronger, and it’s movement smoother.
• Experiment with different tubes. Heavier tubes mainly. A transparent tube would be desirable for aesthetics as well.  

All code from the project can be found here:

• Code: github.com/banditlev/TubeBalancerProject.git