Design
Original Concept
The images above are the original design concept with both the linear actuator and the reaction wheel mounted on the bike. Through extensive testing it was determined that the reaction wheel was not going to be mounted on the bike and was going to be demonstrated as a separate system.
Steering System
Reaction wheel
The reaction wheel was originally intended to be mounted directly to the bicycle and work in conjunction with the steering control to maintain the balance of the bicycle. After determining that the system could be modeled as an inverted pendulum a crude model one was built to determine proof of concept.
Once the proof of concept was shown scalability had to be determined. Was is possible to build a reaction wheel that could effectively control the bicycle we had using a limited budget? To try to determine the size of motor, reaction wheel and electrical forces that would be required a larger model was built.
.jpg)
After significant efforts with trying to size motor drivers, motors, power supplies and reaction wheels, it was finally determined that building a reaction wheel that could control something on the scale of the bicycle was out of the scope of the project with the available resources. There were two main reasons the larger reaction wheel did not work. First it was difficult to find a motor driver capable of handling the current required. At start up the motor was drawing up to 80 amps of current to get the reaction wheel moving at the torques required. We were able to find motor drivers capable of handing close to this current, but the ones that did had a long delay (0.5seconds) in direction switching. Obviously this was far too long of a pause for the system to work.
Some of the motor drivers used and tested. The motor drivers' control schemes varied, so each driver had to be researched and effective code written for it.
The other main problem encountered was the MPU6050 accelerometer. When it functioned it grave great feedback, but unfortunately we had difficulty with it. It was undetermined what exactly the issue was with the accelerometer, but under high currents it caused the Arduino controller to crash.
MPU6050
After deciding that the wheel was not going to be mounted on the bicycle it was decided to build another model that was cleaned up and made to run very effectively. By choosing to only build a model things could be scaled down to appropriate levels and a potentiometer could be used for positional feedback of the arm instead of the accelerometer
Cleaned up model for demonstration.