Date of Award
Master of Science in Electrical Engineering
Washkewicz College of Engineering
Two regenerative motor drives, a voltage source converter and a bidirectional buck/boost converter, are studied for energy regeneration and joint trajectory tracking. The motor drives are applied to two different robotic systems—a PUMA560 robotic arm and a hip testing robot / prosthesis system. An artificial neural network controller is implemented with the two motor drives and provides joint trajectory tracking with an RMS error of 0.03 rad. The control signals produced by the artificial neural network contain a large amount of high frequency content which prevents practical implementation. A robust passivity-based motion controller is modified to include information about the motor drives to overcome the limitations of the artificial neural network controller. The modified robust passivity-based controller outperforms the artificial neural network controller by maintaining a 3 V RMS error between the voltage generated by the converter and the desired voltage while maintaining comparable trajectory tracking. The high frequency content of the robust passivity-based controller contains less high frequency content than the artificial neural network controller. The modified robust passivity-based controller is implemented inside the semiactive virtual control energy regeneration framework to demonstrate energy regeneration with one of the motor drives. The motor drive implemented with the energy regeneration framework shows that energy can be regenerated while using the bidirectional buck/boost converter.
Barto, Taylor, "Design and Control of Electronic Motor Drives for Regenerative Robotics" (2017). ETD Archive. 958.