Optimal Low- Power Control Strategies for an Autonomous Piezoelectric Micro-Robot.

Abstract

This dissertation presents optimal control strategies for autonomous operation of a piezoelectric micro-robot with a strict energy budget and payload constraints. Due to the capacitive nature of piezoelectric actuators, traditional analog amplifier based controllers and pulse width modulation based controllers use excessive amounts of energy compared to the actual physical work done by the actuators. Hence, an optimal on-off controller was developed which is the building block for the control strategies proposed in this dissertation. Strategies are presented for minimizing energy consumption during both the charge and discharge phases of actuator and when using feedback.

First, an optimal on-off controller and a partial charge recovery controller are presented. These controllers are developed for driving simple movements of actuators in the robot from a given initial position to a prescribed final position in a given time with minimum number charging of the actuator from external power source. Charge recovery controller recovers drained energy at the actuator discharge phase as well. Conceptually these optimal movements of the actuators can be combined to form an efficient gait sequence for the robot. Compared to traditional controllers, the proposed on-off controller consumes only a fraction of energy and the charge recovery controller uses even lesser; in certain cases charge recovery controller consumption was observed to be one third of on-off controller consumption for the same motion.

Second, a three step strategy is proposed for adding an energy consuming position sensor to the system in such a way that the sensor is turned on sparingly in order to achieve a certain level of error performance. Under certain assumptions, which may be verified relatively simply, the proposed strategy can be used to find optimal time points to turn on the sensor and update input sequences.

Finally, a hybrid energy system architecture containing a small solar cell and a micro-battery which can be used on-board is proposed. A study on selecting optimization parameters based on gait sequences for choosing one of the power sources is presented. Together this is an attempt to form a benchmark for autonomous micro-robotics using piezoelectric actuators from an energy consumption point of view.