SUMMARY

As robots are applied to a greater number of tasks, roboticists are increasingly drawing inspiration from biology. However, most robots still use traditional servomotors to provide actuation, which has little in common with biological actuation. Biological systems produce motion and force using a process known as recruitment, whereby groups of motor units are selectively activated at various instants in time. Units are joined together by flexible cytoskeletal tissue, producing an overall aggregate effect. Actuators that possess these principles inherit certain benefits present in biological motion systems, such as fault tolerance, zero backlash, and reduced risk of damage to the surrounding environment. Piezoelectric cellular actuators are a novel technology that incorporate these biological attributes, namely, a cellular structure and inherent flexibility. This thesis will mathematically characterize these effects. Flexibility arises from compliant mechanisms that amplify the displacement of the piezoelectric material. This research will propose models that can predict the output performance in terms of the displacement and force limits across a wide range of geometries. These models also give rise to a systematic design approach for compliant mechanisms. The cellular structure and operation by recruitment can be mathematically described as a quantizer applied to the control input, complicating the system dynamics. This research will present a control-theoretic treatment for actuators with a cellular structure and develop controllers that can produce smooth motion despite quantization and flexibility effects. Contributions are expected to the areas of compliant mechanisms and control of quantized systems. The techniques and methods developed in the thesis will be demonstrated on a biologically inspired camera positioning mechanism.