MS student Dennis Butts with Advisor A. Banerjee

Inspired by the flexible movements of animals, the robotic spine aims to solve some of the issues with rigidity in robots. This project uses a distributed electromechanical actuation system that can be stacked to form a robotic spine. This electromechanical actuator exchanges torque for total angular flexibility – the total rotational range each joint can achieve – which contrasts with the torque and speed relationship of a motor. When the rotational angle of the distributed actuation system is small, the torque density of the actuator is much larger than that of brushless DC motors. Because animal spines are usually limited in rotational capability, the actuation system is well-suited for robotic spine applications. Additionally, the current through a motor is often limited to avoid nonlinearities between torque and current. However, limiting the electromechanical actuation system’s current is not necessary since the relationship between current and torque is already nonlinear.

My research focuses on the successful implementation of an agile snake robot using the proposed distributed electromechanical actuation system. There are a few challenges with implementing this robotic system. Any new actuator setup requires careful consideration of the material and mass properties to optimize mechanical efficiency in locomotion. Additionally, the actuator’s highly nonlinear torque-current relationship makes traditional control approaches difficult. The small airgap further complicates the control since the controller needs to be extremely precise. My research attempts to integrate modeling, hardware, and control design to successfully implement an agile snake robot using the distributed electromechanical actuation system.

Electromechanical actuation system acting as a snake robot.