Evolution of a jump in an articulated leg with series-elastic actuation

The remarkable ability of humans and animals to perform dynamic maneuvers, such as a jump, is largely attributed to series-elastic elements in skeletal muscle. Both the degree of elasticity and the coordination of muscular contractions have been shown to impact jump performance. The objective of this paper is to use a genetic algorithm (GA) to optimize the control and actuator parameters of a series-elastic actuator (SEA), which is functionally analogous to skeletal muscle, in an articulated leg to produce the highest jump. Similar to skeletal muscle, the control and stiffness of the SEA is found by the GA to affect jump performance, yielding solutions with biological properties. In particular, the jumps evolved by the GA made use of the stretch and shortening cycle of the series-elasticity, which is commonly seen in nature to increase the force of an explosive movement. The model studied in this paper is of a prototype leg with series-elastic actuation. A detailed leg and actuator model was developed to include the important electrical and mechanical properties of the DC motors as well as the characteristics of the motor amplifiers.