Optimization of linearly approximated musculoskeletal robots targeting elliptical trajectory generation based on topological derivative

Abstract

Topology optimization is a method for determining the optimal structure for a particular objective based on physical equations and optimization mathematics. It has been applied to a wide variety of physical and engineering problems, and its application to soft robot design has attracted much attention in recent years. We specifically focus on the use of topology optimization in the design of musculoskeletal robots. In this study, we first propose a simple representation model of musculoskeletal structures suitable for topology optimization. This model represents an artificial muscle that behaves differently in time due to multiple expanding domains. Then, we present topological derivatives, which are design sensitivities for optimizing the robot’s structure. An optimization method using these topological derivatives is applied to the design of a mechanism that creates elliptical trajectories, and a reasonable optimal solution is obtained. This provides a promising direction for future research in the field of soft robotics.