Control of thin McKibben muscles in an antagonistic pair configuration

Abstract

This paper proposes an antagonistic proportional integral (API) control strategy for thin McKibben muscles (TMM) in an antagonistic pair configuration. The study intends to explore a model independent control strategy for TMMs in an antagonistic pair assembly on a TMM platform which mimics the human finger at metacarpophalangeal (MCP) joint. Set-point tracking is achieved by eliminating dead zones due to muscle–tendon slack and using a mode selector for switching of flexion and extension muscles to provide smooth tracking. To test the effectiveness of the proposed method, experiments are carried out on a TMM platform and injected with step and sinusoidal input signals. Experiments conducted showed the TMM platform able to produce high accuracy and fast rise time by implementing the API control scheme. The control method can be very useful in other TMM applications requiring antagonistic muscle actuation.