Reconfigurable Mechanical System Design for Tracking an Ankle Trajectory Using an Evolutionary Optimization Algorithm

Abstract

This paper presents a procedure for designing a mechanical training and rehabilitation gait system centered on ankle trajectories for children with cerebral palsy or any other psycho-motor limitation. With the aim of developing a reconfigurable device adjustable to the anthropometric characteristics of the user, an analytic model for the generation of ankle trajectories was elaborated, taking as a base the experimental data reported in literature. A dimensional synthesis for a mechanism to follow the drop type trajectories was carried out as a constrained numerical optimization problem that was solved with both the mathematical programming method and an evolutive algorithm, sequential quadratic programming (SQP) and differential evolution (DE), respectively. The comparative analysis of the results from both methods for this case study shows that DE outperforms SQP because of the limited feasibility space derived from the problem constrains and boundaries. The best result from DE was simulated with a computer-aided design package using a real size model for manufacturing.

Keywords

• Trajectory tracking
• optimization
• four-bar mechanism
• differential evolution