Design, Analysis, and Multicriteria Optimization of an Overground Pediatric Robotic Gait Trainer

Abstract

Robotic gait training has been gathering much attention and has demonstrated potential to improve walking patterns for children with cerebral palsy. This paper proposes a novel pediatric robotic overground gait trainer (Pro-GaiT) based on a nonanthropomorphic linkage design and walking frame. The robot configuration is a 2 degree-of-freedom 5 bar linkage “end effector” and considers the human as part of the kinematic chain. This permits the user to move relative to the robot, inherently handling joint misalignments (unlike exoskeletons), to simplify clinical application while still allowing various gait patterns to be accurately achieved. To successfully achieve this complex motion, the kinematic design is critically analyzed to ensure that desired motion and forces can be imparted on the user. A thorough design analysis is conducted, and a genetic algorithm is used to optimize the robot configuration in terms of workspace, singularities, and physical size of the robot. The device was experimentally validated with four healthy subjects, as well as with clinical data from three children with CP that walk with crouch gait with varying human-robot misalignment and joint spasticity. The results demonstrate that the Pro-GaiT has achieved the desired clinical specifications.