Optimisation and Control of an Overground Paediatric Robotic Gait Trainer

Abstract

Robotic gait training has become increasingly popular and has shown to be able to improve the gait patterns of children with cerebral palsy. This research details the continued development of a novel paediatric robotic overground gait trainer (PRO-GaiT), a device based on a non-anthropomorphic linkage mechanism and frame. The device consists of a two degree of freedom five-bar linkage with a footplate end-effector driven by two motors. The design allows for various gait patterns to be accurately generated while permitting the user to move freely relative to the device. The original design of the PRO-GaiT was reconsidered and the device geometry determined via a new multiobjective optimisation problem. The device was optimised with respect to kinematic criteria as well as the actuator torques required to move a user through a normal gait pattern. Different multiobjective optimisation configurations were solved to evaluate the total problem space. The results of the optimisation configurations were evaluated and a final solution chosen based on the design criteria. The new design was found to better meet design specifications when compared to the original design and proved the feasibility of using five-bar linkage mechanisms in gait rehabilitation devices. The new design of the device was validated using simulation and a prototype manufactured. Different control algorithms were developed for the PRO-GaiT to move a user through a normal gait pattern and to allow a user to move unhindered in the device. These algorithms were trajectory tracking, impedance control, zero-impedance control, and force-field impedance control. The algorithms were validated using simulations and device testing. Results indicated the device is capable of guiding a user along a reference gait trajectory while considering user intent. The development of the automated control of motorised wheels gave overground capabilities to the device. Walking trials were completed with six participants. Participant feedback showed the device is easy to walk in and it was found that over time participants adapted to the device. Overall, the results of the simulations, device testing, and walking trials demonstrated that the PRO-GaiT meets the design criteria and can be adopted for clinical use in the future.