No abstract

Gait training via a wearable device in children with cerebral palsy (CP) offers the potential to increase therapy dosage and intensity compared to current approaches. Here, we report the design and characterization of a pediatric knee exoskeleton (P.REX) with a microcontroller based multi-layered closed loop control system to provide individualized control capability. Exoskeleton performance was evaluated through benchtop and human subject testing. Step response tests show the averaged 90% rise was 26 ± 0.2 ms for 5 Nm, 22 ± 0.2 ms for 10 Nm, 32 ± 0.4 ms for 15 Nm. Torque bandwidth of P.REX was 12 Hz and output impedance was less than 1.8 Nm with control on (Zero mode). Three different control strategies can be deployed to apply assistance to knee extension: state-based assistance, impedance-based trajectory tracking, and real-time adaptive control. One participant with typical development (TD) and one participant with crouch gait from CP were recruited to evaluate P.REX in overground walking tests. Data from the participant with TD were used to validate control system performance. Kinematic and kinetic data were collected by motion capture and compared to exoskeleton on-board sensors to evaluate control system performance with results demonstrating that the control system functioned as intended. The data from the participant with CP are part of a larger ongoing study. Results for this participant compare walking with P.REX in two control modes: a state-based approach that provided constant knee extension assistance during early stance, mid-stance and late swing (Est+Mst+Lsw mode) and an Adaptive mode providing knee extension assistance proportional to estimated knee moment during stance. Both were well tolerated and significantly improved knee extension compared to walking without extension assistance (Zero mode). There was less reduction in gait speed during use of the adaptive controller, suggesting that it may be more intuitive than state-based constant assistance for this individual. Future work will investigate the effects of exoskeleton assistance during overground gait training in children with neurological disorders and will aim to identify the optimal individualized control strategy for exoskeleton prescription.

Introduction Wearable robotic exoskeletons offer the potential to move gait training from the clinic to the community thereby providing greater therapy dosage in more naturalistic settings. To capitalize on this potential, intuitive and robust interfaces are necessary between robotic devices and end users. Such interfaces hold great promise for research if they are also designed to record data from the robot during its use. Methods We present the design and validation of an open source graphical user interface (GUI) for wireless operation of and real-time data logging from a pediatric robotic exoskeleton. The GUI was designed for trained users such as an engineer or clinician. A simplified mobile application is also provided to enable exoskeleton operation by an end-user or their caretaker. GUI function was validated during simulated walking with the exoskeleton using a motion capture system. Results Our results demonstrate the ability of the GUI to wirelessly operate and save data from exoskeleton sensors with high fidelity comparable to motion capture. Conclusion The GUI code, available in a public repository with a detailed description and step-by-step tutorial, is configurable to interact with any robotic device operated by a microcontroller and therefore represents a potentially powerful tool for deployment and evaluation of community based robotics.