Ganesh Roy, Dinesh Bhatia, and Subhasis Bhaumik
Biomechatronics, lower-limb exoskeleton, gait training, control algorithm, BCI∗ Department of Aerospace Engineering & Applied Mechanics,IIEST, Shibpur, West Bengal 711103, India; e-mail: [email protected]∗∗ Department of Instrumentation Engineering, Central Instituteof Technology Kokrajhar, Assam 783370, India∗∗∗ Department of Biomedical Engineering, North Eastern HillUniversity (NEHU), Shillong, Meghalaya, India; e-mail:bhatiadinesh@rediffmail.com∗∗∗∗ IIEST, Shibpur, India; e-ma
In recent years, the design of an assistive robot for mankind has become an important area to robotics and biomechatronics researchers. The design of such a device involves many issues, including gait biomechanics, mechanical architecture, control system design, selection of actuator, sensor details, power storage, etc. The present work dealt with all the issues carefully to develop a rehabilitative trajectory-controlled lower-limb exoskeleton (TCLLE) prototype. The TCLLE prototype has been developed and tested with three trajectory control algorithms. The control circuit is built around a microcontroller and a master controller environment. The instantaneous inverse kinematic model has been developed for the single and double support phases of the exoskeleton to determine the appropriate joint space configuration for the gait trajectory. The mean steady-state errors obtained between reference and joint space trajectories are 0.0331◦, −0.0667◦, and 0.9112◦ for ankle, knee, and hip joints, respectively. Finally, a brain–computer interface (BCI) system using neural networks (NNs) for communicating with TCLLE has been addressed. Average classification accuracy of 91.6% has been obtained as the NN classifier performance during the triggering operation.
Important Links:
Go Back
