Feedback Control of an Exoskeleton for Paraplegics: Toward Robustly Stable Hands-free Dynamic Walking

This manuscript presents control of a high-DOF fully actuated lower-limb\nexoskeleton for paraplegic individuals. The key novelty is the ability for the\nuser to walk without the use of crutches or other external means of\nstabilization. We harness the power of modern optimization techniques and\nsupervised machine learning to develop a smooth feedback control policy that\nprovides robust velocity regulation and perturbation rejection. Preliminary\nevaluation of the stability and robustness of the proposed approach is\ndemonstrated through the Gazebo simulation environment. In addition,\npreliminary experimental results with (complete) paraplegic individuals are\nincluded for the previous version of the controller.\n