Saeed Jamali; Sajjad Taghvaei; Seyyed Arash Haghpanah
Volume 4, Issue 3 , September 2017, , Pages 70-75
Abstract
Background: Sit-to-stand motion is a frequent and challenging task in daily lifeactivities especially for elderly and disabled people. Central nervous system usesseveral strategies ...
Read More
Background: Sit-to-stand motion is a frequent and challenging task in daily lifeactivities especially for elderly and disabled people. Central nervous system usesseveral strategies for sit-to-stand movement. Many studies have been conductedto understand the underlying basis of the optimal approach. Reinforcementlearning (RL) is a suitable method for modeling the control strategies that occurin neuro-musculoskeletal system.Methods: In this paper a dynamic model of human sit-to-stand was derived, andkinematic data of a healthy subject has been extracted in this task. An optimalcontrol problem was formulated considering minimum energy and Q-Learningmethod has been utilized to find the optimal joint moments during sit to standmovement.Results: The simulation results have been compared to the experimental data.The lower extremity joint angles have been simulated and tracked the actualhuman angles extracted from the experiments. Also the joints moments showeda satisfactory precision by the proposed approach.Conclusion: An RL-based algorithm was used to model the human sit-to-stand,in which the model explores the state space with a Markov based approach andfinds the best actions (joint moments) at each state (posture). In this approach themodel successfully performs the task while consuming minimum energy. Thiswas achieved by updating the algorithm in every trial using a Q-learning method.
