Trajectory optimization for an active full lower body exoskeleton
Shahriar Sheikh Aboumasoudi (ET-BE and ET-MS3), Arvid Q. L. Keemink (ET-BE), Hakan Koroglu (ET-MS3), and Herman van der Kooij (ET-BE)
Abstract
Introduction:
Full lower body exoskeletons can help people with paraplegia to stand up and walk again. One way to achieve this goal is through having the robot follow a prespecified trajectory. However, due to the complexity of these dynamical systems stemming from the fact that they have inherent underactuation and that they make and break contact with the environment as they walk, hand-crafting such trajectories is a cumbersome task. In this research, we try to automate this task by means of formulating the design process as an optimization problem that can be solved using an off-the-shelf solver.
Method:
In order to achieve this goal, first the constraints corresponding to a stable gait are identified and mathematically described. Then all of these constraints for the whole period of the gait are gathered together and an objective function that reflects the desired overall behavior of the gait e.g. mechanical energy efficiency, is defined. This way, an optimization problem that has as its variables the joint trajectories, required joint torques of the robot, and the ground reaction forces associated with the gait, is constructed. Next, an off-the-shelf solver can be employed to solve this optimization problem.
Results:
The aforementioned method has been used to generate trajectories for a full lower body exoskeleton and their feasibility was shown by implementing them on the exoskeleton.