University of Twente Student Theses

Login

Design of a Momentum-Based Optimal Controller for a Lower Limb Humanoid

Rijt, J.T.J. van de (2022) Design of a Momentum-Based Optimal Controller for a Lower Limb Humanoid.

[img] PDF
3MB
Abstract:Between 250.000 and 500.000 suffer a spinal cord injury (SCI) per year worldwide. Demographics show that young people are among the groups most at risk. Trials suggest that early treatment mayimprove neurological recovery, however, full recovery is often not possible. In the majority of cases, the resulting immobility will lead to a significant reduction in quality of life. Exoskeletons could be used to combat immobility and improve quality of life. The Biomechanical Engineering group at the University of Twente has been developing a lower limp exoskeleton designed to enable individuals with incomplete or complete SCI. While the exoskeleton is mechanically functioning and a trajectory generator is being developed, a controller that adequately executes the trajectory, while maintaining balance is still absent. This thesis explores the option of using a Momentum Based Controller. The problem is approached by a two-dimensional lower limb humanoid model with point feet. For this model, a simple trajectory generator is created that translates a CoM trajectory into momentum rate and feet acceleration trajectories. A quadratic program is used to reconcile the trajectory goals with the dynamics and find the optimal joint torque while minimizing the energy expenditure. Using this system, the control parameters are analyzed during two situations: walking at a constant velocity and recovering for a push. The results show that all tracking goals can be used as decision variables and none have to be fully constrained. For the cost, a high priority should be given to feet and chest angle tracking. If these are adequate, the momentum tracking can be done with a lower cost and balanced with operational costs. While the model used is relatively simple compared to the exoskeleton, the recommendations made can still be applied to more complex situations and give useful insight into the control priorities of human-like walking.
Item Type:Essay (Master)
Faculty:TNW: Science and Technology
Subject:52 mechanical engineering
Programme:Biomedical Engineering MSc (66226)
Link to this item:https://purl.utwente.nl/essays/92234
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page