About Physical Human Robotic Interaction for Assistive Exoskeletons
Time: Wed 2021-10-20 13.00
Location: 13:202, University of Gävle, Kungsbäckvägen 47, SE-80176, Sweden, Gävle (English)
Subject area: Computer Science
Doctoral student: Nauman Masud , Robotik, perception och lärande, RPL
Opponent: Professor Shaoping Bai, Aalborg University, Denmark
Supervisor: Professor Magnus Isaksson, University of Gävle; Ass. Professor Christian Smith, Robotik, perception och lärande, RPL
Abstract
The thesis work has contributed to the field of assistive robotics. The physical interaction between the exoskeleton and human has been studied by considering performance both at the joint as well as task space level of the exoskeleton. For ensuring safe and enhancing physical human interactions for elderly persons, special consideration has been given to problems due to the weight and number of actuators of the exoskeleton. Specific scenarios have been formulated to investigate fundamental requirements and where innovations have been developed for originality and academic content after the initial phases of the investigations.
Research on the lower active degree of freedom serial robotic manipulators has hence gained importance from the perspective of developing assistive exoskeletons that are light and can provide effective assistance to the user despite being less dexterous as compared to the high degree of freedom counterparts. Control methodologies have been investigated and developed for low active degrees of freedom exoskeleton that can ensure stable and safe human interaction. With this focus in mind, a specific strategy has been proposed to compensate for the nonlinear dynamics of the human exoskeleton system at the joint level. Furthermore, active compliance through impedance control in conjunction with passive compliance has been proposed to provide safe human interaction. The interactive human-machine-impedance-loop with a human as a dynamic environment (which contrasts with the existing approaches) and exoskeleton as a controlled impedance has also been investigated for stability and performance. This, in turn, has provided the sound-realistic basis for the development of cascaded strategies to ensure safe interaction between humans and the exoskeleton. A Hybrid switching control strategy has also been developed to simultaneously improve the load torque compensation performance as well as the stability of the human-exoskeleton system in case of actuator saturation. Methodology for proper selection of joint actuators along with a framework for finding the desired assistive forces based on the actual end-user group data has also been developed. A distributive controller area network-based control architecture has also been proposed for a lower-body exoskeleton. Lower and upper body exoskeleton test rigs and prototypes along with the associated hardware have been developed in tandem to verify the proposed strategies both at the joint and task space level. A new control strategy capable of imparting simultaneous impedance-based force tracking control for both the compliant contact supports of the lower-body exoskeleton(in task-space) using DOB-based-DLTC (at joint-space) has also has also been proposed