Simulation Based Optimal Design of Exoskeletons to Reduce Metabolic Cost and Improve Energy Efficiency

Official URL: https://risc01.sabanciuniv.edu/record=b2583861_(Table of contents)

Wearable robotic assistive devices have the potential to improve the metabolic efficiency of human locomotion. Developing exoskeletons that can reduce the metabolic cost of human locomotion is challenging, since there is no systematic mechatronic design approach for the design of such exoskeletons. A systematic design approach necessitates a means for a rigorous and fair comparison of effects of different exoskeleton designs and assistance torque profiles on the human metabolic cost of locomotion. Conducting such investigations through human subject experiments with physical devices is generally infeasible, while design studies relying on musculoskeletal models hold high promise in providing effective design guidelines, as the effect of devices and various assistance torques on muscle recruitment and metabolic cost can be studied systematically. In this thesis, a simulation-based design approach is introduced to systematically design exoskeletons that reduce the metabolic cost of locomotion. Along these lines, a Pareto optimization approach is proposed to enable rigorous and fair comparisons of effects of different exoskeletons designs and assistance torques on the metabolic cost of locomotion, under some realistic physical limits on actuator torques. The proposed systematic mechatronics system design approach is demonstrated by introducing a bi-articular exoskeleton design and comparing its efficiency with commonly used mono-articular exoskeletons. In particular, the power consumption of and metabolic rate reduction due to assistance provided by bi-articular and mono-articular hip-knee exoskeletons are optimized simultaneously during unloaded and loaded walking conditions, and rigorous comparison among such devices is presented. Furthermore, the effect of regeneration on the power consumption of exoskeletons and the detrimental effects of inertial properties of exoskeletons on the metabolic cost of locomotion are studied by superposing these effects on the Pareto-front curves. Our results explain the effect of heavy loads on the optimal assistance profiles and provide guidelines on choosing the optimal device configurations under actuator torque limitation, device inertia, and regeneration effects. The multi-criteria comparison of devices indicates that on the one hand, similar assistance levels can be achieved by both exoskeletons; on the other hand, mono-articular exoskeletons demonstrate better performance on reducing the peak reaction forces, while the power consumptions of bi-articular exoskeletons are less affected by the loading of subjects. Furthermore, the inclusion of device inertia results in significantly less detrimental effects on the metabolic cost of subjects and does not affect the Pareto-optimality of solutions for bi-articular exoskeletons, while non-dominated configurations are significantly affected by the device inertia for mono-articular exoskeletons.