Research

Real-Time Locomotion Transitions Detection

Abstract: Accurate detection of locomotion transitions, such as walking to sitting or stair ascent, is essential for effective control of lower-limb exoskeletons, as each mode requires specific assistance. This study presents a real-time transition detection approach for assistive devices. Link-1 Link-2


ExoRecovery: Push Recovery with a Lower-Limb Exoskeleton based on Stepping Strategy

Abstract: Introduced a control framework for omnidirectional recovery step planning, optimizing step duration and position based on real-time forces. Using an impedance controller and human-like foot trajectory mapping, this approach enhances exoskeleton stability and safety under unforeseen forces. Validated through simulations and human subject experiments. Link


Partial Assistance to Gait and Balance during Daily Living Activities

Abstract: Aimed at improving the quality of life for people with walking impairments, this project developed controllers for collaborative assistance during daily activities. The exoskeleton provides partial support to gait and balance, aligning with user movement intentions. Link


Co-Design Optimization of a Variable Stiffness Actuator

Abstract: This study focused on optimizing a variable stiffness actuator at the system level, comparing outcomes with optimal control inputs. Multiple system-level optimal designs were created, resulting in a unique variable stiffness actuator system model with compact features. Link


Two-Port Analysis of Stability and Transparency in Series Damped Elastic Actuation

Abstract: This project developed non-conservative passivity bounds using frequency-domain passivity analysis and added dissipative elements to improve rendering performance. Link-1 Link-2

ExoRecovery




Cable-Driven Aerial Camera System

Abstract: A cable-driven aerial camera system was developed to overcome the limitations of stationary cameras in capturing smooth and dynamic footage for large-scale events such as sports and concerts. The system, designed for a 4m x 3.5m workspace, utilizes four cables to control the camera's position and orientation with 1 cm accuracy.