HERO Glove Insight: Utilizing Computer Vision and Force Sensors for Object-Specific Force Control

This research developed and integrated new mechatronic features to the Hand Extension Robot Orthosis (HERO) Glove soft exoskeleton to improve the assistive abilities of the device. The added features include low weight actuators, anchoring structures, force sensors, a camera and a cascaded computer vision driven Proportional Integral Derivative (PID) algorithm. Supporting electronics were developed to ensure that the integrated system can provide accurate grip force control based on situational awareness. The upgraded version of the HERO glove was evaluated and showed a 55% increase in grip strength, improved durability, and new force sensing and control capabilities as accurate as 0.1 Newtons. The algorithm uses computer vision software to recognize the object that the user intends to grip and then adjusts the target grip force commanded during autonomous gripping. This enhanced system, HERO Glove Insight is designed to overcome usability challenges in grasp stability and grip force modulation faced by stroke and spinal cord injury survivors when utilizing soft hand exoskeletons to enable their independence in home and community settings.