2021 IEEE International Conference on Real-time Computing and Robotics (RCAR) | 2021
Terrain environment classification and recognition for soft lower-limb exosuit
Abstract
Reliable environmental context prediction is critical for wearable robots (exoskeletons or prostheses) to assist terrain-adaptive locomotion. Inspired by the mechanism of human perception of the environment, the vision sensor can be introduced into the soft lower limb exosuit robot so that future environmental information can be used to improve the control. In this paper, a stable terrain classification and recognition system (TCRS) is designed for lower-limb soft exosuit robots. When the exosuit robot perceives the environment, it uses the residual network to train the terrain classifier to classify the terrain. During walking, the wearable soft exosuit robot collects the indoor environment, preprocesses, and marks the sampled images separately. The experimental results show that the classification accuracy is more than 90%, and the highest is 97.5%. The results in this study may lead to novel context recognition strategies in reliable decision-making, efficient sensor fusion, and improved intelligent system design in various applications.