Terrain Detection for Lower Limb Prostheses Using Depth Camera Assistance

This study proposes a novel method utilizing depth cameras to assist lower limb prosthetic systems in terrain recognition and gait prediction. The algorithm, based on geometric methods, accurately classifies terrain types and calculates terrain parameters such as step height, width, and slope. The experimental validation demonstrates that this algorithm possesses strong classification capability and maintains terrain parameter calculation errors at a low level, thereby achieving accurate terrain reconstruction. Additionally, the present study introduces a camera-based method for stride length calculation, which under certain conditions can compute precise stride lengths, thus contributing to enhancing the stability of prosthetic limb control. This study marks a promising advancement in the use of visual sensors for prosthetic devices, paving the way for enhanced user comfort and stability during walking activities.