Closed-chain leg sliding sensation dataset for different terrains

This dataset contains ground surface characteristic recognition data based on tactile sensing, capturing mechanical vibration signals from different terrain surfaces (such as grassland, sand, gravel, concrete, and rock) using tri-axial accelerometers. The data has undergone systematic processing including outlier removal, abrupt change detection, cycle segmentation, and standardization to form structured terrain feature cycle samples. Each sample contains complete three-channel (x, y, z) mechanical response time-series data, which can be directly used for the development and verification of terrain classification and recognition algorithms. The dataset comprises thousands of high-quality cycle samples, providing a valuable experimental foundation for research in robot perception, all-terrain mobility, and intelligent navigation.