Deep Learning for Human Locomotion Analysis in Lower-Limb Exoskeletons: A Comparative Study

Omar Coser^1,2*Christian Tamantini^2,3Matteo Tortora^1,4Leonardo Furia¹Rosa Sicilia¹LOREDANA ZOLLO²Paolo Soda^1,5

• ¹Unit of Computer Systems & Bioinformatics, Università Campus Bio-Medico di Roma,Via Álvaro del Portillo, 21, Rome, Italy
• ²Research Unit of Advanced Robotics and Human-centred Technologies, Campus Bio-Medico University, Rome, Lazio, Italy
• ³National Research Council (CNR), Roma, Lazio, Italy
• ⁴University of Genoa, Genoa, Liguria, Italy
• ⁵Umeå University, Umeå, Västerbotten, Sweden

Wearable robotics for lower-limb assistance have become a pivotal area of research, aiming to enhance mobility for individuals with physical impairments or augment the performance of able-bodied users. Accurate and adaptive control systems are essential to ensure seamless interaction between the wearer and the robotic device, particularly when navigating diverse and dynamic terrains. Despite the recent advances in neural networks for time series analysis, no attempts have been directed towards the classification of ground conditions, categorized into five classes, and subsequently determining the ramp's slope and stair's height. In this respect, this paper presents an experimental comparison between eight deep neural network backbones to predict high-level locomotion parameters across diverse terrains. All the models are trained on the publicly available CAMARGO 2021 dataset. IMU-only data equally or outperformed IMU+EMG inputs, promoting a cost-effective and efficient design. Using three IMU sensors, the LSTM model achieved high terrain classification accuracy (0.94 ± 0.04) and accurately estimated ramp slopes (1.95 ± 0.58°), while the CNN-LSTM model proving to be the most effective for stair height estimation-achieved a precision of 15.65 ± 7.40 mm. As a further contribution, SHAP analysis justified sensor reduction without performance loss, ensuring a lightweight setup. The system operates with 2 ms inference time, supporting real-time applications. The code is available at https://github.com/cosbidev/Human-Locomotion-Identification.