Home | Help Center

Endless possibilities in academia

Gait prediction for lower limb exoskeleton robots based on real-time adaptive Kalman filtering

Haonan Geng1, Xudong Guo1, Fengqi Zhong2, Haibo Lin1, Guojie Zhang3, Qin Zhang4, Jiaheng Chen1


1School of Health Science and Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China. 2CloudSemi, Pudong New Area, Shanghai 200120, China. 3LingYuan Iron and Steel CO., LTD, Lingyuan 122500, Liaoning Province, China. 4Medical Engineering Department of Northern Jiangsu People’s Hospital, Yangzhou 225001, Jiangsu Province, China.


Address correspondence to: Xudong Guo, School of Health Science and Engineering, University of Shanghai for Science and Technology, No.516 Jungong Road, Shanghai 200093, China. Email: guoxd@usst.edu.cn.


DOI: https://doi.org/10.61189/164995qvdasw


Received September 29, 2024; Accepted December 3, 2024; Published March 31, 2025


Highlights

● The paper develops a gait prediction control strategy for lower limb exoskeleton robots using a real-time adaptive Kalman filtering algorithm, with public gait data from a Clinical Gait Analysis serving as input.

● The model incorporates motor rotation angle, angular velocity, and angular acceleration as core parameters, calculated based on the principles of uniformly accelerated motion. It achieves gait prediction by initializing parameters, calculating Kalman gain, correcting measurements, and updating the covariance matrix.

● A control strategy guided by normal gait parameters enables the exoskeleton to transition efficiently into the desired motion state during startup and gait phase switching. The system employs a microcontroller and Raspberry Pi as its control core, integrated with Bluetooth communication for effective robot control.

Abstract

This paper presents a gait prediction method for lower limb exoskeleton robots using a real-time adaptive Kalman filtering algorithm. The exoskeleton robot targets two user groups: individuals with impaired lower limb motor function requiring rehabilitation training, where the device aids in muscle exercise during walking to facilitate recovery, and healthy individuals using it as a wearable assistive device. To enhance movement intention prediction and improve human-machine coordination, this study focuses on the gait prediction algorithm for walking assistance in healthy users and proposes a gait prediction control strategy based on normal gait orientation. The control system utilizes a microcontroller and Raspberry Pi as its core, enabling functional mode selection through multi-sensor data fusion and effective control of the robot via Bluetooth communication. By comparing the original model algorithm with the proposed real-time updating Kalman filter algorithm, the latter demonstrates feasibility, achieving a prediction error within 1°. This validates the model’s effectiveness in real-time gait prediction.

Keywords: Gait prediction, lower limb exoskeleton robot, Kalman filtering

Geng HN, Guo XD, Zhong FQ, Lin HB, Zhang GJ, Zhang Q, Chen JH. Prog in Med Devices 2025 Mar;3(1): 57-65. doi: 10.61189/164995qvdasw

Copy