Motion Intention Estimation Algorithm for Intelligent Rollator Users Based on Multi-sensor Fusion Technology

The intelligent rollator serves as a crucial assistive tool in various applications, aiding individuals with mobility impairments in maintaining independence and overcoming daily challenges. Among these applications, achieving an accurate estimation of the user's motor intention is crucial for enabling precise control of the intelligent rollator. We present the design of an intelligent rollator tailored for individuals with lower limb mobility issues. The rollator, equipped with 3D force sensors and 2D LiDAR, collects data on the user's walking status. However, the algorithm for estimating the user's motion intention faces challenges due to susceptibility to interference from odometer errors in the original linear velocity estimation method. To address this challenge, we introduce a method for estimating the user's intention linear velocity based on the relative position of the user's legs. Additionally, to address the challenge of accurately estimating the intended angular velocity of individuals with lower limb mobility issues using only 3D force sensors, we introduce an angular velocity estimation method based on multisensor fusion technology. Moreover, to address the lag induced by algorithm data processing, we integrate the Informer model into the algorithm to overcome challenges related to real-time human-computer interaction. The results indicate that the proposed linear velocity estimation method achieves an accuracy of 98.44%, while the angular velocity estimation algorithm exhibits an accuracy of 96.6%. As evidenced by the EMG(ElectroMyoGraphy) signals, users can use the proposed intelligent walker more easily. Furthermore, the integration of the Informer model into the intelligent rollator further enhances user convenience. Several experiments are conducted to confirm the effectiveness of the proposed method.