Intelligent prosthetic hand for blind people served by fusion of multi-modal information and method for controlling the same

By integrating multimodal information into an intelligent prosthetic hand control method, and utilizing the SSD-MobileNet target detection and tracking algorithm, combined with proximity, electromyography, and tactile feedback, the problem of hand gesture selection and switching timing for visually impaired individuals is solved, achieving highly precise and natural control effects.

CN117653429BActive Publication Date: 2026-06-12TONGJI UNIV
View PDF 0 Cites -1 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
TONGJI UNIV
Filing Date
2023-12-08
Publication Date
2026-06-12

AI Technical Summary

⚠Technical Problem

Existing intelligent prosthetic hands have poor control effects, especially for visually impaired people, as it is difficult to achieve healthy and natural operation behaviors in terms of selecting and switching control gestures.

⚗Method used

The intelligent prosthetic hand control method integrates multimodal information. It uses the SSD-MobileNet target detection algorithm to detect the control object, combines the target tracking algorithm to achieve pixel-level feature alignment, selects appropriate control gestures, and performs visual servo control through proximity feedback, electromyography control, and tactile feedback.

🎯Benefits of technology

It enables highly precise and natural control for visually impaired individuals wearing intelligent prosthetic hands, improving the accuracy and naturalness of control.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117653429B_ABST
    Figure CN117653429B_ABST
Patent Text Reader

Abstract

The application discloses an intelligent artificial limb hand for the blind fusing multi-modal information service and a control method thereof, and comprises the following steps: S1, target detection of a control object is carried out through an SSD-MobileNet target detection algorithm, and a bounding box and a semantic label of the control object are acquired; S2, pixel-level feature alignment is realized through a target tracking algorithm based on information of the bounding box; S3, a suitable control gesture is selected through a matching strategy based on the semantic label; and S4, visual servo control is carried out through fusion of proximal sense feedback information, electromyographic control and tactile feedback information.
Need to check novelty before this filing date? Find Prior Art