Artificial lower limb motion intention recognition algorithm based on BPSOGWO-KNN

Abstract

An artificial lower limb motion intention recognition algorithm based on BPSOGWO-KNN belongs to the technical field of mode recognition, and comprises the following steps: firstly, extracting multi-sensor data in a 200ms time window when a weighing sensor value is just greater than 16N in each gait cycle, denoising and removing abnormal data, and adding a classification label to normal data; extracting seven time domain feature values of data of each dimension in the time window, and performing feature selection by using a BPSOGWO-KNN algorithm; performing optimization training on a nearest neighbor value K in the KNN classifier and a feature weight in an optimal feature subset selected by a BPSOGWO-KNN algorithm by using a BBO algorithm; inputting the nearest neighbor value K and the feature weight value obtained through optimization into an improved KNN classifier; according to the method, redundant feature values are removed, different weights are given according to different contributions of the feature values, the distance mean values of K neighbor values closest to the target under each classification category are compared, pattern recognition is carried out, the algorithm accuracy is greatly improved, and the use safety of a lower limb amputation patient is guaranteed.

Description

technical field [0001] The invention belongs to the technical field of pattern recognition, and in particular relates to a BPSOGWO-KNN-based motion intention recognition algorithm for a lower limb prosthesis. Background technique [0002] Lower extremity amputations are a growing problem requiring advanced technology to help disabled people restore natural gait function in a variety of terrains; most commercially available prosthetic solutions are passive and simple mechanisms make them easy to implement in the clinic However, there are some problems. For example, amputee patients will consume more energy than normal people, causing compensatory movement and causing new damage to other uninjured tissues or parts; powered prosthetics will reduce compensation strategies and increase their own The selected walking speed and the range of comfortable walking speeds, but more advanced and intuitive control strategies are needed before this device can be widely adopted, and it is e...

AI Technical Summary

Problems solved by technology

[0004] Due to the various defects of the existing algorithm, the effect in practical application is not ideal. The present invention proposes to use the swarm intelligence optimization algorithm for feature screening, and again uses the intelligent optimization algorithm to perform weight training on the selected features, and provides a method based on BPSOGWO-KNN Motion Recognition Algorithm for Lower Limb Prosthetics

Images