Lower limb motion intention recognition method based on empirical rule in combination with machine learning

Abstract

The invention discloses a lower limb motion intention recognition method based on empirical rules in combination with machine learning, and belongs to the technical field of mode recognition. The method comprises the following steps: acquiring data generated by a knee joint angle sensor, a weighing sensor and an IMU sensor, denoising and removing abnormal values, and then designing three classifiers, and carrying out human body intention recognition through empirical threshold judgment and an improved weighted KNN algorithm. According to the method, seven common motion modes can be accurately identified only by using a small number of mechanical sensors, the data volume of a training set required for improving the weighted KNN algorithm is greatly reduced, the operation time of the algorithm on an STM32 single chip microcomputer is shortened, and real-time prediction of the motion state of the human body is ensured. On the basis of experiments, human body intention recognition is carried out in combination with empirical rules and a machine learning algorithm, and the method aims at promoting development of commercial prostheses and facilitating daily use of lower limb amputation patients.

Description

technical field [0001] The invention belongs to the technical field of pattern recognition, and in particular relates to a method for recognizing motion intentions of lower limb amputee patients based on empirical rules combined with machine learning. Background technique [0002] Accurately judging the exercise intention of lower limb amputee patients can effectively reduce the risk of falls and enable them to complete their daily exercise needs. Commercial prosthetics mostly use empirical threshold judgments as the basis for switching motion states, while some scholars try to use machine learning classification algorithms for intention recognition. Since the machine learning algorithm requires a large amount of data for offline training, and the trained model also has a large number of parameters, there are still some challenges in using STM32 for real-time predictive recognition. The KNN algorithm is a classic machine learning algorithm, which has the advantages of high ...

AI Technical Summary

Problems solved by technology

Many algorithms can achieve an accuracy rate of more than 95% for switching between sports modes and modes, but an error rate of about 5% will still cause falls, and redundant sensors will also burden amputees

[0004] Due to the various defects of the existing algorithms, the effect is not ideal in practical applications, and the algorithm needs to be improved

Images