Joint angle error compensation experiment device and method based on nerve network

Abstract

The invention discloses a joint angle error compensation experiment device and method based on a nerve network and relates to the field of wearable exoskeleton. The device includes four bands, two installing plates, a first gyroscope, a second gyroscope, an encoder and a signal collection plate. The two installing plates are connected with each other at the ends, and the two installing plates canrotate around a joint of the two installing plates; the four bands are fixedly installed on the inner side walls of the installing plates, and two bands correspond to one installing plate; the first gyroscope is fixedly installed on the outer side wall of a top installing plate; the second gyroscope is fixedly installed on the outer side wall of a bottom installing plate; the encoder is fixedly installed on the outer side wall of the joint of the two installing plates; and the signal collection plate is fixedly installed between the first gyroscope and the encoder. According to the joint angleerror compensation experiment device and method based on the nerve network, a gait walking experiment is beforehand performed, a learning sample is obtained, an optimal model parameter is obtained through nerve network training, and the accurate measurement of knee joint angles is achieved through the wearing of two gyroscopes and trained nerve network models after an exoskeleton structure is worn by people.

Description

technical field [0001] The invention relates to the field of a wearable exoskeleton, in particular to a neural network-based experimental device and method for joint angle error compensation. Background technique [0002] The human gait information collection system needs to collect accurate knee joint angles of the human body, which are mostly used in the evaluation of human motion ability, wearable exoskeleton gait tracking control, etc. Rehabilitation training exoskeletons, disabled walking exoskeletons, weight-bearing exoskeletons and other exoskeleton equipment require a coordinated gait pattern between the human body and the exoskeleton equipment, but there are indelible differences between the two, so the wearers of the above exoskeleton equipment Acquisition of gait information and analysis of gait data classification, tracking and prediction are required in order to realize gait tracking and control of lower extremity exoskeleton and improve its dynamic stability. ...

AI Technical Summary

Problems solved by technology

[0003] Image sequence analysis, using computer vision technology to detect motion and moving objects from image sequences and perform motion analysis, tracking or recognition on them, real-time detection of gait data requires high cost, large amount of data processing and serious delay; EMG signal detection , through the myoelectric sensor to detect the myoelectric signal on the surface of the human body, and to detect and analyze the movement gait. The sensors and supporting devices used are expensive, and the signal recognition and classification are difficult, and the detection of the myoelectric signal is subject to the detection position, sweat, temperature, etc. etc., it is easily affected by signal interference, and the repeatability is not high; angle / angular velocity detection: through the integral calculation of the output data of angular velocity sensors such as gyroscopes, the calculation of the angle of human motion, and then the prediction and analysis of human gait tracking, etc., due to the use of The integral algorithm leads to the continuous accumulation of errors, which gradually increases the error of the joint angle; the accelerometer detection method: uses the acceleration sensor to detect the gravity vector component, calculates the movement angle of the human body limbs, and then performs gait tracking and analysis operations, etc., but this method only Applicable to static calculation, not extended to dynamic joint angle calculation

Images