Prediction of Limb Joint Angles Based on Multi-Source Signals by GS-GRNN for Exoskeleton Wearer

Abstract

To enable exoskeleton wearers to walk on level ground, estimation of lower limb movement is particularly indispensable. In fact, it allows the exoskeleton to follow the human movement in real time. In this paper, the general regression neural network optimized by golden section algorithm (GS-GRNN) is used to realize prediction of the human lower limb joint angle. The human body hip joint angle and the surface electromyographic (sEMG) signals of the thigh muscles are taken as the inputs of a neural network to predict joint angles of lower limbs. To improve the prediction accuracy in different gait phases, the plantar pressure signals are also added into the input. After that, the error between the prediction result and the actual data decreases significantly. Finally, compared with the prediction result of the BP neural network, GRNN shows splendid prediction performance for its less processing time and higher prediction accuracy.

Continuous Estimation of Knee Joint Angle Based on Surface Electromyography Using a Long Short-Term Memory Neural Network and Time-Advanced Feature

“…If , then . where m is the output unit number, n is the input unit number, and a is a constant between [ 1 , 10 ]. where n is the input unit number.…”

“…In recent years, the exoskeleton robot has attracted the attention of researchers from all over the world due to its broad application prospects in the fields of power assistance, disability assistance, and rehabilitation [1][2][3][4][5]. Human-machine cooperation technology can effectively improve exoskeleton comfort, which has become a current research focus in the exoskeleton field [4,6].…”

Estimation of Knee Joint Angle Using Textile Capacitive Sensor and Artificial Neural Network Implementing with Three Shoe Types at Two Gait Speeds: A Preliminary Investigation

“…Measuring joint angles with those sensor technologies is still under a few limitations related to sensitivity, low accuracy, nonlinearity, complex algorithms, and they require accurate sensor alignment [ 20 ]. Recently, advancements in new technologies that use a machine [ 21 , 22 ] and deep learning approach [ 23 , 24 ] with sensor technologies have contributed to developing various algorithms that provide accurate and effective predictions for lower limb joint angles giving reliable results for further analyses. Moreover, to make the system more convenient and easier to implement, the user avoids finding the optimal sensor placement, so the current method of joint angle assessment could conduct through foot plantar pressure and a deep learning algorithm.…”

“…Sivakumar et al (2019) indicated that the ground reaction force could have the capability to assess ankle joint angle relied on the ANN approach. Moreover, integrating foot pressure data with multi-source data, including surface electromyography (sEMG), hip joint angle, provides high accuracy in building a deep learning model in estimating lower limb joint angles [ 23 ]. The model provided a better correlation efficiency and short calculation times compared to a backpropagation neural network [ 23 ].…”

“…Moreover, integrating foot pressure data with multi-source data, including surface electromyography (sEMG), hip joint angle, provides high accuracy in building a deep learning model in estimating lower limb joint angles [ 23 ]. The model provided a better correlation efficiency and short calculation times compared to a backpropagation neural network [ 23 ]. The foot plantar pressure has been shown the capability, although it can stand alone or integrate with multi-sources signals, to serve as input data for deep learning to assess the joint angles.…”