BPNN-Based Real-Time Recognition of Locomotion Modes for an Active Pelvis Orthosis with Different Assistive Strategies

Abstract

Real-time human intent recognition is important for controlling low-limb wearable robots. In this paper, to achieve continuous and precise recognition results on different terrains, we propose a real-time training and recognition method for six locomotion modes including standing, level ground walking, ramp ascending, ramp descending, stair ascending and stair descending. A locomotion recognition system is designed for the real-time recognition purpose with an embedded BPNN-based algorithm. A wearable powered orthosis integrated with this system and two inertial measurement units is used as the experimental setup to evaluate the performance of the designed method while providing hip assistance. Experiments including on-board training and real-time recognition parts are carried out on three able-bodied subjects. The overall recognition accuracies of six locomotion modes based on subject-dependent models are 98.43% and 98.03% respectively, with the wearable orthosis in two different assistance strategies. The cost time of recognition decision delivered to the orthosis is about 0.9[Formula: see text]ms. Experimental results show an effective and promising performance of the proposed method to realize real-time training and recognition for future control of low-limb wearable robots assisting users on different terrains.

Machine Learning Approaches for Activity Recognition and/or Activity Prediction in Locomotion Assistive Devices—A Systematic Review

“…Ai et al 2017 [9] 70.8% TT (4)/Healthy (1) Ankle Prosthesis Beil et al 2018 [10] 90.9% Healthy (10) Exoskeleton Chen et al 2013 [11] 72.7% TT (5)/Healthy (8) Ankle Prosthesis Chen et al 2014 [12] 79.2% TT (1)/Healthy (7) Ankle Prosthesis Chen et al 2015 [13] 77.3% TT (1)/Healthy (5) Ankle Prosthesis Du et al 2012 [14] 75.0% TF (9) Ankle Knee Prosthesis Du et al 2013 [15] 45.8% TF (4) Ankle Knee Prosthesis Feng et al 2019 [16] 77.3% TT (3) Ankle Prosthesis Godiyal et al 2018 [17] 86.4% TF (2)/Healthy (8) Ankle Knee Prosthesis Gong et al 2018 [18] 86.4% Healthy (1) Orthosis Gong et al 2020 [19] 86.4% Healthy (3) Orthosis Hernandez et al 2012 [20] 37.5% TF (1) Ankle Knee Prosthesis Hernandez et al 2013 [21] 54.2% Healthy (1) Ankle Knee Prosthesis Huang et al 2009 [22] 81.8% TF (2)/Healthy (8) Ankle Knee Prosthesis Huang et al 2010 [23] 79.2% TF (1)/Healthy (5) Ankle Knee Prosthesis Huang et al 2011 [24] 83.3% TF (5) Ankle Knee Prosthesis Kim et al 2017 [25] 63.6% Healthy (8) Exoskeleton Liu et al 2016 [26] 70.8% TF (1)/Healthy (6) Ankle Knee Prosthesis…”

Design of Decision Tree Structure with Improved BPNN Nodes for High-Accuracy Locomotion Mode Recognition Using a Single IMU

“…The BPNN algorithm has been widely used in pattern recognition and system modeling [ 34 , 35 ]. Compared to other algorithms, e.g., SVM, QDA, and LDA, the BPNN algorithm has a better effect in locomotion mode recognition [ 36 ]. In this study, the BPNN algorithm was improved by adaptive learning rate and forgetting factor which can accelerate the convergence, and the IBPNN algorithm as the DTS node has good identification ability and convergence speed.…”

Design of Common Data Distributed Loading in Marketing Real Time Cost Control System

“…In order to collect the remote real-time cost control system of electric power marketing, so as to optimize the design of data distribution, firstly, through data collection, the big data sample analysis model of marketing real-time cost control system is established. The collected multi-dimensional power marketing remote real-time cost control database includes [4] . The number of storage resources of each power marketing remote real-time cost control information is x1,x2,...,xn.…”