Abstract:This article describes a multimodal fusion data acquisition and processing system about electromyography for dynamic movement recognition and bioelectrical impedance for key posture recognition. In addition, a new dynamic–static fusion algorithm strategy is designed. Experimental results demonstrate that proposed algorithm significantly improves the accuracy of movement recognition, and has been successfully applied to the remote control of Universal Robot. Accurate recognition of human dynamic movement is essential for seamless human–machine interaction (HMI) across various domains. However, most of the existing methods are single‐modal movement recognition, which has inherent limitations, such as limited feature representation and instability to noise, which will affect its practical performance. To address these limitations, this article proposes a novel fusion approach that can integrate two biological signals, including electromyography (EMG) and bioelectrical impedance (BI). The fusion method combines EMG for capturing dynamic movement features and BI for discerning key postures representing discrete points within dynamic movements. In this method, the identification of key postures and their temporal sequences provide a guiding framework for the selection and weighted correction of probability prediction matrices in EMG‐based dynamic recognition. To verify the effectiveness of the method, six dynamic upper limb movements and nine key postures are defined, and a Universal Robot that can follow movements is employed for experimental validation. Experimental results demonstrate that the recognition accuracy of the dynamic movement reaches 96.2%, representing an improvement of nearly 10% compared with single‐modal signal. This study illustrates the potential of multimodal fusion of EMG and BI in movement recognition, with broad prospects for application in HMI fields.

A hybrid HMM/SVM classifier for motion recognition using μIMU data

Towards Multi-Classification of Human Motions Using Micro IMU and SVM Training Process

Real-time Recognition of Multi-category Human Motion Using Μimu Data

Towards HMM based human motion recognition using MEMS inertial sensors

PCA/ICA-based SVM for Fall Recognition Using MEMS Motion Sensing Data.

Classify Motion Model via SVM to Track Underwater Maneuvering Target

HMM-based IMU Data Processing for Arm Gesture Classification and Motion Tracking

Real-Time Onboard Human Motion Recognition Based On Inertial Measurement Units

An SVM-Based Pedestrian Gait Recognition Algorithm Using a Foot-Mounted IMU

Machine Learning-Based Human Motion Recognition via Wearable plastic Fiber Sensing System

Motion Based Event Recognition Using HMM

Multi-Classification Algorithm for Human Motion Recognition Based on IR-UWB Radar

A Scalable and Wearable Self-Sensing IMU Sensor Network for Personalized Human Motion and Deformation Capture

Human action recognition in videos using hybrid motion features

Wi-Motion: A Robust Human Activity Recognition Using WiFi Signals

IMU-Based Gait Phase Recognition for Stroke Survivors

IMU-based Modularized Wearable Device for Human Motion Classification

Gait Recognition Using Optical Motion Capture: A Decision Fusion Based Method

On the Design of a Wearable Multi-sensor System for Recognizing Motion Modes and Sit-to-stand Transition

Research on motion recognition based on multi-dimensional sensing data and deep learning algorithms

A High‐Precision Dynamic Movement Recognition Algorithm Using Multimodal Biological Signals for Human–Machine Interaction