A novel framework for designing a multi-DoF prosthetic wrist control using machine learning

Continuous joint velocity estimation using CNN-based deep learning for multi-DoF prosthetic wrist for activities of daily living

Introduction

Myoelectric control of prostheses is a long-established technique, using surface electromyography (sEMG) to detect user intention and perform subsequent mechanical actions. Most machine learning models utilized in control systems are trained using isolated movements that do not reflect the natural movements occurring during daily activities. Moreover, movements are often affected by arm postures, the duration of activities, and personal habits. It is crucial to have a control system for multi-degree-of-freedom (DoF) prosthetic arms that is trained using sEMG data collected from activities of daily living (ADL) tasks.

Method

This work focuses on two major functional wrist movements: pronation-supination and dart-throwing movement (DTM), and introduces a new wrist control system that directly maps sEMG signals to the joint velocities of the multi-DoF wrist. Additionally, a specific training strategy (Quick training) is proposed that enables the controller to be applied to new subjects and handle situations where sensors may displace during daily living, muscles can become fatigued, or sensors can become contaminated (e.g., due to sweat). The prosthetic wrist controller is designed based on data from 24 participants and its performance is evaluated using the Root Mean Square Error (RMSE) and Pearson Correlation.

Result

The results are found to depend on the characteristics of the tasks. For example, tasks with dart-throwing motion show smaller RSME values (Hammer: 6.68 deg/s and Cup: 7.92 deg/s) compared to tasks with pronation-supination (Bulb: 43.98 deg/s and Screw: 53.64 deg/s). The proposed control technique utilizing Quick training demonstrates a decrease in the average root mean square error (RMSE) value by 35% and an increase in the average Pearson correlation value by 40% across all four ADL tasks.

Feasibility Evaluation of Online Classification-based Control for Gross Movement in a 2-DoF Prosthetic Arm

Regression and classification models have been extensively studied to exploit the myoelectric and kinematic input information from the residual limb for the control of multiple degree-of-freedom (DoF) powered prostheses. The gross movement control of above-elbow prostheses is mainly based on regression models which map the available inputs to continuous prosthetic poses. However, the regression output is sensitive to the variation in the input signal. The myoelectric signal variation is usually large due to unintentional muscle contractions, which can deteriorate the user-in-the-loop performance with respect to the offline analysis. Alternatively, the classification models offer the advantage of being more robust to the input signal variation, but they were predominantly used for fine motor functions such as grasping. For gross motor functions, the discrete output may cause issues. Therefore, this work attempts to investigate the feasibility of utilising the classification model to control a 2-DoF transhumeral prosthesis for gross movement. The performance of 6 able-bodied subjects was evaluated in performing reaching and orientation matching tasks with a prosthetic arm in a virtual reality environment. The results were compared with the case of using their intact arms and existing results using the regression model. Our findings indicate that the classification-based method provides comparable performance to the regression model, making it a potential alternative for gross arm movement in multi-DoF prosthetic arms.

Sensor Selection With Composite Features in Identifying User-Intended Poses for Human-Prosthetic Interfaces

A Human-Prosthetic Interface (HPI) serves to estimate and realise the limb pose intended by the human user, using the information obtained from sensors worn by the user. In recent studies, the HPI maps multi-joint limb poses (i.e. coordinated movement of the body and limbs) to the inputs of multiple sensors. This is in contrast to the conventional methods where each degree of freedom of the powered prosthesis is mapped to the input of one/a pair of sensors. In this approach, it is necessary to systematically select sensors that carry the most information for the intended set of poses, to improve system accuracy and/or minimise the number of sensors, thus the complexity, in the prosthetic system. In this paper, sensor selection process is systematically formulated to maximise the information contained in the input features for a given number of sensors. Most importantly, it accounts for composite features, which are features requiring information from multiple sensors. Such composite features exist and are important in HPIs as we seek to capture coordinated motion involving movements of multiple limb and body segments. A non-convex optimisation problem is formulated which accounts for the constraint introduced by the composite features. A projection matrix is utilised as the optimisation variable to select intended features for evaluation. The problem is solved by the proposed Sensor Selection with Composite Features (SS-CF) algorithm which adapts convex-relaxation techniques. The SS-CF is benchmarked against HPI with expert-selected sensors in the literature and against a greedy heuristic method. The outcome demonstrated the efficacy of the SS-CF algorithm.

Artificial Neural Network-Based Activities Classification, Gait Phase Estimation, and Prediction

Gait patterns are critical to health monitoring, gait impairment assessment, and wearable device control. Unrhythmic gait pattern detection under community-based conditions is a new frontier in this area. The present paper describes a high-accuracy gait phase estimation and prediction algorithm built on a two-stage artificial neural network. This work targets to develop an algorithm that can estimate and predict the gait cycle in real time using a portable controller with only two IMU sensors (one on each thigh) in the community setting. Our algorithm can detect the gait phase in unrhythmic conditions during walking, stair ascending, and stair descending, and classify these activities with standing. Moreover, our algorithm is able to predict both future intra- and inter-stride gait phases, offering a potential means to improve wearable device controller performance. The proposed data-driven algorithm is based on a dataset consisting of 5 able-bodied subjects and validated on 3 different able-bodied subjects. Under unrhythmic activity situations, validation shows that the algorithm can accurately identify multiple activities with 99.55% accuracy, and estimate ([Formula: see text]: 6.3%) and predict 200-ms-ahead ([Formula: see text]: 8.6%) the gait phase percentage in real time, which are on average 57.7 and 54.0% smaller than the error from the event-based method in the same conditions. This study showcases a solution to estimate and predict gait status for multiple unrhythmic activities, which may be deployed to controllers for wearable robots or health monitoring devices.