Research On Key Technologies Of Sensing And Control For Intelligent Lower Limb Prosthesis

The amount of the handicapped people in China is approximately7.8%of the total population. They suffer from the defects of the bodies physically and mentally. As the compensation tools for walking function, Lower limb amputee can restore the movement ability of amputees. of amputees. The use of traditional mechanical prosthesis is restricted by rigid gait, high metabolic energy consumption and difficulty in gait velocity adjustment. Intelligent lower limb prosthesis controlled by MCU can change gait velocity optionally and accommodate changing walking condition and the gait is more like that of a healthy human. In addition, intelligent prosthesis is much better in mobility, adaptation to complex environment and coordination with healthy leg than common mechanical prosthesis.It has been a great endeavor for information exchange and coordinated control between human and his/her intelligent prosthesis. For above-knee prosthesis, amputee couldn’t control knee joint movement. To achieve symmetrical stable walking, intelligent prosthesis must be able to perceive the movement intention of amputee so as to actuate its knee joint accordingly when walking on various terrains with different cadences and stride length. The information for movement intention estimation of lower limbs can be divided into three categories: kinetics, kinematics and electromyography. The method of acquiring kinematics by using inertial sensors has the advantages of simple, reliable and low-cost, and has been wildly used in either research or implementation. In this dissertation, we focus on perception and control of intelligent prosthesis, and make deep study on inertial sensor based gait analysis by combining theory and practice. The main work of the research includes:Corresponding to deficiency in perception and control for non-intelligent prosthesis, a novel perception and control scheme based on inertial sensors is proposed to deal with difficulties in installation, low reliability and unnatural human-machine interaction.Lower limb kinematics acquired by inertial sensors is a classical non-linear, non-stationery signal. A novel solution for signal local mean is proposed to deal with problems of boundary effect and imprecise approximation of local mean function. Experimental results demonstrate the effectiveness of the proposed method. The method is then used in de-noising of signals acquired from inertial sensors. Experimental results validate that this algorithm could effectively analyze the noise and drift trends of inertial sensor.A new gait event detection method is proposed in this dissertation based on sagittal plane swing angular velocity and acceleration energy extreme. This method could be successfully used in gait event detection of different subjects on different terrains with various motion patterns. A gait feature extraction algorithm based on linear discriminant analysis is proposed and the features are then classified by support vector machine. The98.2%of average accuracy is obtained for different subjects on different terrains. The proposed scheme can achieve high classification precision and good stability.Zero velocity update method is improved in this dissertation by using Gaussian Process. Experimental results show the proposed scheme has advantages of high precision and insensitivity to gait cycle. It can effectively improve the precision of gait quantification and can provide correct gait parameters for gait planning.An online gait synthesis scheme for intelligent prosthesis is proposed in this dissertation. By employing discrete-time Fourier series, the samples were extracted into features, and further dimensionally reduced by using PCA to simplified features. By using ANNs, the nonlinear functional relations between gait parameters (i.e. cadence and stride) and simplified features that can be used to reconstruct angles of hip and knee joints were established. Walking experiments at slow, intermediate and fast speeds were conducted to validate the effectiveness of the proposed scheme. The results indicate that the synthesized gait is smooth, efficient and human-like.An intelligent prosthesis-amputee interaction scheme based on coordinated control is proposed. To realize symmetrical stable walking, master-slave coordinated control strategy is proposed. According to information acquired by gyroscopes, intelligent prosthesis can recognize the motion intention of healthy leg and reconstructed its kinematic variables based on inter-joint coordination. By employing iterative learning control, gait tracking of intelligent prosthesis to healthy leg was archived. Real environment walking experiments and simulation validated the correctness and effectiveness of the proposed scheme.