Design and Control of an Upper Limb Bionic Exoskeleton Rehabilitation Device Based on Tensegrity Structure

Abstract

Upper limb exoskeleton rehabilitation devices can improve the quality of rehabilitation and relieve the pressure of rehabilitation medical treatment, which is a research hotspot in the field of medical robots. Aiming at the problems such as large volume, high cost, low comfort, and difficulty in promotion of traditional exoskeleton rehabilitation devices, and considering the lightweight, discontinuous, high flexibility, and high biomimetic characteristics of tensegrity structure, we designed an upper limb bionic exoskeleton rehabilitation device based on tensegrity structure. First, this article uses mapping methods to establish a mapping model for upper limb exoskeletons based on the tensegrity structure and designs the overall structure of upper limb exoskeletons based on the mapping model. Second, a bionic elbow joint device based on gear and rack was designed, and the stability of the bionic elbow joint was proved using the positive definite matrix method. This device can simulate the micro displacement between bones of the human elbow joint, improve the axial matching ability between the upper limbs and the rehabilitation device, and enhance the comfort of rehabilitation. Third, an impedance control scheme based on back propagation (BP) neural network was designed to address the low control accuracy of flexible structures and patient spasms. Finally, we designed the impedance control scheme of the PSO-BP neural network based on a fuzzy rehabilitation state evaluator. The experimental results show that the exoskeleton rehabilitation device has good flexion motion stability and assist ability and has significant advantages in volume and mobility. The control strategy proposed in this paper has high control precision and adaptive ability and has potential application value in the field of medical rehabilitation.

Figure 1

Bionic elbow joint. (a–d) Mechanical mapping of upper limb. (e and f) Bionic elbow joint mapping mode based on tensegrity structure. (g–i) Structural optimization of bionic elbow joint. (j and k) Stability analysis of bionic elbow joint structure. (l) The locking structure of bionic elbow joint.

Figure 2

Structure of upper limb exoskeleton rehabilitation device. (a) Overall structure. (b) Wrist structure. (c) Forearm rotation structure. (d) Forearm structure. (e) Upper arm structure. (f–k) Optimization process of forearm rotation structure.

Figure 3

Control scheme. (a) The impedance control scheme of PSO–BP neural network based on fuzzy rehabilitation state evaluator. (b) The input–output relationship of fuzzy rehabilitation state evaluator. (c) The impedance parameter controller based on PSO–BP neural network. (d and e) Simulation result.

Figure 4

(a–d) Tensegrity structure stability test. (a–c) Horizontal forces. (d) Vertical forces. (e–g) Motion speed test. (h) Load test.

Figure 5

(a and b) Rehabilitation trajectory test. (a) Motion smoothness tests. (b) Motion trajectory test. (c–e) Wearing comfort test. (c) The wearing structure. (d) The export results of pronation and supination. (e) The export results of flexion movement.

Figure 6

(a–d) Flexion motion stability test. (a) The structure of stability test device. (b) The maximum driving force. (c) The minimum driving force. (d) Capture images of flexion motion. (e–i) Assist ability test. (e) The structure of assist ability test device. (f) The maximum driving force. (g) The minimum driving force. (h) The variation of tension meter reading F and angle R. (i) The angle–tension curve.