The Cooperative Control For Lower Extremity Rehabilitation Robot

Rehabilitation training is a common way to patients who have lower limb motor dysfunctions. Due to the limitation of the traditional human-based rehabilitation training method, Lower Extremity Rehabilitation Robot(LERR) working in an automatic training mode has been developed and widely used in recent years. Along with the increase of the aged population in our country, the incidence of motor dysfunction is continually raised, which calls for a huge demanding of the robot-based rehabilitation products in our society. However, in the field of rehabilitation robotics, the market in our country is mainly dominated by foreign products and lacks our own mature products. Therefore, it is urgent to develop the LERR with our independent intellectual property rights.Developing an automatic rehabilitation device based on robotic technology to replace the traditional human-assisted walking training approach performed by therapists has become the current research focus in the robotics field. Walking rehabilitation training robot system is a such automatic device, in which different control strategies can be implemented to make the patients’ lower limbs move along a normal reference gait trajectory and thus scientific rehabilitation training for patients with lower limbs motor dysfunctions can be realized. LERR system can increase the training time, improve the effect of rehabilitation and reduce the doctor’s work strength. Therefore, based on the scientific training using the LERR system lower extremity motor functions of patients are eventually recovered.This thesis aims at the dynamics modeling of Lower Extremity Rehabilitation Robot system and the development of the corresponding cooperative control methods. Due to the difficulty to establish the accurate dynamics model of LERR system in reality, a gait trajectory adaptive cooperative control method based on a robust PID controller is proposed to ensure the system stability, achieve the desired gait trajectory adaptation and tracking performance even under system model uncertainty. A Youla-Kucera parameterized adaptive regulator for LERR based on the robust PID controller is also proposed for the post-rehabilitation training stage, which is designed to enhance patients’ muscle strength and achieve joint trajectories in a full active training mode. In the aspect of the treadmill speed control, an acceleration constrained treadmill speed cooperative controller is developed to adjust automatically the treadmill speed according to the patient’s movement intention in a more safe way and thus improve the effectiveness of the overall patient’s rehabilitation training process.In chapter 1, the review of the development of the LERR at home and abroad is briefly introduced. The robot controllers and the corresponding control strategies are also introduced and analyzed. In chapter 2, The dynamics of the LERR system is modeled. For the no-cooperative training mode, the dynamic model of LERR is established using the Lagrange method. While for the cooperative training mode, the dynamic model is established using the Newton-Euler approach and thus the active force between the robot and the patient can be easily analyzed. In chapter 3, the invertibility decoupling method for the nonlinear systems is first introduced, and then is applied to the LERR system to obtain the decoupled sub-systems. With the growth of the patient active participation training awareness and the resulting patient-robot active interaction force, the impedance controller is designed to deal with the patient active interaction intention under the assumption that the system dynamic model is perfectly known. Meanwhile, in order to extract the patient’s movement intention online, an adaptive cooperative algorithm based on the impedance control is also proposed and the designed controllers are finally verified in the Matlab/Simulink and Adam co-simulation environment. In chapter 4, A gait reference trajectory adaptation control method based on a robust PID controller which takes into account of the model uncertainty of the robot system is proposed. This approach is developed to make the inner control system stable even under system model uncertainty, and thus the desired gait trajectory tracking and adaptive adjustment can be achieved. The controller parameters are optimized using the properly formulated linear matrix inequalities. In the post-rehabilitation training stage, the patient’s locomotor has gradually recovered and the patient could walk in a full active training mode. Therefore,an Youla-Kucera parameterized adaptive regulator for LERR based on the robust PID controller is further developed to control the robot to track the fully unknown patient’s trajectory with a necessary assistant training force. The performances of the proposed adaptive controller are verified in the MATLAB/Simulink. In chapter 5, the cooperative treadmill speed control problem is studied. In the traditional rehabilitation system, the treadmill speed is usually set to a fixed value or adjusted manually by the therapist, whereas patient-determined phases of accelerations and decelerations cannot be automatically performed. Therefore, a treadmill speed adaptation controller for LERR, which automatically adjusts the treadmill speed based on the measured interaction force between the patient and the treadmill, is proposed. The designed treadmill speed adaptation controller can make the treadmill realize the automatic speed adjustment according to the patient’s movement intention and thus achieve the cooperative control between the patient and the treadmill. In order to effectively restrain the treadmill acceleration and avoid the secondary damage to the patient, an acceleration suppression module is included in the designed controller. The performance of the controller is also verified in the MATLAB/Simulink to show the effectiveness of the designed controller. In chapter 6, the design of the control system hardware is first introduced, and then the control software is designed based on the existing hardware. The performances of designed controllers are verified experimentally using the developed LERR system. The introduction of the hardware part mainly includes the control system hardware components and the electrical layout design of the control system. In the aspect of software part, the control system software components, motion control program, data acquisition and analysis processing, control algorithm program, coordinated control program and the human-machine interface are elaborated. Based on the established platform of software and hardware of the control system, the experiments of the position control, the impedance control and the designed gait trajectory adaptation control based on robust PID control are implemented, respectively. The feasibilities of the designed controllers used for the cooperative training way are verified and evaluated in the experiment. In chapter 7, the research conclusion of this thesis is summarized and the future research directions are suggested.In this thesis, the key issues on the rehabilitation robot for the cooperative gait training have been studied deliberately. These works can provide the key technology for developing the application-oriented Lower Extremity Rehabilitation Robot system to improve the quality of rehabilitation training process.