| Human exploration of the robot has more than 3,000 years of history and has never changed,however,the production of the world's first industrial robot has so far only 60 years.With nearly 60 years of rapid development,making robots from scratch to everywhere.This has become a true portrayal of today's life and brought unprecedented improvements to the national economy and people's lives.For a long time,numerous scientific researchers have been eager to realize the close contact between robots and humans.We hope that robots can help us in our daily life and work with us.Traditional manipulator mostly perform the position control in a fixed configuration and the joints work in a single active mode,which has been successfully operated in a controlled environment such as a factory production line.Outside the controlled environment,traditional manipulator can perform the complex operational tasks only when operated by an operator.Up to now,the robot has gone beyond the structured factory environment and entered a complex environment to achieve human-computer collaboration.Traditional industrial manipulator with position control is not suitable for complex operation in human environment.Therefore,the robots in contact with the environment face new opportunities and new challenges as well.The rapid development of robotics in these few decades relies mainly on the rapid development of other disciplines,which are mutually reinforcing with the development of robotics.For example,bionic science,artificial intelligence,computer science and other fields,the robot system can provide an experimental platform for its theoretical verification,and robotics can not be separated from the support of these disciplines.Since the concept of modular reconfigurable manipulator has been proposed,a large number of researchers in related fields have devoted themselves to the research of position control and have made great achievements.However,in practical engineering applications,such as opening and closing operations,search and rescue and assisted surgery and rehabilitation training case,manipulator working in the complex and precise operation by environmental constraints,single position control can not satisfy the need of practical engineering application,more researchers locked the control target on its position control while achieving accurate control of its end force.Due to the increasing investment in reconfigurable module robots and even working directly with people,their life expectancy and reliability requirements continue to increase.Referring to all of theavailable literature concerned,the study on position/force fault tolerant control problem of the constrained modular and reconfigurable robot is still rare at home and abroad.In addition,modular and reconfigurable robot(MRR)may need transform its configuration to adapt to different mission requirements during the mission.Therefore,the decentralized control strategy is more suitable for the control problem of the MRR's modular design ideas.However,the model decomposition,namely how to use only local joint information to map the endpoint contact force to each joint,is always a bottleneck which limits the application of decentralized control strategy to the position/force control problem of the constrained MRR.In this paper,the torque sensorless robust fault-tolerant control method of reconfigurable manipulator system is deeply investigated.This paper mainly studies the model description of constrained reconfigurable manipulator with and without actuator fault,the torque sensorless force/position decentralized control method,the force/position robust decentralized fault-tolerant control method,the system health monitoring method with fault-tolerant controller,and the fault tolerance control method with consideration of actuator saturation.The organization of this thesis is as follows,in which the main works are included.1.The background and significance of the research are expounded,and the research status of modular reconfigurable manipulator and the technology of fault tolerant control are reviewed.2.The Newton-Euler iterative algorithm is used to get the traditional dynamic model and fault model of modular reconfigurable manipulator,and it is characterized as a set of subsystems set up by coupling torque.In addition,a joint torque estimation method based on harmonic drive model is presented.By using this method,a relatively simple model expression method is obtained compared with the traditional model.3.A joint torque estimation method based on the harmonic drive model is proposed for the subsystems dynamics model of the constrained reconfigurable manipulator,which based on the motor-side and link-side position measurements along with harmonic drive model.Combining the estimated joint torque and dynamic output feedback technique,a decentralized position/force control strategy is presented.In addition,the non-fragile robust control technique and dynamic output feedback are combined to further improve the control performance of the system,considering the perturbation of the controller parameters.4.A robust decentralized fault-tolerant control strategy is developed forconstrained reconfigurable manipulator system with non-affine actuator fault.First,a modified sliding mode controller is selected for constrained reconfigurable manipulators fault-tolerant control,which can guarantee force/position tracking performance.Then,the unknown non-affine actuator fault is compensated by adding an adaptive compensation term in independent subsystem.Finally,simulations are performed to verify the advantages of the proposed method.5.Considering the motor output limit,a novel distributed module system operation health monitoring scheme for modular and reconfigurable robots is proposed,assuming that the modular and reconfigurable robots system is controlled by a distributed robust fault tolerant controller.In order to avoid measuring or estimating the joint acceleration of the modular and reconfigurable robots,the command torque is filtered and compared with a filtered torque estimate,based on which a performance index that reflects the operation health of an MRR module is developed.6.A novel decomposition-based distributed robust fault tolerant control method is proposed for modular and reconfigurable robots with consideration of actuator saturation.The actuator saturation is avoided and the original task is completed in the recovery stage by introducing a novel radial basis function neural network-based compensation scheme.The RBF is only used to compensate actuator saturation,and has a simple architecture and a fast training process.Any delay caused by radial basis function does not affect the fault tolerant control action.In the final,the conclusion and the perspective of future research are given at the end of this thesis. |
PDF Full Text Request