Research On Control Strategy Of Manipulator Based On Event-triggered Mechanism

The robot industry has developed rapidly in recent years,and it has played an increasingly important role in industrial production and social life.The manipulator is a complex controlled object with multiple inputs and multiple outputs,strong nonlinearity and strong coupling,and has strong model uncertainty,so its control problem has also become a hot spot.The traditional periodic sampling control will waste a lot of system resources,such as network bandwidth,node energy and computing resources.On the premise of ensuring system performance,the event trigger mechanism can adjust the sampling interval in real time according to the state of the controlled object for on-demand control,which effectively reduces the consumption of system resources.Therefore,this paper studies the control problem of the manipulator based on the event-triggered mechanism.The main contents include:The mechanical arm system is modeled based on the Euler-lagrange equation,and an event triggering tracking control strategy is proposed for unknown external disturbance.Trigger the sliding mode control law and its trigger conditions by designing the corresponding event to ensure the anti-interference ability of the system.The switching gain of the control law is designed to be related to the system status,ensuring the desired tracking performance of the system.The Lyapunov stability theory deduction controller generates sufficient conditions for the desired tracking performance,and calculates a non-zero down boundaries of the event interval to ensure that the proposed event triggering conditions do not produce Zeno phenomena.The effectiveness of the proposed control method is verified by the simulation of the two-link robot arm.Considering the model uncertainty of the robotic arm system,a neural network control strategy based on adaptive event trigger mechanism is proposed.Under this trigger mechanism,the controller uses a neural network to process the number of events needed to approximate the neural network by designing the conditions triggered by the corresponding adaptive event,to reduce the calculation resource consumption.Based on the traditional neural network controller,this article also proposes a new weight update for non-cycle adjustments to neural network weights on the trigger time.In order to analyze the stability of the system,the event trigger system is modeled as a nonlinear pulse dynamic system,and the Lyapunov theory is used to prove the local final boundaries of the system.The event triggered condition is related to the system status and event trigger error,in order to avoid unnecessary events in the final boundary,use the dead zone operator to reset the event trigger error to zero.Then,by calculating the non-zero down boundary of the event interval,the proposed adaptive event trigger condition does not generate Zeno phenomena.Simulation studies verify the effectiveness of the proposed control method.For distributed multi-robot arm systems,a distributed event trigger control policy is proposed for communication between the robot arm,and the local controller update of the single robot arm.A distributed consistency control protocol is proposed,and then the distributed event trigger function of the multi-robot arm system is designed based on system status information.It is also given sufficient conditions that can ensure the trigger consistency of the multi-robot arm system event,and it demonstrates that the event trigger time sequence does not occur in Zeno.The effectiveness of the proposed control method is verified by the system composed of multiple two-link robotic arm.