Design Of Remote Monitoring System For Unmanned Surface Vessel Based On Data-Driven

Unmanned surface vessel(USV)is one kind of surface warship that is an unmanned and intelligent platform on the water.It has great performance in stealth,flexible and automatic navigation that can finish a multitude of extremely dangerous missions that human can't complete.It plays an irreplaceable role in contemporary ocean war and marine resources exploration.With the continuous improvement in the intelligence of USV,the new and higher-level requirements have also been put forward for remote control platform.The remote control platform has the highest level of control over the USV.The operator can monitor the state of the USV through the remote control platform and control the navigation state of the USV.Since USV usually work in the case of over-the-horizon,it can only judge its working status and navigational attitudes based on telemetry data.In particular,it is critical to identify features of state in telemetry data to identify abnormal states and abnormal trends.Firstly,this paper describes the relationship between the upper and lower computers of the remote control platform and the unmanned craft,and emphatically introduces the control system of the HEU-2 unmanned craft.In order to enable the unmanned boat to pass its own steering angle control system,the desired course sent by the remote control platform is completed.Analyzing and deducing the operational response equation of unmanned surface vessel: Nomoto model.The unknown parameters in the model were identified by the actual ship test.Finally,the PID autopilot was selected as the unmanned boat heading control,and the rationality of the control algorithm was verified through simulation.Secondly,in the full analysis of USV control characteristics and functional requirements,clear the design goals.Based on the remote control platform software system developed by MFC,the function of each sub-module was introduced in detail.At the same time,a robust and efficient communication protocol was designed.The format and content of the message were elaborated,and the remote-operation-based platform was introduced.The USV's operation flow.Finally,the hardware part of the remote control platform was designed based on the idea of modularization.Thirdly,the basic principle of nonlinear classification of SVM and the solution strategies of common multi-class classification problems is introduced,especially the construction strategy of multi-classifier based on SVM.In order to make multi-classifier achieve good classification effect,the particle swarm optimization algorithm(PSO)is used.The basic principles and operation steps of SVM attribute parameter optimization lay a theoretical foundation for the subsequent construction of fault diagnosis model.Finally,the unmanned state data features were excavated and the USV faults diagnosis model was constructed.Firstly,the telemetry data obtained by the experiment is preprocessed and filtered.The Chebyshev fitting algorithm is used to fit various states and the feature vector is used to replace the original data.The Binary Tree is used to construct the USV fault diagnosis model and the PSO algorithm is used.The SVM performs parameter optimization and verifies the accuracy of the diagnostic model through classification accuracy.Use online diagnostics to verify the accuracy of the model.