Research On ROV High Precision Control Method Based On Multi-information Fusion

With the development of marine resources,many underwater equipment is also changing,as an important tool of marine development-remote control shallow water robot(ROV),because of its many advantages are widely used.However,due to the complex environment underwater,the control of ROV has always been the focus and difficulty.At present,many control methods are applied to the control of underwater robots,but the most commonly used common closed-loop PID(Proportion Integration Differentiation)control method can not meet the high-precision control needs.Therefore,the high-precision control of underwater robots is studied.Through the research of the track-type underwater robot designed and produced,the structural design and control logic of the underwater robot are introduced,the dual coordinate system system is set up for the underwater robot and the space motion equation is established for the underwater robot according to its underwater force,the transfer function of the underwater robot is further obtained,and the sensor data is fused by the improved Kalman filter to obtain more accurate information on the state of the underwater robot.The state information is used as the input data of the improved BP neural network to make continuous prediction of the underwater robot state,and the prediction data is constantly corrected as the feedback information of PID controller,so as to obtain the high-precision control system of multi-information fusion.The track and wireless control of track-type underwater robot are experimented,and the stability and reliability of track and wireless control are verified to meet the design requirements.Under the condition of pool,several groups of experiments were carried out on depth and hem direction control.Through the analysis of the data,it can be seen that: First,the control real-time and stability of underwater robots are related to the prediction time of BP neural network,and secondly,in depth control and heteronormation control,the prediction time is different,in depth control,Predictive time selection of 0.3s can make the control stability and real-time to achieve the best state,in the direction control,the prediction time selection of 0.2s can make the control stability and real-time to achieve the best state;It is proved that the high-precision control method of multi-information fusion studied has significant advantages.