Robust Control On Mining Moving Robot For Deep Seabed Mining

Exploration technique on deep seabed is a kind of complicated high technology and a new important project which follows aerospace exploration faced in the world. It has important strategic significance for our country's mineral resource continual utilization and development.Deep seabed mining robot works in 6000m-depth seabed unknown complex environment with very dilute and soft bottom, no natural light, high pressure, and strong disturbance. Moreover, the special design of mining robot leads to its strong nonlinearity. Deep seabed mining system requires mining robot works efficiently following desired path with certain accuracy in unknown complex environment. As the carrier of deep seabed mining system, the control quality of mining robot has direct influence on our countries ocean exploitation strategy. The mining robot is operating in the deep-sea sediment, which shows itself weak shear intensity and varies randomly. Compared with the similar type of tracked-vehicle on land, it is easy to slip seriously, even easy to sink into the sediment and lose its ability to run. However, the robot's special design and special working environment cause the complexity of control problem. Therefore, the research on control of mining robot for deep seabed mining is an important problem among many challenges of deep seabed mining technique. It is high time to research how to control the slip to safeguard the running of the robot. The research has great practical significance and great scientific significance.The dissertation mainly studies the robust control problem of mining robot for deep seabed mining aims to providing some previous theoretical research and technical support for deep seabed mining.The major innovation research achievements include:(1) With the analyzing mechanisms of deep seabed mining robot deeply, considering mining robot belongs to a kind of tracked vehicle, analyzing robot's walking characteristics, using differential steering, the mining robot dynamics model is established. Mathematical model of electro-hydraulic proportional valve, plunger-type variable displacement pump, hydraulic motor are built. The motion system of deep seabed mining robot is described effectively.Influential factors of mining system motion states and dynamic characteristics are analyzed. Dynamic equations of soft pipe and rigid pipe are established, fluid force analysis on soft pipe is performed, force on soft pipe end is given, and the mining robot yaw control mathematical model is established.(2) On the basis of deep seabed mining robot localization system functional requirements, the localization system frame is proposed. Aiming at complex localization problem, an integrated location method based on pseudo long baseline and dead reckoning is proposed, and the Kalman filter structure is established. To improve precision, adaptive Kalman filtering algorithm is applied for data fusion and amending filter gain. The feasibility of this method is tested by simulation research results, estimation accuracy has a great enhancement comparing with existing method. A new location approach is put forward.(3) On the Analysis of the principle of anti-skid control of the track reveals the relationship of the slip ratio and adhesion coefficient, the optimal slip ratio of deep-sea sediment is gained according to the ratio and the adhesion coefficient curve, an anti-slip controller is developed.(4) Deep seabed mining robot works with high slip and great uncertain moving state change, the key moving parameters are difficult to measure. Path tracking control approach is studied based on mechanical analysis and hydraulic driving principle analysis. The parameters including tracks slip rate and effective radiuses of driving wheels identification model is built, it builds basis for optimal unbiased estimation of deep seabed mining robot key moving parameters. Controller is designed based on efficient key moving parameters estimation, position error and yaw angle, simulation experiment results show the effectiveness of the method.(5) Considering the uncertainty of mining robot control model, structure parameter and external environment, combining robust control with neural network, path tracking adaptive robust control approach is concerned. RBF neural network is used to learn the unknown bounds of system uncertainties adaptively, adaptive heading robust control algorithm is presented based on status feedback and virtual input. And control method on the crawling of deep seabed mining moving robot is described after discussion on fuzzy rule between virtual input and tracks velocity. The feasibility of this method is tested by simulation research results.(6) Experiments have been made on seabed mining robot vehicle model simulation system. Velocity changing curves, yaw angle changing curves, yaw angle error curves and path tracking curves test the feasibility and validity of mining robot control system.