Agent-based Motion Control System Architecture Of Autonomous Underwater Vehicle

With the development of exploring and utilizing ocean source, Autonomous Underwater Vehicle (short for AUV) which could finish autonomous mission process is paid more and more attention. As an artificial intelligence system, AUV has high independence, reliability and adaptability to ocean environment. An efficient architecture of AUV play a significant roll in achieving those properties. Motion control system is one of the two subsystems for AUV. The research on constitution and realization of motion control architecture has significance.A kind of new motion control architecture is realized for a developed AUV called ZT-AUV. Conditioned by task demand, sensors and actuator configuration, the motion control architecture is constructed based on both behaviorism method and agent technology. The motion control architecture is divided into four parts including blackboard system, behavior element agent group, reflection behavior agent group and execution agent. The blackboard system is not only information processing and management center, but also agents'behavior control center. As the executable unit of motion controller, behavior element agent group makes AUV achieve three kinds of motion including surge, yaw and heave by certain control algorithm. Reflection behavior agent group is the unit by which the behavior of AUV can be achieved in another way, and it works when the system has fault. Execution agent finally drives the actuators of the system. The motion control architecture has good robustness, flexibility and expansibility. The knowledge and technology refer to the architecture is discussed in theory and engineering field.In the blackboard system part, a kind of data process is given aiming at the motion information of AUV. Simulations are performed to research the sensor and actuator fault diagnosis by sliding-mode observer. Strong Tracking Filter (short for STF) algorithm is improved according to data characteristic of under water sensors. The improved STF algorithm is utilized to realize data fusion of multi-sensor system of AUV.In the behavior element agent group part, intelligence algorithm used in motion control is discussed. An Immune Particle Swarm Optimization (short for IPSO) algorithm is utilized to design fuzzy neural network controller. A generalized predictive control method based on model of support vector machine is studied by simulation. In experiments, yaw controller performs bad at a high speed using traditional S surface control method, to solve this problem, a improved method is proposed based on speed compensation; and windup often occurs in the traditional S surface controller too. An anti-windup method based on motion state evaluation is proposed by adjusting integration term dynamically.In the reflection behavior agent group part, a fault tolerance control method based on force and moment redistribution is discussed as actuator fault occurs. Three fault execution schemes are proposed in different conditions of actuator fault.In the execution agent part, a force and moment distribution scheme is discussed. For the composite actuator configuration, different force and moment distribution schemes are discussed and formulas are obtained.