Underwater Vehicle’s Intergrated Positioning Design Based On Vision And Inertial System

Recently, sonar, long baseline, doppler velocitylogger and other related acoustic equipment are widely used among the underwater vehicles (ROV), however such usage is limited by expensive cost, vulnerability in signal transition and blocked application interface, which cannot be applied in small range with high precise positioning demand. Another widely used navigation system, inertial measurement unit-IMU, is able to acquire vehicles’attitude and position through integrating its angle rates and accelerations, but integration will accumulate error along the time.Based on the ROV’s positioning technology research above, the integrated positioning system was promoted, combing camera electronica compass and IMU with Kalman optimal algorithm. With this strategy the vision positioning system would fix the drift conducted from inertial system and provide a small range with higher precise positioning.The whole article consists of six chapters as follows,The first chapter introduces the current scenario of ROV’s positioning system around the world. Based on the advantages and disadvantages analysis of different kinds positioning system, meaning and content within the research are investigated.The second chapter describes architecture of integrated positioning system.1. Inertial positioning module, which is applied with strapdown inertial navigation algorithm as the fundamental positioning method;2. Aided position module, including vision positioning in displacement fusion and compass in attitude fusion, meanwhile the vision algorithm and compass’s interference are analyzed;3. Pan-tilt Auto track system, the reverse kinetics among the tracking process and control strategy of motors are analyzed.The third chapter focuses on solving the low precise of the IMU, especially the random noise along the output. Time serial method is investigated to build the output model and the AR(3) model is verified.The fourth chapter demonstrates the multi-sensor fusion algorithm based on Kalman Filter. Two kinds of integrated position, displacement and attitude, are modeled in both system states equation and measurement equation.The fifth chapter describes the test organization, including the ROV body, Polaris, the platform, control circuit based on AVR MCU and the fusion process in MATLAB. With the comparison in the test, it is proved that the integrated positioning system based on vision, compass and IMU is effective in improving the positioning precise.