Localization and Mapping for Deep-Sea Mining Vehicles With Forward-Looking Sonar

Abstract

Abstract The deep-sea mining vehicle is one of the critical equipment of the deep-sea mining system, which is used to collect manganese nodules or other minerals on the seafloor. When the deep-sea mining vehicles works, a reliable localization system is essential. In addition, the ability to sense and model the surrounding environment in real time is also necessary for the safe and efficient operation for deep-sea mining vehicles. This study developed and validated a localization and mapping algorithm. The localization algorithm uses extended Kalman filtering fusing inertial measurement unit, ultrashort baseline, a compass, track encoders and forward-looking image sonar. The mapping algorithm relies on forward-looking image sonar and the position output by the localization algorithm. The environment map is a grid map to meet the requirements of obstacle avoidance and path planning for deep-sea mining vehicles. A deep-sea mining vehicle named ‘“Pioneer I” which was designed and developed by Shanghai Jiao Tong University has been tested in the South China Sea. All the data of sensors in the sea trials were recorded with time synchronization based on the software framework. Therefore, the sensor data from the sea trials can be used as input for validation of the proposed localization and mapping algorithm.