Real Time Path Predicting for Autonomous Underwater Vehicle Using Support Vector Regression Machines

Autonomous underwater vehicles (AUV) are unmanned underwater vessels to investigate sea environments, oceanography and deep-sea resources autonomously. Navigation of underwater vehicles is a very demanding task, especially in confined environment. In order to avoid different risk, the path information must be communicated between sensors and monitor immediately, but the data transfer rate are very low in the sub-sea acoustic communication channel, it is impossible to receive these information immediately, so a model based on support vector regression machines for real time path predicting is proposed in this paper. This model can update according the new data form the sensors. Experiments show this method can have a good performance to predict the path of AUV.