Path planning algorithms for unmanned surface vehicle formation in maritime environment

The research into unmanned surface vehicles (USVs) has received increasing attention in recent years due to the maturity of the technology. Potential deployments of USVs can been seen through both civilian and military applications with the benefits of improved mission efficiency and decreased resource costs. However, it should be noted that current USV platforms are generally of small size with low payload capacity and short endurance. To improve the effectiveness, there is a trend to deploy multiple USVs as a formation fleet. This thesis therefore primarily investigates the path planning problem of USV formation. Overall, algorithms have been developed based on the leader-follower control strategy and adopt the fast marching method (FMM) as the base algorithm. To make the algorithm more suitable for practical maritime navigation, the FMM has been first improved and redeveloped by considering the dynamic characteristics of the USV and making the generated path compliant with the USV's turning constraints. Next, to solve the problem of avoiding moving obstacles in the environment, a constraint fast marching method (CFMM) has been proposed to model the dynamic behaviour of moving ships. The CFMM generates effective ship domain and collision avoidance areas of moving ships according to different velocities so protecting the USV formation from collision. The uncertainties associated with the maritime environment have also been investigated, and a Kalman filter based trajectory tracking algorithm (KFTTA) has been designed and developed to obtain the accurate navigation information of moving ships. The KFTTA can be integrated with the formation path planning algorithm to improve its effectiveness and efficiency. All algorithms have been tested and verified using computer based simulations. In addition, full scale experiments on a practical USV have also been carried out to test capability in water.