Adaptive agent tracking approach for oil contamination in water environments

Abstract

The online monitoring of water environments is urgently needed. A feasible and effective approach is the use of agents. Water environments, similar to other real-world environments, present many changing and unpredictable situations. To ensure flexibility in such an environment, agents should be prepared to deal with various situations. In this study, we focused on an adaptive agent tracking approach for oil contamination. An integrated tracking framework, which is used to track the moving contour of oil pollution via a system comprising multiple unmanned surface vehicles, is proposed. The zigzag, unmanned underwater vehicle-gas, cloverleaf trajectory and curvature-weighted deployment algorithm methods are employed with consideration of their suitability to our approach. A cyclic particle swarm optimisation–Kalman method is also proposed. The possible position of moving vertices is predicted by the Kalman filter, and an objective search region is generated around the centre position. Moreover, particle swarm optimisation is performed to search for the best target position in this region. This particle swarm optimisation–Kalman method is circle operated to compensate for the deficiency of a few agents. To evaluate the approach, we conduct usability and performance simulations.