Cooperative Target Enclosing Control for Multiple Unmanned Surface Vehicles with Unknown Dynamics and Safety Assurance
Fanghao Huang,Xuanlin Chen,Zheng Chen
DOI: https://doi.org/10.1109/tiv.2024.3431949
IF: 8.2
2024-01-01
IEEE Transactions on Intelligent Vehicles
Abstract:Cooperative control of multiple unmanned surface vehicles (USVs) is essential for addressing increasingly complicated maritime tasks. This paper presents a safety-critical framework for cooperative target enclosing control of USVs, which considers the moving target with unavailable velocities, unknown dynamics of USVs, safety issues including the internal and external collision avoidance, and control input constraints. Specifically, a cooperative target estimator is designed that utilizes the inter-USV communication to estimate the position and velocity of a target with unknown velocities. Then, a cooperative target enclosing controller based on USV kinematics is designed to achieve both the distance keeping and evenly space circumnavigation with neighbors. To guarantee the safety issues for USVs with unknown dynamics during the target enclosing, a radial basis function neural network (RBFNN) based controllers combined with a safety-critical filter using input-to-state safe based control barrier function (ISSf-CBF) are developed. The RBFNNs are employed to approximate the unknown USV dynamics, and the network outputs serve as the basis for the safety-critical filter. This filter handles the impact of approximation errors and external disturbances on safety issues while achieving the collision avoidance, thus ensuring the safety and reliability during the target enclosing process. Comparative simulations and real platform experiments are implemented to verify the effectiveness of proposed safety-critical cooperative target enclosing control framework for USVs.