Objectives  This paper proposes a real-time obstacle avoidance method based on artificial potential field, in order to cope with the many potential obstacles in complex underwater environment which affects the movement plan for multi-autonomous underwater vehicle(AUV)cluster.

  Methods  Firstly, a formation method based on dynamic network topology is adopted, and the AUV is regarded as a node in the network. The potential field function is set to meet the formation requirements. Then, based on the artificial potential field method, the potential field function is established for the region where both targets and obstacles exist simultaneously. Afterwards, the potential field function is upgraded to an exponential function, such that the AUVs can be planned online for real-time accomplishment of the mission of obstacle avoidance for multi-AUV cluster. Finally, 10 AUVs and 6 obstacles are simulated with Matlab software.

  Results  The simulation results show that with this method, each AUV can successfully avoid obstacles and reach the safe area at the target point.

  Conclusions  The artificial potential field function method may enable the multi-AUV to accurately avoid obstacle in real time. The advancement of this technology has important and positive significance for improving military operational capability.