Adaptive navigation control of a bionic robotic fish in complex Karman vortex street flow fields using an LSTM-DDPG hybrid strategy

Changhui Zheng, Peigang Jiao, Honghao Xu, Yiheng Zhang, Jiaxin Shi

Bionic robotic fish suffer from poor navigation robustness, high energy consumption, and control lag in unsteady Karman vortex street flow fields. This study proposes a hybrid adaptive navigation strategy combining long short-term memory (LSTM) and deep deterministic policy gradient (DDPG) to achieve autonomous, efficient, and stable motion control. The method constructs a closed-loop framework of local flow field perception, temporal prediction, and continuous flexible control, eliminating dependence on pre-built flow field models. A Karman vortex street simulation platform is developed using the immersed boundary-lattice Boltzmann method (IB-LBM), and a multi-level reward function is designed to balance accuracy, stability, and energy efficiency. Numerical simulations and physical prototype experiments are conducted under Reynolds numbers Re = 500, 800, and 1000, with comparisons to PID, DQN, and MPC. Results show that the LSTM-DDPG strategy significantly improves navigation precision and anti-disturbance ability while reducing energy consumption. The average task completion rate reaches 88.3%, and average energy consumption is 5.2 J/m, which is 45.3% lower than conventional PID control. This method provides a feasible solution for robust and energy-efficient navigation of bionic robotic fish in complex ocean environments.