.png){kind=icon}
.png){kind=icon}
Return
Path planning based on improved deep Q-network algorithm for bionic robotic fish with ocean currents
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2025.131173.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Dynamic integrated reward mechanism. In order to solve the multi-objective optimization problem for the BRF path planning, a dynamic integrated reward mechanism is presented to improve the optimization performance in the environment with ocean currents. • Dynamic two-step action-selection strategy. This strategy is proposed to adjust the balance between exploration and exploitation, which can not only guarantee the agent full exploration of the complex environment in the early training stage, but also make explicit exploitation of high-value actions in the later training stage to improve the learning efficiency and convergence speed. • Double level dynamic learning rate. The double-level dynamic learning rate is proposed in this paper, which mitigates the overfitting, improves generalization ability and enhances robustness for the model. This is helpful to obtain the optimal BRF path in different unknown environments.
Keywords:
BRF path planning
dynamic integrated reward mechanism
two-step action-selection strategy
double-level dynamic learning rate
multi-objective optimization
Journal
IF:
6.5
Papers:
2.5W
Citations:
6.5W