Reinforcement Learning-Based Autonomous Navigation and Obstacle Avoidance for USVs under Partially Observable Conditions

Table 1

The overall network structure of the UANOA algorithm, which consists of a perception part and a decision part.

Network structure	Network type	Number of neurons	Activation function
First layer	Linear	256	Relu
Second layer	Linear	256	Relu
Third layer	Linear (concat)	131	Relu
Fifth layer	GRU	128	Tanh
Sixth layer	Linear	128	Relu
Seventh layer	Linear	128	Relu
Eighth layer	Linear	13	Softmax