Two-Loop Acceleration Autopilot Design and Analysis Based on TD3 Strategy

Table 1

Dimensions of each layer of the actor- and critic-networks.

Layer	Actor-network	Critic-network
Input layer	3 (state dimension)	5 (state dimension+action dimension)
Hidden layer 1	256	256
Hidden layer 2	256	256
Output layer	2 (action dimension)	1 (action value function dimension)