Algorithm 1 Perception and Decision-Making Process

Require: Sensor data I from cameras, LiDAR, radar, etc. Ensure: Detected objects and driving decisions   1:Input: Sensor data I   2:Output: Bounding boxes B, object classes C, confidence scores S, and action $a_t$   3:Step 1: Object Detection   4:$B,C,S\leftarrow M\left(I\right)$           ▹ Apply selected CNN/Transformer model M   5:Step 2: Decision Making   6:Initialize state $s_0$ based on $B,C,S$   7:Define MDP components $(S,A,T,R)$   8:Compute optimal policy ${\pi }^{*}=arg{max}_{\pi }\mathbb{E}\left[{\sum }_{t=0}^{\infty }{\gamma }^{t}R(s_t,a_t)\mid \pi \right]$   9:Select action $a_t$ based on ${\pi }^{*}\left(s_t\right)$ 10:Return:$B,C,S,a_t$