Algorithm 4 Reward function for the vehicle agent (Code Listing 4: Reward Function)

1 REWARD (st, at){

2 vmin ← 5

3 vbrake ← 20

4 vmax ← 18

5 ɸthreshold ← 0.05

6 reward ← 0

7 if (vt > vbrake and acct = Brake){

8    reward ← reward + 1}

9 else if (vt > vbrake and acct = accelerate){

10    reward  ← reward − 1}

11 else if (vt > vmax and acct = Brake){

12    reward ← reward − 1}

13 else if (vt > vmax and acct = Accelerate){

14    reward ← reward + 1}

15 else if (ɸt < −ɸthreshold and steer t = TurnRight){

16     reward ← reward + 1}

17 else if (ɸt < −ɸthreshold and steer t = TurnLeft){

18    reward ← reward − 1}

19 else if (ɸt > −ɸthreshold and steer t = TurnLeft){

20    reward ← reward + 1}

21 else if (ɸt > ɸthreshold and steer t = TurnRight){

22    reward ← reward − 1}

23 else if (vt < vmin and acct ≠ Accelerate){

24    reward ← reward − 1

25 return REWARD

26 }