Algorithm 1 Award accumulated in each IA’s observation-action cycle

for i = 1:20 if ray[i].hit = true Add reward(−0.5/ray[i].distance); end end if velocity ≤ 0 AddReward(−0.1 · f); if velocity > 0 AddReward(2 · f); if 2 < velocity < 12 AddReward(velocity · f); if 12 < velocity ≤ 27 AddReward(0.1 · f); AddReward(−0.1 · acceleration); AddReward(−0.01 · steer · steer);