前言 A high-level decision agent trained by deep reinforcement learning (DRL) performs quantitative interpretation of behavioral planning performed in an autonomous driving (AD) highway simulation. T