驾驶人驾驶决策机制遵循最小作用量原理

为提升智能汽车的自主决策能力，使其能够学习人的决策智慧以适应复杂多变的道路交通环境，需要揭示驾驶人决策机制。首先通过对自然驾驶数据的分析，发现在车辆行驶过程中能够反映驾驶人决策行为的主要运动特征参数存在极值现象，而产生极值现象的内在动因是驾驶人遵循“趋利避害”的基本决策机制，即驾驶过程中驾驶人力图实现机动性和安全性综合性能最优。受自然界包括物理和生物行为上的众多极值现象遵循最小作用量原理的启发，提出驾驶人决策机制遵循最小作用量原理的假设。随后建立抽象描述驾驶过程的物理模型，并提出最小作用量决策模型（Least Action Decision-making Model，LADM），通过与传统驾驶决策模型（经典跟车模型和换道模型）对比，分析结果显示LADM模型更具通用性。最后开展了实车试验，采集20名驾驶人在自由行驶、跟车行驶和邻车切入3种工况下的试验数据，分析计算并检验了不同驾驶人行车过程的理论最小作用量和实际作用量。试验结果表明：驾驶人在驾驶过程中的实际作用量与最小作用量之间无显著性差异，体现出驾驶人在行车过程中对安全和高效具有共性追求，验证了驾驶人决策机制遵循最小作用量原理。

Decision-making Mechanism of the Drivers Following the Principle of Least Action

It is necessary to determine the decision-making mechanism of drivers to improve the autonomous decision-making ability of intelligent vehicles so that they can learn human decision-making intelligence to adapt to complex and varied road traffic environments. In this paper, naturalistic driving data is analyzed. It is found that there are extreme value phenomena in the significant motion parameters, which can reflect the decision-making behavior of drivers during the driving process. The fundamental cause of the extreme value phenomena is that human drivers satisfy the primary decision-making mechanism of “achieving gains and avoiding losses.” Specifically, human drivers try to achieve the optimal overall performance of mobility and safety while driving. Inspired by the fact that several extreme phenomena occur in nature, including physical and biological behaviors, a hypothesis that the decision-making mechanism of human drivers follows the principle of least action was proposed. Then, a physical model that theoretically describes the driving process was established, and a least-action decision-making model (LADM) was developed. Compared with the traditional driving decision models (the classic car-following model and lane change model), the LADM model is generally more applicable. Finally, experimental field investigations were carried out to collect driving data from 20 drivers under three operating conditions, i.e., free driving, car-following, and cut-in scenarios. By analyzing the theoretical least action and actual action of different human drivers during the driving process, no significant difference was observed between the actual action and least action of human drivers. These results indicate that the human drivers typically seek both safety and high efficiency while driving.Furthermore, experimental results also demonstrate that the decision-making mechanism of human drivers follows the principle of least action.