EI、Scopus 收录
中文核心期刊

基于可拓决策法的车辆自适应避撞控制方法研究

EXTENSION-DECISION-BASED ADAPTIVE COLLISION AVOIDANCE CONTROL FOR VEHICLES

  • 摘要: 车辆的避撞控制可以有效避免或缓解车辆的碰撞事故, 是自动驾驶汽车的关键控制技术之一. 各种交通条件、不确定的道路附着系数以及复杂的液压制动执行系统都会降低避撞控制的有效性. 因此, 本文提出了一种基于可拓决策法的自适应避撞控制, 该控制方法对路面附着系数具有自适应性, 并能够精确控制制动系统的制动液压力. 首先, 设计了滑模观测器来估计轮胎纵向力, 并基于观测得到的轮胎纵向力, 进一步提出带遗忘因子的递推最小二乘法估计道路附着系数. 其次, 基于递推最小二乘法的估计值, 提出了基于路面附着系数自适应调节的自适应避撞控制方法, 该方法基于可拓决策法的先决判定决定当前时刻应采用何种避撞控制策略, 即采用可拓决策方法判断进行点刹预警制动、全制动或者不制动. 再次, 通过对执行系统-电控液压制动系统进行精确的液压控制, 实现主动减速控制效果. 最后, 采用软件联合仿真手段对上述方法进行了验证, 结果表明所提出的算法在避撞行驶工况中具有良好的避撞效果.

     

    Abstract: Vehicle collision avoidance control can effectively avoid or mitigate vehicle collision accidents, which is one of the key control technologies of autonomous vehicles. Various traffic conditions, uncertain road adhesion coefficient and complex hydraulic brake execution system will reduce the effectiveness of collision avoidance control. Therefore, an adaptive collision avoidance control (ACAC) based on extension decision method is proposed in this paper. The control method is adaptive to the road adhesion coefficient and can accurately control the brake fluid pressure of the braking system. Firstly, the tire longitudinal force is estimated by the sliding mode observer (SMO), The designed SMO uses only the signal measured by sensors on board, i. e. the wheel angular speed to estimate the longitudinal tire force. Based on the observed tire longitudinal force, the road adhesion coefficient is estimated by the recursive least square method with forgetting factor (FFRL). Secondly, based on the estimated value of the road adhesion coefficient by FFRL, an ACAC method based on adaptive adjustment of road adhesion coefficient is proposed. This method determines which collision avoidance control strategy is adopted, that is, the extension decision method is used to judge whether short point braking, full braking or no braking. The impending collision time is defined as the main characteristic quantity, and the collision danger distance is the auxiliary characteristic quantity. A two-dimensional extension set is established based on the motion relationship between the ego vehicle and preceding vehicle and it can be divided into three different domains: the classical domain, the extension domain, and the non-domain. Thirdly, the effect of active deceleration control (ADC) is realized through precise hydraulic control by the executive system of electronic hydraulic braking system. Finally, the above methods are verified by manner of software joint simulation. The results show that the proposed algorithm has good collision avoidance effect in collision avoidance driving conditions.

     

/

返回文章
返回