驾驶员意图与行驶环境的统一识别及实时算法

INTEGRATED INFERENCE OF DRIVER'S INTENTIONS AND DRIVING ENVIRONMENT AND REAL-TIME ALGORITHM

针对以往意图识别存在的问题,根据车辆运行过程中的动力学特性,定义了平直道路匀速运动的平衡节气门开度和相对节气门开度,并在此基础上分析了驾驶员期望加速特性,给出了相互关联的五个分类以及各个分类下的换挡操作特征。分析表明,行驶环境在换挡过程中可以等效为驾驶员期望加速特性,从而提出了驾驶员意图与行驶环境的统一识别方法。根据相应的实车试验数据和经验收集,制定了模糊推理规则库,建立了多输入单输出的模糊推理模型,并将该模型移植到单片机控制系统中,实现了驾驶员意图识别的实时算法。经过对实车试验数据进行的识别验证,该算法能够比较准确完整地识别驾驶员意图。

Referring former methods of driver＇s intention inference, according to vehicle dynamics, the balance throttle open on plane and straight road and relative throttle open are defined. Based on these, driver demand acceleration characteristics are discussed and driver intentions on plane and straight road are classified to five relative styles. Driver usual operation characteristics under these five classic conditions are summarized. It is proved that driving environment is equivalent to certain driver intention and they can be integrated to infer. A multi-input single-output fuzzy reasoning model is established and the fuzzy rules are based on test data and driver experiences. This model can be applied to micro controller unit and an improved real-time algorithm is brought up. Validated with the test data, this model can infer the intentions quite well.