论文部分内容阅读
针对车辆质心侧偏角难以精确测量和不可观测性等问题,提出以可观测数据为基础,采用相关向量机对车辆质心侧偏角进行估计。首先,针对非线性车辆模型建立运动微分方程,构建可观测变量横摆角速度、方向盘转角以及车速与质心侧偏角之间的映射关系;其次,利用可观测变量的观测值以及映射矩阵完成对相关向量机模型的训练,通过边缘似然数求解确定模型超参,利用大量实车样本对模型进行测试;最后,采用Carsim软件进行仿真分析,并对横摆角速度、方向盘转角与侧偏角之间的关系进行分析,结果表明:在质心侧偏角出现尖峰处估计值与仿真值存在一定差异性,估计值整体上能够很好的逼近仿真值;横摆角速度、方向盘转角与质心侧偏角之间存在相似的变化规律。
In view of the difficulty of accurately measuring the side slip angle of the vehicle center and the unobservable nature of the vehicle center, this paper proposes to estimate the center of mass of the vehicle using the correlation vector machine based on observable data. Firstly, a differential equation of motion is established for the nonlinear vehicle model, and the relationship between the observable variable yaw rate, the steering wheel angle and the vehicle speed and the center of mass roll angle is established. Secondly, the correlation between the observed variables and the mapping matrix The training of vector machine model, through the edge likelihood analysis to determine the model hyper parameters, the use of a large number of real vehicle samples to test the model; Finally, the use of Carsim software for simulation analysis, and the yaw rate, steering wheel angle and slip angle The results show that there is some discrepancy between the estimated value and the simulation value of the spikes at the center of mass, and the estimated values can approximate the simulation value as a whole. The yaw rate, steering wheel angle and center of mass There is a similar pattern of change.