多传感器融合下质心侧偏角和路面附着系数估计

Multi-Sensor Fusion Based Side Slip Angle and Road Friction Estimation

智能驾驶的发展将由驾驶员承担的行车安全责任转移到自动驾驶系统上,因此,智能车的安全运行越来越重要。质心侧偏角和路面附着系数是智能车辆安全规划和控制的重要参数。本文融合摄像头和惯性测量单元(IMU)信息,提出了不依赖于轮胎信息的质心侧偏角和路面附着系数估计框架。首先,基于车辆运动学模型,利用摄像头和IMU信息,运用卡尔曼滤波估计智能车辆质心侧偏角;其次,于车辆动力学模型,利用IMU信息,运用改进的扰动观测器估计车辆前后轴侧向力;最后,基于估计的质心侧偏角和轮胎侧向力,融合摄像头识别的路面状态,利用高斯牛顿法,鲁棒估计了路面附着系数。仿真分析显示,本文提出的估计框架可以精确地估计质心侧偏角和路面附着系数。

The development of intelligent vehicles transfers the responsibility of driving safety from driver to the automatic driving system,leading to the increase of importance for intelligent vehicle driving safety.Side slip Angle and road adhesion coefficient are very important for intelligent vehicle's safe planning and control.In this paper,a framework for estimating side slip angle and road friction coefficient is proposed by fusing camera and inertial measurement Unit(IMU)information.Firstly,based on the vehicle kinematics model,Kalman Filter is utilized to estimate the side slip angle with camera and IMU information.Secondly,based on the vehicle dynamics model,an improved disturbance observer is implemented to estimate the lateral force of front and rear axles using IMU information.Finally,based on the estimated side slip angle and tire lateral force,the road friction coefficient is estimated by using Gaussian-Newton method with road condition from camera as constraints.Simulation analysis shows that the proposed estimation framework can accurately estimate the sideslip angle and road friction coefficient.