非线性传感器的融合在多小车平台中的应用  

Application of Nonlinear Sensor Fusion for Multi-Vehicle Teams

作　　者：张菡 李大卫 郑卫刚[2] Zhang Han;Li Dawei;Zheng Weigang

机构地区：[1]墨尔本大学电气和电子工程系 [2]武汉理工大学工程训练中心

出　　处：《变频器世界》2019年第11期65-70,共6页The World of Inverters

摘　　要：使用一种价格低廉的光强传感器来定位多小车组中各个成员的相对位置。根据光强传感器的读数与光源的距离和角度之间的变化关系,为高度非线性的传感器建立合适的含参量数学模型。采用MOTION CAPTURE系统获取多组传感器与光源处于不同距离、不同角度时的光强数据。对实验数据处理之后,采用最小二乘法的方式拟合得出最终的传感器模型。建立小车运动的空间状态方程,基于传感器的数学模型结合粒子滤波算法设计出状态估计器,来估计多小车平台中,单个小车相对于周围小车的距离、方位角以及小车的运动朝向。结果表明,所设计的粒子估计器可以有效地解决廉价传感器读数漂移和受噪声干扰的问题,从而准确地估计出系统中小车的状态。Developing a novel low-cost sensor system in the multi-vehicle teams.Vehicles inside the teams could communicate and sensor the relative position about one another simultaneously.Establishing a parametric sensor model which could divulges the relationship between sensor measurement and the orientation of light source.MOTION CAPTURE system was used to obtain large number of trajectories in different distance and orientation angle.Applying least square algorithm to fit the sensor model.The state space model of the vehicle was established to design an effective estimator.In this essay,the particle filter was implemented to estimate the relative distance and orientation angle between the vehicle members as well as the heading angle of each vehicle.It could be revealed that the particle filter which is designed in this essay could solve the sensor drift and noise perturbation problem efficiently.

关 键 词：传感器融合 状态估计器 非线性系统 小车定位 群蜂行为 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]