【作者】 孙海荣；

【导师】 杨鹏；

【作者基本信息】 河北工业大学 ， 控制科学与工程， 2011， 硕士

【摘要】 当前移动机器人应用的研究重点之一是多传感器信息融合理论与目标定位技术相结合的方法,多传感器信息融合技术能够综合多个传感器提供的各个侧面信息,用以提高目标定位的精度,其定位性能更全面、更准确。本文提出了一种抗野值的多次卡尔曼滤波的机动目标自适应跟踪方法。当测量值中存在野值时该方法自适应地修正传感器测量值,以达到修正Kalman滤波新息正交性的效果,并且根据目标的运动状态,调整观测噪声协方差的估计值,使其更符合实际的观测情况,同时解决了大步长滤波器的滞后性问题。理论和仿真结果的研究表明,本方法对提高机动目标的定位精度有着明显的优势。本文所作的主要工作如下:首先,对信息融合的基本原理、意义、融合模型和估计理论作了较详细的介绍。其次,针对卡尔曼滤波算法中野值及观测噪声协方差影响滤波精度的问题,提出了一种自适应的卡尔曼滤波算法,通过仿真实验,验证了该方法良好的定位效果。然后,在传统分布式融合结构的基础上提出了基于卡尔曼滤波的分布式两层融合方法,通过仿真实验,验证了该方法的有效性。最后,基于现有的实验平台设计了应用信息融合技术的移动机器人目标定位系统实现方案,及算法流程。该多传感器的信息融合系统将听觉定位、视觉和超声传感器结合起来形成了一种新的模型,为移动机器人的目标定位提出了一种创新性的新研究思路。更多还原

【Abstract】 Nowadays, target locating with multi-sensor information fusion technology is one of the key technologies in researching on application of mobile robots. Multi-sensor information fusion technology is able to combine every aspect of information provided by several sensors; and increase the precision of target locating. The performance of the locating is more accurate and all-sided.In this paper, an adaptively target tracking method based on double-Kalman filter in existence of outliers has been proposed. If there are outliers in measurements, the method can adaptively adjusting the measurement, to ensure the property of the innovation, meanwhile, it can change the measurement noise covariance based on the movement state of target, and it also solves the problem of the hysteretic of the big step filter. The simulation results proved that, the method is higher precision on target tracking. The main work of the paper is as follows:Firstly, the basic principle, significance, fusion model and estimation theory are introduced more detailed.Secondly, because the outliers and the measurement noise covariance infect the precision of Kalman filter, an adaptive Kalman filter is proposed. The simulation shows that the effect is good.Thirdly, based on the normal distribution fusion structure, a multi-sensor optimal information decentralized fusion filtering with a two-layer fusion structure and Kalman filter is given for discrete time linear stochastic control systems. The simulation shows that the algorithm is effective.Finally, Based on the Experimental platform, designed a set of information fusion locating system, and target calculation program flow chart. It combines visual sensors, audition sensors and the ultrasonic sensors to formation a new model of localization. And, proposed an innovative new method for mobile robot locating target.更多还原