基于最适带宽的核密度估计源搜索方法  

Source Search Method for Kernel Density Estimation Based on Optimal Bandwidth

作　　者：曹凯[1] 杨海宇 李长红 杨帆[1] CAO Kai;YANG Hai-yu;LI Chang-hong;YANG Fan(School of Electronic Information Engineering,Xi'an Technological University,Xi'an Shaanxi 710021,China;Northwest Institute of Mechanical&Electrical Engineering,Xianyang Shaanxi 712099,China)

机构地区：[1]西安工业大学电子信息工程学院,陕西西安710021 [2]西北机电工程研究所,陕西咸阳712099

出　　处：《计算机仿真》2024年第3期441-447,共7页Computer Simulation

基　　金：陕西省科技厅重点研发计划项目(2022GY-238);陕西省技术创新引导专项(2022QFY01-16)。

摘　　要：针对发生放射性物质泄漏事故后的辐射源定位问题,提出一种基于最适带宽计算的核密度估计移动机器人源搜索方法,通过对传统带宽计算方法进行改进从而提高核密度估计的精度。利用搭配传感器模块的移动机器人在环境场中采集少量观测数据,计算出概率密度模型并建立空间场的密度分布图,估计出密度最大的位置设为移动机器人的运动方向,移动机器人在运动的过程中继续采集数据并再次进行核密度估计更新带宽以及场密度分布图,最终实现源定位。在室内环境下利用光源模拟辐射源进行定位实验,实验结果验证了所提出方法的有效性。For the problem of locating radiation sources after a radioactive material leak,this paper proposes a mobile robotic source search method based on improved nuclear density estimation.The accuracy of kernel density estimation has been improved by refining the traditional bandwidth calculation method.The mobile robot equipped with the sensor module collects a small amount of observation data in the environmental field.Then,the probability density model is calculated and the density distribution map of the space field is established.The maximum density position is estimated as the movement direction of the mobile robot.In the process of moving,the mobile robot continues to collect data and perform kernel density estimation again.Meanwhile,the bandwidth and field density distribution map are updated to achieve source search and location.In the indoor environment,a light source was used to replace the radiation source,experimental results verified the effectiveness of the proposed method.

关 键 词：源搜索 核密度估计 最适带宽 核密度分布图 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]