Research On Autonomous Source Finding Method For Nuclear Radiation Detection Robot

Nowadays,the nuclear industry technology has obtained unprecedented development,but the loss and theft of radioactive sources have occurred.The early work of source search is mainly human.It is often carried out by wearing protective clothing and holding radiation detector.The way of searching radiation source will harm the health of the staff.With the further development of robot technology,the search of radiation source is gradually replaced by robots.How to search the radiation source by the robot equipped with radiation detector is an important research content of radiation emergency treatment.This paper uses the detection robot in radiation environment to search the radiation source.The research contents can be divided into three parts: static and dynamic positioning of radiation source,trajectory planning of robot,source search simulation and real environment verification.(1)For the location algorithm of radioactive source,it can be divided into two kinds.One is the static radiation source localization algorithm based on MCMC,which requires a human robot to make the radiation detector collect the data of multiple observation points in the source searching environment,and invert the parameters of the radiation source based on the Bayesian framework to get a Markov chain about the parameters of the radiation source,After several iterations,the parameters of the radiation source are estimated,and the MCMC is improved based on the environment.The second is a dynamic source localization algorithm based on improved particle filter.The algorithm uses the idea of generating particles to narrow the range of particles by means of particle importance sampling,particle resampling and particle adaptive updating,and then estimates the source parameters through the final convergent particles.MCMC based localization algorithm is suitable for relatively simple environment with small obstacle area.Particle filter based localization algorithm is suitable for relatively complex environment with high real-time positioning requirements.(2)In the process of dynamic source searching,the trajectory of the radiation source needs to be planned.The artificial potential field algorithm is suitable for trajectory planning of robot.In the static source search algorithm,the robot needs to be placed in the nuclear environment.Firstly,the environment is discretized and divided into several regions with equal size.The observation data is measured randomly in each region.The Markov chain is established and the source position is estimated by using these data.Finally,an improved artificial potential field method is used to plan a path from the starting point to the estimation point.In the dynamic source search algorithm,it is necessary to make the robot move forward for a distance and collect a small amount of observation data,then estimate the source position by particle filter.By improving the artificial potential field,the robot can move forward a certain distance and measure and estimate the source again.Repeat the search for many times until the distance between the source estimation position and the robot position is less than the threshold.The improved artificial potential field algorithm belongs to the local path planning algorithm,which is not only suitable for the static source seeking process of robot,but also for dynamic exploration.(3)Finally,virtual simulation and physical verification are carried out for the two improved algorithms.The results show that both the static MCMC method and the dynamic particle filter algorithm can search the radioactive source.