Research On Radioactive Source Search Method Based On Particle Filter

The research on the prediction and estimation of radioactive sources using Bayesian theory is of great significance in nuclear emergency treatment.The traditional radioactive source search method requires personnel to enter.Considering from the perspective of radiation protection and nuclear emergency,there are some problems,such as the search personnel are injured by radiation and unable to deal with the loss of radioactive source in emergency.At present,the detection method based on mobile platform is welcomed by many research fields of emergency response.How to use the land-based mobile platform to carry out the localization and search of radioactive sources with small simple radiation detectors,so as to improve the efficiency of emergency response and reduce personnel injury,is an important research content in the field of nuclear emergency response.This paper mainly focuses on the location search of unknown radiation sources in unknown radiation environment,and uses particle filter and its improved algorithm to realize mobile radiation source location search.On the basis of reading a large number of domestic and foreign literature and conducting research,this paper mainly carries out some research work in these aspects:(1)Based on the radiation spatial statistical characteristics of radioactive sources,the Bayesian theoretical model of radioactive source location search is established,and the radioactive sources with unknown location are located by gradually introducing the observation information.It provides a Bayesian theoretical basis for the radiation source location model of subsequent particle filter.At the same time,the autonomous localization algorithm of the robot is also studied,which provides the basis for the autonomous search of the robot.(2)The theory of particle filter algorithm is deeply studied.Aiming at the defects of standard particle filter algorithm,traceless transformation and divide and conquer strategy are introduced σ The point distribution strategy is improved to optimize the performance of traceless transformation.Finally,the algorithm before and after optimization is simulated and analyzed by one-dimensional nonlinear unsteady growth model and single station single target tracking model.Simulation results show that the optimized algorithm has better performance in computing time,particle diversity,error and so on.(3)Through the research content of the first part and drawing lessons from the research at home and abroad,a radiation source location model suitable for particle filter is established.Through the method of particle filter,the problem that it is difficult to obtain the analytical solution by calculating the infinite integral of a posteriori probability in Bayesian model is solved.In the third step,the particle filter algorithm is used to optimize the number of radial sources,and then the particle filter algorithm is introduced to calculate the effective number of radial sources The simulation results are analyzed from the angle of error transformation.Experimental results show that the optimized particle filter algorithm has better performance.