Research On Radioactive Source Search Method Based On Mobile Robot

With the rapid development of the worldwide nuclear industry,radioactive source loss events occur frequently.At present,manual search for uncontrolled sources is one of the main methods,and the search personnel are vulnerable to radiation damage.Today,The development of robotics is changing with each passing day.In order to improve search efficiency and reduce the risk of personal injury,using intelligent robots equipped with small detectors is an important research content in the field of nuclear emergency.According to the needs of the national “13th Five-Year Plan” nuclear energy development research project “Key Technology Research of Nuclear Emergency Disposal Robot”,a three stages search routine is proposed to find out of control radioactive sources: the inversion of radioactive source parameters,gradually approaching the radioactive source and avoiding obstacles.(1)Based on the statistical characteristics of the radiation space of point sources,a Bayesian inference model for the inversion of radioactive source parameters is established,including two models: joint reasoning of multiple measurement points and progressive reasoning.In order to obtain the result of the joint reasoning,the traditional Metropolis–Hastings sampling method is improved,and the adaptive MH is used to estimate the parameters of the radioactive source.Combining robots to search for radioactive sources is a gradual process,and the particle filter algorithm is used to estimate the source parameters.Since the estimation of the source parameter is a static parameter estimation problem,it will lead to the sample impoverishment effect.By adding a small amount of noise to the particles and using MH algorithm to improve the particle sampling importance sampling process,the progressive estimation of the source parameters is completed.(2)The search process of the radioactive source is a path planning problem in which the global target point is unknown,and the robot can obtain a series of local measurement information on the moving path.The sub-target points are set by the idea of the rolling window method,and the robot approaches the radioactive source through continuous iteration of the sub-target points.The principle and process of setting subtarget points by using local measurement information are studied.The navigation system of source-seeking robot is designed,including three main modules: sourcing path planning module,trap escape module and strategy control module.(3)The problem of obstacle avoidance in the process of robot sourcing is studied.Because the radioactive source search robot completes the final global path planning through continuous local path planning,it mainly studies the obstacle avoidance problem under the local path plan,which is realized by improving the traditional artificial potential field method.The improved artificial potential field method can make the robot reach the target point near the obstacle.(4)Due to the particularity of the radioactive source,computer simulation experiments and physical tests were carried out on the theoretical methods.The result proves that the robot which is equipped with a small radiation detector,combined with Bayesian theory,can complete the search of unknown radioactive sources,which provides a feasible solution for radioactive source search.