Research On Multi-robot Formation Oriented To Location Of Nuclear Radioactive Sources

With the wide application of nuclear-related technologies in the civilian areas,nuclear technology has brought considerable conveniences for people in their daily lives,but it can’t be overlooked that the risks of nuclear accidents have also been increasing day by day.At present,robots are mainly employed to locate and discover uncontrolled radioactive sources.However,when a single robot is used for nuclear radioactive source detection,the system has low fault tolerance.Using multiple robots to work in a nuclear radiation environment in formation can improve the efficiency of out-of-control radioactive source positioning,reduce the large deviation of single robot nuclear radioactive source parameter estimates,and increase the fault tolerance of the system.The work of this paper is mainly focused on the research of multi-robot formation to realize the location of nuclear radioactive sources,mainly from the following aspects:First,in terms of multi-robot formation,a distributed formation control method is proposed: in terms of formation maintenance,the moving speed of robots is controlled based on three rules of separation,alignment,and formation;in terms of formation changing,the main goal is to prevent collisions among robots,and to realize the change of new formation.The distributed formation control method calculates the linear velocity and angular velocity required by the robot motion,and controls the specific motion of the robot through the differential robot motion model.Second,the improved particle filter is used to estimate radioactive sources’ parameters based on the spatial statistics of the point radioactive source.Since the position of the radioactive source and its radioactivity will remain unchanged in a short period of time,the particle degradation phenomenon will occur in the estimation of static parameters of the particle filter,which needs to be solved by the resampling method.In the resampling process,the improved mean resampling method is used to update the weight,and the average value of the estimated radioactive source parameters is used to increase the diversity of particles and improve the accuracy of the particle filter to estimate the radioactive source parameters.Based on the multirobot formation and the estimation of radioactive sources’ parameters by particle filter.The estimated radioactive source parameters are used to change the moving target point of the formation,and finally reach the estimated radioactive source position.Finally,the method proposed in this paper is verified through simulation and experiments conducted in the actual environment.The results show that the mean re-sampling method proposed in this paper improves the accuracy of estimation over radioactive sources’ parameters compared with the traditional re-sampling method in the particle filter system.Moreover,the proposed distributed formation controller also presents a better performance in terms of trajectory accuracy controlled by formation and efficiency of formation changing than the trajectory tracking method proposed recently.In addition,in the process of multirobot formation,the formation can estimate the parameters of the nuclear radiation source and control the robot formation to move in the direction of the nuclear radiation source,and finally arrive near the actual location of the nuclear radiation source.