Research On SLAM Method Of Cardinality Balanced Multi-Bernoulli Filtering Based On Multi-Platform Cooperation

Since the Industrial Revolution,human society has entered a stage of rapid development.Land resources can no longer meet the needs of human development,and the ocean contains rich resources,which has accelerated the pace of exploration of the ocean.Underwater robots have received much attention and research because they can replace workers to complete various underwater tasks,and have become an extremely important means for countries to explore the ocean.Navigation and positioning technology is one of the key technologies necessary to explore the ocean.It is very important for underwater robots working in unknown complex underwater environments,and determines whether underwater robots can successfully complete their tasks.Excellent navigation methods can improve the accuracy of underwater robot navigation and positioning,and efficiently complete seabed exploration tasks.Simultaneous Localization and Mapping method is the key technology for realizing robot autonomy.This method effectively combines map construction and self-positioning.It is a good method for underwater robots to complete seabed terrain detection.This is also the main research content of this article.This paper mainly studies the SLAM algorithm that under the frame of random finite set,and proposes the cardinality balanced multi-bernoulli filtering SLAM method under the framework of random finite set.Based on this method,A SLAM method of cardinality balanced multi-bernoulli filtering based on multi-platform cooperation was proposed.The main content of this paper is showed as following.First of all,to against the problems like high complexity,the need for data association,and large amount of calculation of traditional EKF-SLAM,Fast SLAM and other methods with,A SLAM method of cardinality balanced multi-bernoulli filtering under the theoretical framework of random finite sets was proposed in this paper.This method does not need to perform data association,nor does it need to use approximate strategy for particle weight calculation,which reduces the calculation complexity of the algorithm,the unbiased potential estimation was utilized in the potential equalization multi-Bernoulli filtering method to avoid excessive estimation of the number of map features.This method improves the accuracy of map feature estimation,thereby improving the overall accuracy of map estimation.Secondly,to against the feature that the SLAM method under the framework of random finite set cannot provide a priori information,A SLAM method based on multi-platform cooperation based cardinality balanced multi-benulli filtering was proposed in this paper.Inthis method,multiple underwater robot platforms jointly detect the surrounding environment in the same unknown environment.During the detection process,the underwater robot platforms share the map feature data information of each detection to increase the a priori information of the map.This method improves the estimation accuracy of the map feature of the underwater robot at the current moment,thereby the estimation accuracy of the entire map is improved.Finally,the estimated features of each platform in the CBMBer-SLAM method based on multi-platform cooperation are fused to obtain the final map feature estimate and the accuracy of map feature estimation is further improved.