Research On SLAM Algorithm Of Robot In Complex Scenes

Simultaneous Localization and Mapping(SLAM)is the key technology for mobile robots to achieve autonomous navigation.The application environment of mobile robots is now more complicated.In a large-scale or dense map environment,the uncertain interference of the external environment makes the robot sensor noise characteristics abrupt.The traditional mobile robot SLAM algorithm will have problems such as decreased estimation accuracy and real-time degradation,which seriously affects the practical application.Therefore,how to improve the positioning,mapping accuracy,robustness and real-time performance of the mobile robot SLAM algorithm in complex scenes is of great theoretical significance and practical value for promoting the development of the mobile robot industry and realizing the autonomous navigation of mobile robots.Firstly,the motion control model and sensor observation model of the mobile robot SLAM system are discussed and established,and the mathematical model of the SLAM problem is described.A unified simulation model is established for the subsequent algorithm research.Secondly,the EKF-SLAM algorithm based on extended Kalman filter is deduced in detail,and its defects are analyzed.The data association part of EKF-SLAM algorithm is studied in depth,and the calculation of nearest neighbor data association algorithm in large-scale environment is considered.The problem of large size and low precision is proposed by local association strategy and dynamic adaptive association threshold criterion.The nearest neighbor data association algorithm is improved.The improved data association algorithm matches possible map features in a local map,and the associated thresholds follow the pose.The state estimation error changes and the correlation error is avoided due to the constant correlation threshold.The simulation results verify the effectiveness of the improved algorithm.Finally,the FastSLAM algorithm based on particle filter is deeply studied.In the actual environment,the assumption of steady-state noise in the FastSLAM algorithm is often not satisfied.External disturbance will cause the system model to change,the sensor noise characteristics will be abrupt,and the difference between the proposed distribution functionand the true posterior probability distribution will be increased.The problem of particle degradation is aggravated,and the diversity expression of particles is also affected,and the problem of particle depletion occurs.Aiming at the above problems,the adaptive multi-fading UKF is used to obtain the proposed distribution function,and the PSO algorithm of the exception mechanism is introduced to optimize the particle distribution,so that the proposed distribution function can truly reflect the posterior probability distribution of the system,so that the FastSLAM algorithm can deal with the unsteady noise.The ability to apply the mobile robot SLAM algorithm in complex scenarios is also significant.The effectiveness of the improved FastSLAM algorithm is verified by simulation platform and real environment standard dataset.