Research On Improved RatSLAM Algorithm Based On WIFI

Simultaneous Localization and Mapping(SLAM)is a self-localization and incremental map construction of robots in some known or completely unknown environments by on-line measurement and position estimation of sensors carried by the robot itself..SLAM is a hotspot and a difficult point in the research field of intelligent mobile robots.It is considered to be the key to truly realize the robot's autonomous positioning and navigation.Mobile robots are essentially mobile sensor platforms.Sensors vary in type and capability,but they all have limitations of use and varying degrees of noise.The traditional probability-based SLAM method is based on the Bayesian state estimation theory.Only the recursive relationship before the data is considered,and there is error accumulation.The long-time large-scale scene navigation will cause map inconsistency.With the in-depth study by scholars at home and abroad,Australian scholar Milford et al.proposed a bionic navigation algorithm based on the RatSLAM model.This algorithm is a purely visual positioning and navigation algorithm.For positioning navigation in complex environments,accuracy is not easy.High,poor environmental adaptability,low success rate of visual image matching,and poor robustness.The indoor positioning technology based on WIFI signal has become a research hotspot in the field of indoor environment location perception due to its wide range of use,low cost and portability.However,in the complex and variable indoor environment,the positioning certainty and anti-interference ability are better,difference.In response to the above problems,this paper has made the following related research.Firstly,a new image matching method that replaces the absolute difference sum(SAD)matching in the original RatSLAM model is adopted.Firstly,the HSV image features are used to perform rough matching of the global features of the image,and then the SURF and ORB fusion algorithms are used to further perform local features.Accurate matching,assisted RatSLAM model to better complete template matching,and thus achieve more accurate correction of the activity of the pose cell network,and obtain an excellent path experience map.Secondly,an improved location fingerprint recognition algorithm based on WIFI is adopted.The direction angle of the mobile robot and the number of access points are added to the offline database.The online location stage is combined with the K-means clustering algorithm for location estimation.This makes the positioning accuracy significantly improved.Finally,a complementary sensor fusion system is designed based on the inherent characteristics of the sensor to apply the appropriate sensor modality at the appropriate time.The image template matching algorithm or the WIFI fingerprint information matching algorithm applied to a single camera is extremely demanding for the matching accuracy requirement,the threshold values are all a fixed value,and the error matching is easy;the image information and the WIFI fingerprint information are adopted.The matching mechanism converts the fixed threshold into the interval threshold;the image information obtained by the camera is used for image matching,and once the camera information is indeterminate,the WIFI template matching is started immediately;the two cooperate with each other and use the coordination mechanism to call at different times.Appropriate template matching makes the matching accuracy significantly enhanced,resulting in a more accurate experience map.In order to verify the feasibility and effectiveness of the proposed algorithm,the "Voyager?" mobile robot was used to collect the experimental data of the experimental platform,and the data information was transmitted to the upper computer to complete the MATLAB simulation experiment.The experimental results confirm that the improved algorithm is better than the improved algorithm,the positioning accuracy is obviously improved,the generated experience graph is more accurate,and the system robustness is significantly enhanced.