Pedestrian Indoor Positioning Algorithm Based On Multi-sensor Information Fusion

In recent years,with the rapid development of Internet technologies and smart mobile terminals,the research of Location-Based Service(LBS)has been rapidly carried out throughout the world,and the application range has gradually expanded.Outdoors,the Global Positioning System(GPS)can perform long-term high-precision positioning and has been developed relatively well.Most of the positioning technologies based on satellites and wireless signals require some additional equipment to assist in positioning.In some indoor environments of emergency rescue,there are no corresponding infrastructures on the site or the signal will be seriously disturbed due to sheltering by buildings,etc.,and there is a large positioning error.The inertial sensor-based positioning method can estimate the relative position information of the nodes to be located,but there is a relatively large cumulative error.In order to satisfy people's positioning requirements,designing a high-precision,low-cost and robust indoor positioning technology is a key issue to be solved.This paper considers the positioning accuracy and cost comprehensively,mainly for the positioning requirements of underground roadways or high-rise buildings,and proposes a positioning scheme that combines multi-sensors information fusion and map prior knowledge.This scheme inherits the characteristics of low sensor power consumption and high accuracy of CSS wireless positioning,combining inertial sensors and wireless positioning.First,the PDR algorithm is used to obtain the initial trajectory of indoor personnel.Then,a positioning calibration method based on a priori features and ToF is introduced to correct and update the position of the pedestrian and reduce the cumulative error.Specifically,in the positioning and calibration stage,Random Forest(RF)and a method of setting a threshold value are used to classify and mark special directions for the traveling direction of indoor personnel,and special points are matched with maps to calibrate the position coordinates of pedestrians at particular points.Among them,in order to improve the classification accuracy,a Linear Discriminant Analysis(LDA)method was used to process the feature set.This method can effectively improve the accuracy of the classifier.In the entire positioning process,this paper also incorporates a wireless round-trip Time of Flight(ToF)ranging technology to intermittently correct pedestrian position information in real time to maximize positioning accuracy.The multi-sensors information fusion indoor positioning algorithm proposed in this paper can adapt to a variety of harsh environments.In order to eliminate the cumulative error of positioning,the position calibration method designed in this paper makes full use of indoor map information to intelligently identify the direction of movement of indoor personnel and corrects and updates the position of pedestrians.The experimental results show that the positioning scheme designed by this paper can achieve more than 97% accuracy of pedestrian recognition,can effectively reduce the cumulative error,improve the positioning accuracy,and can achieve reliable real-time positioning.It is in line with the current development trend of indoor positioning.