The Application Of Multi-Sensor Information Fusion Algorithm For Indoor Positioning System

With the development of information society and popularity of smart mobile devices, the demand of personal positioning had reached a new level. The location based services (LBS) had also been considerably developed in many fields, such as navigation, tracking, tour guides, etc. The foundation and key technology of LBS is navigation technology, including outdoor and indoor positioning. The traditional methods using GPS technology and mobile communication methods have sophisticated algorithms and high accuracy in outdoor environments. However, their performances in the indoor environment are unsatisfactory. In recent decades, the indoor positioning technology is in continuous development and progress, but most of these methods suffer from several disadvantages. Thus, we need to focus on the accuracy and continuity of indoor pedestrian positioning.In this paper, we present an indoor positioning scheme utilizing smartphones and multi-sensor information fusion algorithm. This method is based on the improved pedestrian dead reckoning algorithm and magnetic map matching algorithm. First, this paper researches and summarizes the law of pedestrian gait. Then the traditional pedestrian dead reckoning algorithm has been optimized and improved with higher adaptability and effectiveness. At last, this paper proposes a magnetic map matching algorithm based on particle filter to analyze and calibrate the results obtained from the pedestrian dead reckoning algorithm.The improved pedestrian dead reckoning algorithm could calculate pedestrian’s walking track. After data analysis and processing of the built-in accelerometer and gyroscope, the system could draw the travel distance and direction. Then the system uses magnetic map matching algorithm to calibrate the positon information obtained from the pedestrian dead reckoning part. This will eliminate cumulative error and gain higher accuracy. Once the pedestrian movement switches to stationary, the improved pedestrian dead reckoning algorithm would send relevant information to the magnetic field calibration module. The pedestrian’s final position will be calculated using real-time data of magnetic field sensor and pre-existing map of magnetic field. This paper also puts forward a floor detection algorithm based on the pressure sensor, which could be used in the three-dimensional indoor positioning field.This paper designed and implemented the system using multi-sensor information fusion algorithm. After field test and simulation verification, we can draw a conclusion that the proposed method can deal with complex human motion and various phone-holding manners. The proposed method does not need any other infrastructure (such as additional receiver and emitter hardware) except a smartphone. It also has many advantages such as low energy consumption, high precision and high efficiency.