Research On Active Indoor Positioning System Based On Inertial Measurement

With the rapid progress of microelectronics and communication technologies, real-timepositioning and navigation technology has been developing very quickly and finding moreand more applications. However, in the indoor environments, the most important places wherewe live and work, major positioning systems such as the Global Positioning System (GPS) areencountering positioning difficulty or accuracy problems due to the signal block by buildingsand the multi-path distortion effect. Many indoor positioning systems with differentperformance and application scenarios have been developed based on different techniques,but deployment difficulty and coverage range are still common issues to hinder theirapplications. As an autonomous positioning method independent of extra infrastructure, theinertial positioning technology has unique advantages in deployment and coverage range. Onthe other hand, owing to the continuous development of smart mobile devices and the sensortechnology, current smart mobile devices not only have strong computing capacity, but alsohave complete built-in sensors for inertial measurements. Therefore, realizing an inertialindoor positioning system on the smart portable device is highly feasible now. In this thesis,we design and implement an active inertial indoor positioning system based on a popularsmart mobile device. The system can realize relative positioning actively without using extrainfrastructures, having great applicability in the indoor environments.The active inertial indoor positioning system is implemented by using the strapdowninertial positioning method. The basic idea is to obtain the three-dimensional acceleration in afixed coordinate system by combining the built-in accelerometer with the gyroscope inmotion, then the three-dimensional velocity and position information in the same coordinatesystem. To achieve this purpose, we need to fulfill the following three parts of the main work:collecting the inertial measurement data including acceleration and angular velocity first;then processing the data by coordinate transformation and integration to convert theacceleration to the displacement; decreasing the noise and accumulated error during datacollecting and processing stages finally so that the positioning process can be completed andthe positioning results can be reliable and accurate enough. The main research work of thisthesis is described as follows.Firstly, a smart mobile device based on Apple’s iOS platform is selected. The inertialmeasurement data are collected from the built-in accelerometer and gyroscope which arecontrolled by the CoreMotion framework. An extra timer is used to synchronize theaccelerometer and the gyroscope as well as to set the sampling rate, making them to work asan inertial measurement unit. With the help of the magnetometer controlled by theCoreLocation framework, the inertial measurement data in the device coordinate system are then transformed to the geographic coordinate system.Secondly, during the position calculation process, an active inertial indoor positioningalgorithm is proposed in order to improve the positioning accuracy and rate of success. In thisalgorithm a zero velocity update method is adopted. The detection of motionless statusbetween moving steps is judged by combining the magnitude of angular velocity with themagnitude of acceleration and its variance. An extended Kalman filter (EKF) is then designedto track errors of the acceleration, the angular velocity, the velocity and the displacement. Inthe motionless status between moving steps, the actual measurement of velocity feeds theEKF as the error of velocity, and posteriori error estimations of velocity and displacement areupdated by the EKF to correct the velocity and displacement.Finally, in order to verify the effectiveness of the algorithm, simulation is performed onthe MATLAB platform by using the inertial measurement data collected by the smart mobiledevice. The verified algorithm is then adjusted appropriately and applied to a device based oniOS platform to implement the positioning system for final test. The test results show that thesystem is effective to realize relative positioning in typical situations such as the middle orshort distance walking at a normal or slow velocity. The displacement error can be wellcontrolled within15%of the walking distance during a single positioning process, and theerror does not change significantly with time.