Research On Indoor Positioning And Navigation Algorithms Based On Inertial Sensor And Smartphone

Since the signal intensity of satellite is decreased when reaches the ground and it cannot penetrate the building wall,this result in satellite positioning system cannot achieve accurate navigation and location in indoor environment.As an autonomous navigation technology,inertial navigation use inertial sensors and inertial navigation algorithm to realize real-time navigation,because it does not need receiving device in building and is not affected by environmental changes and signal block,inertial navigation has obvious advantages compare to other indoor navigation methods.This study uses micro mechanical inertia measuring element and the sensor in the Android smartphone,designed on the Android platform,proposes an indoor inertial navigation algorithm based on pedestrian movement posture recognition.the main contents are as follows.1.Indoor positioning technology of scientific research institutions at home and abroad in recent years and indoor navigation algorithm on the base of inertial sensors were analyzed and summarized.The principle of strap-down inertial navigation technology and pedestrians dead reckoning navigation were contrasted and analyzed on the base of inertial components.The navigation coordinate,posture method,and the corresponding coordinate conversion method are introduced,which are used in inertial navigation system.The quaternion attitude algorithm was used to obtain the real-time attitude angle of vehicle equipment.2.The corresponding error elimination method and IIR lowpass filtering algorithm was put forward on the base of inertial sensor error analysis.Feature detection and periodic division were analyzed for the pedestrians' acceleration and feet posture information on the base of pedestrians walking gait features.Moving average filtering algorithm and zero speed sampling algorithm was adopted for the mobile electronic compass data correction.3.The fusion algorithms of three conditions(acceleration amplitude?acceleration variance?angular velocity amplitude)zero speed detection and foot pitch attitude recognition was adopted to accurately divide the zero speed point of pedestrian movement,and the zero velocity correction technology was used to reduce the cumulative error caused by acceleration of quadratic integral.Through the experiments and research on stability and regularity of pedestrian pace and step,analyzing and comparing different step model algorithms,the pedestrian frequency stability of inertial navigation calculating results was used to set up inertial navigation step length model.4.In the process of pedestrian's actual walking,through the recognition of pedestrian movement frequency,attitude angle and direction angle,the stride frequency and length model matching was adopted in pedestrian normal speed walking,which effectively reduce the accumulative errors caused by the traditional algorithm of inertial navigation and the Random error caused by pedestrian and Inertial sensor.5.In order to study the controlled experiment in walking,two inertial sensors was worn on different foot side at the same time,the feet of the inertial data was fusion and processing,the optimization of stride frequency estimate step length detection and direction were realized.6.Through the simulation experiment with the MATLAB software,the navigation precision and the feasibility of indoor inertial navigation algorithm based on pedestrian movement state recognition were verified.The hardware platform of indoor positioning system was set up which based on MPU9250 inertial sensor and Android smartphone.The software of indoor positioning navigation system was designed on Android software platform to verify the application performance and real-time performance of the navigation algorithm.