Research Of Indoor Positioning System Based On WiFi And IMU Fusion

In indoor situations,positioning services cannot be effectively applied because there is no GPS signal or the signal is too weak.Indoor positioning scheme based on WiFi is widely used,but its transmitting source is unstable and indoor transmission has problems such as signal attenuation and multipath effect,resulting in poor positioning accuracy of indoor positioning system based on WiFi.The indoor positioning system based on inertial measurement element also has a large cumulative error.These two methods have inherent defects,the positioning effect is not ideal.In this paper,the combined positioning method of WiFi and inertial measurement is carried out to improve the above problems.In the research of indoor positioning system based on WiFi,the experimental area is divided into four area blocks.Through careful arrangement of the locations of WiFi access points,each area block is covered by five WiFi access points.But the larger RSSIs received in each area come from different access points.This paper collects data in multiple periods and multiple batches,establishes a comprehensive fingerprint database,and supplements the missing data.Aiming at the shortcomings of ordinary fingerprint database algorithms,a step-by-step matching online positioning algorithm was proposed.Experiments show that this scheme improves the accuracy of WiFi positioning.In the indoor positioning research based on the inertial measurement unit,the micro-electro-mechanical system sensors are studied from the aspects of gait detection and direction monitoring.In this paper,the wave peak detection method of linear acceleration sensor for gait detection and the direct monitoring method of gait sensor as well as the soft solution of acceleration sensor and gravity sensor for heading Angle detection and the direct monitoring direction sensor are studied.The results are applied to PDR indoor positioning algorithm,and the experiment shows that the one-way accumulation error of this scheme is serious.In the combined positioning system,the extended Kalman filter algorithm is used to fuse the WiFi positioning with the inertial sensor.The position coordinates and heading angle information output by the PDR algorithm are used a s the state vectors of the system,and a state equation is established.The position coordinates output by the WiFi positioning system and the yaw Angle obtained from the transformation of the coordinate system are taken as the observation vector of the system,and establish the observation equation.Based on the extended Kalman filtering principle,the Jacobian matrix of the non-linear function partial derivative is obtained,and the dynamic noise covariance matrix and measurement noise covariance matrix of the filter are obtained based on the positioning errors of the two positioning methods.Experiments show that the positioning accuracy of this scheme is higher than that of systems using WiFi alone and inertial positioning alone.