Indoor Positioning Method Based On Fusion Of Smartphone Inertial Sensor And RSSI

In recent years,with the rapid development of smart cities,people's demand for indoor location services is increasing,and many new indoor location technologies are emerging.However,the positioning methods suitable for smart phone platform are very limited.Pedestrian dead reckoning(PDR)method based on inertial sensor and location fingerprint positioning algorithm based on RSSI have attracted people's attention due to their advantages of easy implementation,low cost and little dependence on the outside world.Location fingerprinting based on RSSI can obtain the absolute location information,and there is no cumulative error,but it is susceptible to the impact of the external environment,while PDR location is not affected by the environment,but the cumulative error will be generated after a long time.In view of this,we obtain the positioning information obtained from the two positioning methods on the smartphone platform,and then use the Extended Kalman Filter algorithm to merge the two advantages,so as to improve the positioning stability and improve the positioning accuracy.The main contents of this article are as follows:(1)In the indoor magnetic interference environment,the existing heading estimation method has a large cumulative error,and a pedestrian heading correction algorithm based on smartphone sensors is proposed:on the basis of traditional complementary filtering,a PI regulator is added,and based on the angular velocity data output by the gyroscope,the error compensation coefficient of the PI regulator is adjusted in real time,and then the heading angle calculated by the pre-processed accelerometer and magnetometer data is compensated,and the compensated heading Angle is fused with the gyro data to obtain a more accurate heading Angle.The experimental results show that in the environment with strong indoor magnetic interference,the average heading error is reduced by 3.4° and 1.8°respectively compared with the traditional complementary filtering algorithm and the nine-axis data fusion algorithm,and the positioning accuracy is improved by 68.4%and 65.9%,which verifies that the algorithm has good applicability to the environment with strong magnetic interference,and can improve the reliability of pedestrian heading angle calculation to a certain extent.(2)Based on the analysis of the existing online matching algorithm for location fingerprint positioning based on RSSI,this paper explores the improvement of the online matching algorithm from the perspective of selecting the optimal reference point:in the offline stage,fuzzy c-means algorithm(FCM)is used to divide the location region and generate the clustering fingerprint.In the online stage,firstly,the region is located,then on the basis of dynamic weighted k-nearest neighbor algorithm(EWKNN)to dynamically select k value,according to the MAC address sequence of the points to be measured and the reference points in the selected area are matched,only the strongest wireless access point(AP)is trusted,further eliminate the more remote reference points,so as to screen out the best reference point among k.The experiment is designed and the accuracy analysis is carried out.The improved algorithm is better than EWKNN algorithm in the aspects of average positioning error,maximum error and error accumulation probability.(3)On the basis of the improvement of the above two positioning methods,fusion positioning is performed by the Extended Kalman Filter algorithm.Unlike classical Extended Kalman Filtering,which continuously integrates the entire positioning system from start to end,we correct it at certain special locations of the test site(corners,large server rooms,etc.where the positioning is not ideal)to improve the accuracy of fusion positioning The experimental results show that the maximum positioning error of this method is reduced by 39.7%based on the classic Kalman Filtering method,and the reliability of the positioning error within 1.5m has been greatly improved,which can obtain more accurate location information of pedestrians to a certain extent.