Research On Indoor Localization Method Of PDR Based On Multi-sensor Fusion

Location Based Service(LBS)technology has been widely used in the fields of navigation,rescue and location tracking.Due to the complicated and volatile indoor environment,satellite signals can't be perceived well.GPS positioning can hardly meet people's positioning needs.Domestic and foreign scholars have proposed a variety of indoor positioning technologies such as infrared positioning,ultrasonic positioning,RFID positioning,WiFi positioning,mobile sensors positioning.These technologies have achieved better positioning results,but they have their own limitations and are difficult to implement on a large scale.In this paper,there are existing issues such as sensor drift and accumulated errors in Pedestrian Dead Reckoning(PDR)based on a variety of sensors embedded in mobile smart phones(accelerometers,gyroscopes,magnetometers,etc.).Solve the above problem.Our specific research work includes the following three aspects:(1)BP neural network model predicts moving distance.The traditional PDR technology calculated pedestrian moving distance through pedestrian steps detection and step estimation,with easy to detect and calculate errors.In this paper,BP neural network model is introduced,and a large amount of data is used to train and test the model to achieve a better prediction effect,which can effectively reducing the calculation error caused by the traditional PDR technology.(2)Design a micro-heading angle fusion algorithm.Pedestrian walking direction can be calculated by the gyroscope and magnetometer,due to the sensor measurement inaccuracy and environmental interference,a single sensor calculation has a larger cumulative error.Therefore,this paper proposed a micro-heading angle fusion algorithm,which divides the pedestrian walking process into four types of micro-scenes and uses three kinds of micro-heading angles to perform weighted fusion to obtain a walking heading angle,which effectively reduces the cumulative error of the sensor.(3)Pedestrian indoor location tracking.We implement location tracking by pedestrian movement distance and heading angle.A total of 16 individuals were tested in this study.The experimental results show that compared to using only a single gyroscope or magnetometer to calculate the direction angle,the positioning error is maintained within the range of 1 to 4 m.There are 7 individuals for the PDR(BP+ Gyroscope)method.The PDR(BP+ Magnetometer)method has only 4 people.The proposed positioning method PDR(BP + Micro-heading Angle Fusion)has 12 individuals,accounting for 75% of the total individuals,the best positioning error is only 2.16 m.It shows that the proposed positioning method can effectively reduce the cumulative effect of errors.