The Research Of UWB/IMU Indoor Pedestrian Positioning Algorithm Based On Particle Filtering

The global navigation satellite system(GNSS)has achieved full coverage of open areas,and the positioning accuracy can reach millimeter level.Due to the influence of indoor environment,it is difficult to guarantee the required positioning accuracy by only relying on a single sensor,and it is easy to have dead space or lost positioning.This paper addresses shortcomings of Ultra Wide Band(UWB)positioning and inertial positioning,proposed corresponding algorithms to improve the positioning accuracy of a single sensor.To achieve people’s positioning needs in different indoor situations,the particle filtering algorithm is used to fuse the two positioning data.UWB positioning technology needs to measure the time of flight of signals in the air.Because of multipath effects and non-visual propagation when obstructions are present,large positioning errors are introduced.The key to improve UWB positioning accuracy is to reduce its ranging error.Therefore,this paper proposed a range value filtering algorithm to eliminate the zero value and correct the abrupt change value.It also uses Newton interpolation to weaken the interference introduced by multipath effect,non-visual propagation and clock error in ranging.Finally,practical tests are conducted with the self-designed UWB positioning module to verify the proposed method.The results prove that the proposed method can effectively reduce the ranging value error.The UWB positioning accuracy is improved in both static and dynamic cases.In the autonomous localization scheme based on Pedestrian Dead Reckoning(PDR)algorithm,the localization error mainly exists in the course calculation and step estimation process.Due to the fact that the integration process of angular velocity,the heading angle is drifted with increasing time.The existing nonlinear step estimation models are empirical models for a single walking trajectory,which cannot effectively adapt to different walking trajectories of pedestrians.Therefore,this paper proposed a heading angle calculation method for the UWB tag position,in which the pedestrian heading angle update is calculated by jointly weighting the heading angle value solved by the Inertial Measurement Unit(IMU)and the heading value of the UWB tag position,thus effectively suppressing the IMU heading drift.Moreover,this paper establishes a step estimation model based on heading information correction to adapt to different walking trajectories of pedestrians.The experimental results show that the pedestrian heading angle values calculated by the weighted fusion do not drift with time,and the modified step length estimation model can more accurately estimate the step lengths for different walking trajectories of pedestrians.Compared with the existing methods,the proposed method effectively improves the accuracy of heading calculation and step length estimation.The PDR positioning accuracy is improved.Finally,this paper proposed a particle filter-based fusion algorithm to fuse the localization information of UWB and IMU.In the indoor scene,pedestrians walk wearing the UWB and IMU positioning modules designed in this paper,and the upper computer terminal saves the positioning information.Through semi-physical simulation,the effectiveness of the fusion algorithm is verified.The fusion localization accuracy is significantly better than that of single sensor localization.The localization system does not need to collect a large amount of offline data.It is also not targeted at specific pedestrians,which is convenient for the popularization and application of the localization technology.