Research On Indoor Location Technology Of UWB Position Fingerprint Fusing Dead Reckoning

From the birth of the concept of robots in the last century to the present,robots with different functions have been manufactured and applied to human production activities and various fields.In recent years,artificial intelligence has brought more attention to robots.Application field of robot more concentrated in the indoor environment,along with the indoor robots of broadening the application field and the improving of the artificial intelligence level made the robot toward the direction of more independent,and robot to complete the autonomous navigation task,it is necessary to provide accurate positioning function,at the same time the robot when performing certain tasks such as inspection,monitoring,etc.,should not only provides a measure of the important information,location information at the same time also is indispensable.Different from wide outdoor conditions,indoor environment is generally complex and changeable,which requires the positioning system to have more stable performance and excellent positioning effect.At present,the general robot positioning in the line-of-sight environment can still meet the positioning requirements,but in the non-line-of-sight environment,especially the dynamic non-line-of-sight environment,the robot positioning accuracy still has challenges.Therefore,it is very practical and necessary to seek more practical and accurate positioning methods.In order to further improve the effect of the location in the robot in a complex indoor environment,this paper puts forward to superior performance in indoor positioning technology of UWB(Ultra Wide Band,UWB)position fingerprint positioning technology combined with dead reckoning positioning technology,make the two complementary advantages positioning technology,Cubature Kalman filtering algorithm is then used to fusion of two kinds of positioning methods,so as to achieve goals for improving indoor robot localization effect.UWB position fingerprint orientation of fingerprint feature information of each reference point in the uses of uwb signal arrival time(Time Of Arrival,TOA),offline phase using fastclustering algorithm to improve the density of the peak position of fingerprint classification,location area,divided into online stage by first class matching judgment target location which area is located in the area,and then in this area with the improved compression perception reconstruction algorithm,compression sampling matching pursuit algorithm estimate the position of target.This method can achieve centimeter-level positioning accuracy under the condition of line-of-sight distance.Considering the frequent occurrence of dynamic non-line-of-sight in indoor environment,which seriously affects the positioning accuracy of position fingerprint,the dead reckoning positioning technology,which is not easily affected by environmental factors,is introduced,and the robot's posture can be detected in real time by the motion sensor carried by the robot itself,which can achieve high-precision positioning within a short distance.However,with the extension of time,the positioning errors of dead reckoning will gradually accumulate,which is not suitable for long-distance positioning.Therefore,an improved volume kalman filter algorithm is adopted in this paper to organically integrate uwb location fingerprinting and dead reckoning location results and sensor information,and further improve positioning accuracy by combining positioning methods to make positioning results more stable and reliable.To verify the accuracy and the feasibility of this indoor positioning method,this paper built the uwb positioning system in the indoor environment,and the ultra broadband mobile label and dead reckoning positioning robot localization experiment modules installed in the robot,the PC real-time experimental data collected and processed in order to draw the robot's movement.After processing and analyzing the experimental results,it is shown that the positioning method proposed in this paper can effectively solve the problem of positioning error caused by dynamic non-line-of-sight,and improve the indoor positioning effect of the robot significantly while ensuring the real-time positioning.