Research On The Technologies Of Personnel Positioning And Tracking And System Design Based On PDR

Location information plays an important role in people’s daily life and work. Sowith the development of societies and technologies, the demand for positioning andnavigation is also growing. However, due to the degree of modernization of the towncontinues to improve, the function of the building is more and more perfect, and itsstructure is also more complex, most of the time people active inside office buildings,museums, libraries, shopping malls and other buildings. They all result in that asimple GNSS has been unable to meet the demand for full location (indoor andoutdoor seamless positioning).With the development of the RF signal processing and wireless communicationtechnologies, a wide range of different levels of precision indoor positioning solutionshave emerged, and they all show a low-cost, low-power, high-precision, and real-timepositioning trends. However, most of these systems require positioning a certaindensity of reference locator beacons in the region to be positioned, even they require alot of preliminary measurements, the positioning accuracy depends on the referencebeacons’ performance and density, and the area that can receive location servicesfrom the system is proportional to the cost of it, so it is not suitable for large-scaleapplications. When it is faced with large-scale applications, the advantage is notobvious. In recent years, with the lower price of inertial measurement unit, the PDRtechnologies based on inertial sensing gradually becomes the solution to indoorpositioning for pedesstrians. The positioning and tracking systems based on the theoryof PDR require smaller equipment, needn’t to arrange locator beacons in the area tobe positioned, and the system costs are also only related to the number of targets to bepositioned, so it is very suitable for large-scale PDR applications. Sine themeasurement accuracy of IMU is not high and the technology of dead reckoning hascumulative errors, the performance of PDR-based positioning system needs to beimproved.For the key issues existing in positioning and tracking system that based on PDRtheory and technology, the studies of this paper as follows:(1) Analysis and recognition for personnel moving action feature based on theparameters of the inertial sensor In oder to accurately calculate the moving distance of the person, PDR system firstneeds to identify the moving actions of personnel. The IMU use inertial measurementunit to collect acceleration and angular velocity data of waist and tibia movement inthe process of moving people, then it establishes a inertial sensing database formoving action recognition. According to the analysis on characteristics of waist andtibia motion in the process of moving, such as ground walking, running, goingupstairs and going downstairs and so on, we proposed a method that combines motioncharacteristics to find a ZUPT point to make a gait division between the continuousprocess of moving. We extract the inertial data characteristics of traveling gait, andbuild an action classifier using support vector machine theory. On this basis, wedesigned an experimental platform and the experimental results show that this methodcan achieve higher accuracy in personnel traveling action classification.(2) Step estimation method based on the model of waist motionPersonnel motion is composed of every single step. The calculation of single steplength has a significant impact on the accuracy of PDR systems. Given that waist islocated near the body center of gravity, movements are relatively stable, and lowerlimb movements are in line with "inverted pendulum" model, we propose theconstruction method of step model based on the motion characteristics of waist, andsuggest that location information of personnel can be obtained based on the stepmodel. This program will place the IMU on the waist and design a calculation methodfor displacement of people walking, then measured the algorithm through manysubjects. The test results show that the total error of displacement of subjects is lessthan6.40%in standard walking process. In this paper, the proposed algorithmexperiment will migrate to smart phones or other portable computing platforms, thenwe can establish a set of PDR systems based on smartphone platforms.(3) Step calculation method based Strap-down Inertial NavigationTo further improve the accuracy of the calculation for a single step, we combine theacceleration and angular velocity of tibia motion during walking with thecharacteristics of it, and make a gait division using ZUPT point, finally calculate thestep based on the principle of SINS within each gait cycle. This paper presents amethod for that how the parameters of SINS will be reset at the ZUPT moment, firstwe will set the initial azimuth unknown, then we deduce the displacement as afunction in triaxial directions within the geographic coordinate system during everywalking and find the characteristic that tibia is perpendicular to the ground at the ZUPT moment as well as on condition that the displacement in vertical direction isalmost zero, we combine with the characteristics of three-dimensional motions duringwalking to approximate and correct the displacement. Finally, we calculate the othertwo displacements parallel to the ground, thereby calculating step. This method canaccurately remove the effects of gravity in SINS calculation process and avoid thediffusion of gravity error from double integral, then the accuracy of the step will begreatly improved. The proposed algorithm has no special requirements for thedifferent way people walk and for the road pavements. And it does not need to set anyassociated parameters for different people, so it has strong adaptability and versatility.(4) Design and construction of personnel location and tracking platform based onPDR teconologiesDue to the lack of mature corresponding experimental platform based on PDR andcombined with the needs of the study, this paper gives the method for systemconstruction in the form of functional modules, then it designs and implements a low-cost inertial measurement unit and data transmission systems. The actual PDR systemplatform proposed in this paper has completed the experimental evaluation of thealgorithm. In addition, PDR algorithm proposed in this article would also betransplanted into the smart phone platform, and it has constructed personnel locationand tracking system solutions based on smartphones.