Research On Multistatic UWB Radar Localization And Tracking In Multipath Environment

With the development of location-based services(LBS), indoor positioning has a very bright future, such as indoor surveillance and mall navigation. Because of its simple structure, anti-multipath, strong penetrability, high ranging accuracy and good coexistence, impulse radio(IR) ultra wideband(UWB) technology has an advantage in indoor positioning. As a result, it has been a hot point of research for decades. However, there are still many problems in the aspects of the UWB propagation channel analysis in complicated environments and practical high-precision ranging methods. The UWB technique can also be used in battlefield surveillance. With respect to the monostatic radar, the multistatic UWB ground surveillance radar system has many advantages, such as good survival, anti-interference and ability of detecting concealed and stealth targets. It is an important developing direction of battlefield surveillance radar. The critical problem and important research subject in short-range battlefield surveillance is how to achieve robust positioning and tracking of moving persons in unknown complex battlefield environment.Combined with several research projects, the thesis studies the problems about the application of the UWB technique in target localization and tracking in indoor and foliage multipath environments. The research subjects include the following aspects.(1) UWB propagation characteristics in complex environments,(2) practical UWB ranging algorithms,(3) theoretical performance analysis of localization and tracking with a multistatic radar system,(4) location ghost elimination and multitarget localization in multistatic localization,(5) robust multistatic tracking in a multipath environment. The research results are validated by simulation and experimental data. They can be applied in UWB localization, multistatic battlefield survillence radadr design and multistatic localization and tracking.1. Some characteristics of the IEEE 802.15.4a standard channel are analyzed for ranging and localization applications, such as the distribution of the location and amplitude of the following path after the the first path(FP), the distribution of the multipath number and their amplitudes within the 2ns interval after the FP, etc. The conclusions lay a foundation for ranging algorithms. A two-stage multipath model is proposed to describe the IR UWB propagation channel between the transmitter, non-cooperative target and the receiver. The UWB through-the-wall propagation characteristics are analyzed based on the experimental data. The results show that there exists multipath in UWB through the wall propagation. After background subtraction, the analysis of the residual echoes proves that there are multipath phenomena associated with the moving target. When the environments change, such as the change of the relative location between the transceiver and the target, the multipaths associated with the target will also change.2. To lower the strict requirement on thresholds in time of arrival(TOA) estimation, a new practical method is proposed to estimate the TOA of the FP with the energy detection receiver. The method utilizes the received signal characteristics, i.e. the randomness of the threshold crossing(TC) noise samples and the definiteness of the FP sample. Based on the entropy, a novel method without thresholds is proposed to detect the FP from a different angle in dense multipath. Based on the analysis of the UWB channels, the dense multipath region has good steadiness, especially near the FP. As a result, the FP can be detected by detecting the multipath region based on the sudden change of the entropy series. To avoid using the complicated chi squared distribution, the power transform is applied to the output of the energy detection(ED) receiver and the best approximate distribution of the transformed data is sought. Based on the best approximate distribution, the TOA is estimated. Similar TOA estimation performance with methods based on complex theoretical chi-squared distribution is achieved. A new TOA estimation method of higher accuracy is proposed based on some delayers after analyzing some parameters of the ED receiver.3. The study systematically analyzes the theoretical performance of two aspects of the multistatic radar system, i.e. observability and Cramer-Rao lower bound(CRLB). Two cases of observability are analyzed, i.e. the observability of instantaneous location and velocity and the observability of different trajectories. Two cases of CRLB are also analyzed, i.e. the CRLB of instantaneous location and velocity estimation and the tracking CRLB of different trajectories. Theoretical analysis and simulation results can provide guidances for system design.4. When the distance-sum and Doppler measurement errors change, their impacts on localization and velocity estimation are analyzed. Based on the analysis results, the principle of velocity consistency is first proposed and deduced. The velocity consistency means that the target location and velocity are unique at one moment. The velocities estimated from different observation combinations should be consistent. When there are false observations, the estimated velocities will be inconsistent. Based on the velocity consistency, the true and false observation combinations can be discriminated. Then multistatic observations can be matched and multitarget localization can be achieved. Simulation results prove its high recognition accuracy for false observation combinations.5. A novel method is proposed to identify the true and false trajectories while tracking based on trajectory connectivity in multipath environments. The multipath echo signals associated with the moving target are taken as echoes from a false target. After seeking all current trajectories, the reflection point is found. The true and false trajectories can be discriminated based on the relationship of propagation paths. Simulation results show the recognition rate is high when the target approaches the baseline. To deal with missing and false observations, a robust tracking procedure is designed with measurements of different dimensions through an adaptive filter structure. To deal with wrong measurements caused by interference, we use the nearest neighbor method and location or velocity consistency to discriminate the true and false measurement combinations in complex environments. To track the zigzag maneuvering soldier, which is common in battlefield, the study proposes a new interacting multiple model(IMM) algorithm based on the piecewise constant velocity(CV) model. A maneuvering model is constructed, which can better describe the true moving state of the person. The model set in the new method can promptly adapt to the change of the moving direction. When the direction changes, the method constantly adjusts the directions of the model set until it reaches a steady state. Then the new estimated state of motion will be closer to the true state. Simulation results show that it can achieve a better tracking performance than traditional maneuvering tracking methods.