Research On The Key Techniques Of Indoor Localization

Accurate indoor geolocation is an important and novel emerging technology for commercial, public-safety, and military applications. However, indoor localization is very challenging due to the complex signal propagation that is caused by obstacles such as walls, clapboard, ceiling and so on. The electromagnetic wave may suffer reflection, refraction, diffraction and may result in the phenomenon of dense multipath arrivals and non line-of-sight propagation both of which will severely degrade the localization accuracy. Due to the large bandwidth, Impulse Radio Ultra-wide (IR-UWB) technology holds the advantages in anti-multipath, anti-interference, penetrability, high-precision ranging, low complexity implementation, low cost, low power consumption and becomes one of the most promising technologies. But at present, the indoor localization based on IR-UWB is still facing lots of problems and needs in-depth researching and improving. This dissertation has been launched to investigate the key technologies of indoor positioning, which is important and valuable for the practical application.First of all, the optimum quantization for the UWB finite resolution digital receiver is studied. Low-resolution quantization is an effective scheme to deal with the large bandwidth in UWB digital receiver; however, traditional uniform quantization will lead to serious information loss. Based on the analysis of fundamental quantization theory and UWB signal detection, two different kinds of optimization problems about optimum quantization are investigated. The first problem comes from the signal detection in noise and connects the quantization parameters with the detection probability. The corresponding optimum quantization thresholds and levels are derived with maximizing the detection probability as the optimization function. The second problem comes from the binary communication and makes the minimization of the bit error rate as the optimization function. The influence of quantization parameters on UWB signal detection and sign detection is also explored, especially under the low signal-to-noise (SNR).Secondly, the estimation of time of arrival (TOA) is studied. Dense multipath arrivals and NLOS propagation are common and changeable in indoor environment, which will decrease the TOA accuracy. Based on the analysis of the characteristic of UWB signal propagation in indoor environment, two robust algorithms for TOA estimation are proposed. The first algorithm combines the nonparametric detection based on conditional tests with low-resolution quantization and degrades the sensitivity of noise parameters on judgment threshold. The second algorithm exploits the information that the optimum threshold is closely related with the SNR and utilizes adaptive threshold with respect to timely estimated SNR.Thirdly, indoor localization method for static nodes is studied. Dense multipath arrivals and NLOS propagation lead to distance measurement errors differing from the traditional model and then can reduce the positioning accuracy. Based upon the research about distance errors, two different types are obtained. The first type is negative error that is caused by false alarm detection of noise samples. The second type is positive error, which is caused by missing detection of the leading path. The negative error is reduced by soft-decision algorithm in which more than one measurement are reserved by each station and selected by minimizing the cost function. The positive error is decreased by public area optimization algorithm in which the point of minimal average distance error is deemed as the node’s position. By combing those two algorithms, a comprehensive location method is proposed to improve the indoor positioning accuracy.Finally, indoor tracking method of moving node is studied. NLOS propagation is one of the key factors that affect tracking accuracy in indoor environments. An adaptive tracking algorithm is proposed to mitigate the NLOS error for indoor mobile localization. The correlation between adjacent NLOS errors in time was analyzed and exploited. A modified extended Kalman filter (MEKF) is presented which includes the NLOS errors as part of the state variables. NLOS identification is achieved based on the state estimation of MEKF. MEKF and NLOS identification are combined to implement the adaptive tracking algorithm. Simulation results demonstrate that the proposed algorithm has better tracking accuracy and adaptability in indoor environments.