复杂室内环境下基于单声阵列的智能移动终端三维定位关键技术

The demand for real-time and efficient indoor localization services on smart devices is constantly increasing. In order to meet this demand, a novel, low-cost and real-time localization framework based on a hybrid Direction of Arrival (DOA) and Time of Arrival (TOA) is proposed in this project. Inside this framework, we focus on investigating the high-accuracy DOA and TOA estimation method, efficient data fusion and tracking theory and dynamic creation of 3D indoor maps to resolve the localization problem in indoor environments, which are full of colored noises, obstacles and suffering from multipath propagations. It aims to improve the overall performance of the localization system from the theoretical study, system innovation, and also optimization of algorithms. .Several approaches are going to be explored in this project. First, in order to address the multipath propagation problem, an anti-Doppler-frequency-shift TOA estimation method based on FrFT and GCC, a fast DOA estimation method based on Co-Prime virtual array geometry, and an NLOS identification method based on statistical characteristics of acoustic channels and its spatio-temporal characteristics are going to be investigated. Second, in order to improve the user capacity of the system, a multi-user cooperative self-localization approach will be studied, and high accuracy self-localization methods will be developed. Third, in order to address the asynchronous data fusion problem from multiple measurements, an estimation correction method based on distributed tracking theory, multi-source data fusion method, track fusion and map information is going to be investigated. Forth, because of dynamic nature of indoor environment, a dynamic indoor map creation method based on mobile crowd-sensing is going to be investigated. Finally a prototype is going to be developed to evaluate the effectiveness of the algorithms and performance of the framework.

面向智能移动终端实时高效的室内位置服务需求，本课题以声信号为基础，结合惯导及地图信息，提出一种基于复合DOA与TOA架构的新型单声阵列定位方案。针对室内复杂的强有色噪声、强多径传输及易遮挡环境，研究高精度DOA及TOA估计方法、高效多源信息融合跟踪理论及室内三维地图的动态构建，从理论研究、体系创新、算法优化等层面综合提高系统性能。针对室内多径传播，研究基于FrFT及GCC的抗多普勒频移TOA估计方法、基于互质虚拟阵列的抗多径DOA快速估计，及基于声信道统计特性与时空联合的遮挡识别方法；针对系统承载能力，研究多用户渐进式协作自定位策略及高精度自定位方法；针对多种量测异步融合问题，研究基于分布式跟踪理论、多源信息融合体系、航迹融合及基于地图信息的航迹修正方法；针对室内地图动态变化的特性，研究基于移动群体感知的室内地图动态构建方法。开发原型系统，检验所提出定位方案的实用性及算法的有效性。

面向智能移动终端实时高效的室内位置服务需求，本课题以声信号为基础，结合惯导及地图信息，提出一种基于复合DOA与TOA架构的新型单声阵列定位方案。针对室内复杂的强有色噪声、强多径传输及易遮挡环境以及室内声音定位中多混叠信号时延估计及遮挡定位问题 ，研究高精度DOA及TOA估计方法、高效多源信息融合跟踪理论及室内三维地图的动态构建，从理论研究、体系创新、算法优化等层面综合提高系统性能。针对室内多径传播，研究室内声音定位中多混叠信号时延估计及遮挡定位研究；针对系统承载能力，研究多用户渐进式协作自定位策略及高精度自定位方法和单声阵列位置的3D自标定算法；针对多种量测异步融合问题，研究强多径环境下的低成本声信号TOA估计方法；针对室内地图动态变化的特性，研究了基于声音的智能移动终端室内定位关键技术研究