Real-time Intelligent Map-matching Algorithms for Advanced Transport Telematics Systems (RiMATTS)

各种交通应用和服务，如按驾驶付费保险计划、导航和路线引导、事故和紧急响应（增强型999紧急服务）、公交车站的公交车到达信息和车队管理，都需要空间和时间位置以及时间信息。这种服务的一个重要组成部分是导航模块，它提供所需的定位数据。许多商业设备可用于支持这种运输系统的导航模块。近年来，大多数商业设备使用GPS技术来获取这样的定位数据。由于全球定位系统存在系统误差和噪声，许多运输服务所需的定位精度无法通过这种装置实现。此外，这样的设备不提供位置解决方案的完整性（置信水平），这对于责任和安全关键应用（例如，随开随付保险方案（由于不正确计费的可能性））和对紧急999呼叫的响应是非常重要的。需要采用将位置数据（来自GPS或其他传感器）与空间道路网络数据集成的地图匹配算法。地图匹配不仅能够识别车辆的实际位置，而且如果有好的数字地图，还可以提高定位精度。当前的地图匹配算法由于与它们相关联的固有限制和约束而不能够在某些操作环境中（特别是在密集的城市区域中）支持某些运输系统的导航模块。除此之外，单个地图匹配算法不能最佳地支持不同操作环境中的运输系统的导航模块。因此，有一个独特的需要，选择一组代表性的地图匹配算法。这些算法通过实验的详细特征是必不可少的，以评估其性能的操作环境中，他们的设计，并确定其局限性。具有进一步增强的现有地图匹配算法的此代表性集合连同可考虑现有地图匹配算法的限制及约束的新地图匹配算法沿着可在大多数操作环境中最佳地支持大多数运输系统的导航模块。因此，本研究项目的主要目标是（1）从现有算法中识别一组代表性的地图匹配算法，（2）开发一种新的地图匹配算法，并解决目标1中在应用和操作环境方面识别的任何差距，（3）发展以知识为本的智能地图配对技术，以找出最佳的地图配对算法（4）展示iMM在不同作战环境中的潜在应用。几个标准将被定义为使用的iMM技术，以选择最佳的算法，在一个给定的操作环境中的特定服务。这些标准将包括作业环境的地理特征（如土地使用、道路网密度和建筑物高度信息），并在必要时包括其他特征（如复杂性和成本）。本研究的开发将在两个层次：（1）算法，（2）实际的导航系统，其中包括算法和导航传感器。预期与实际导航系统相关的成本将相对较低（在每单位500美元的水平）。随着导航传感器芯片和基于MEMS技术的传感器的价格随着时间的推移而降低，预计这一数字将下降。

A variety of transport applications and services such as pay-as-you-drive insurance scheme, navigation and route guidance, accident and emergency responses (enhanced 999 emergency services), bus arrival information at bus stops and fleet management require spatial and temporal location, and time information. One of the important components of such services is the navigation module which provides the required positioning data. Many commercial devices are available to support navigation modules of such transport systems. In recent years, most commercial devices use GPS technology for acquiring such positioning data. Since GPS suffers both systematic errors and noise, the required positioning accuracies of many transport services cannot be achieved by such devices. Moreover, such devices do not provide integrity (the level of confidence) of position solutions which is very important for liability and safety critical applications such as pay-as-you-drive insurance schemes (due to the possibility of billing incorrectly) and responses to emergency 999 calls. A map matching algorithm that integrates the locational data (from GPS or other sensors) with the spatial road network data needs to be employed. Map matching not only enables the physical location of the vehicle to be identified but also improves the positioning accuracy if a good digital map is available. Current map matching algorithms are not capable of supporting the navigation modules of certain transport systems in some operational environments (specifically in dense urban areas) due to the inherent limitations and constraints associated with them. In addition to this, a single map matching algorithm cannot optimally support the navigation module of a transport system in different operational environments. Therefore, there is a distinct need to select a set of representative map matching algorithms. The detailed characterisation of these algorithms through experiments is essential to evaluate their performance in the operational environments for which they were designed and to identify their limitations. This representative set of existing map matching algorithms with further enhancements, along with a new map matching algorithm that can take into account limitations and constraints of existing map matching algorithms, could optimally support the navigation modules of most transport systems in most operational environments. Therefore, the main objectives of this research project are to (1) identify a set of representative map matching algorithms from existing algorithms, (2) develop a new map matching algorithm and to address any gaps identified in objective 1 both in terms of applications and operational environments, (3) develop a knowledge-based intelligent map matching (iMM) technique to identify the best map matching algorithm (achieved in objectives 1 and 2 above) suitable for an operational environment, and (4) demonstrate a potential application of iMM in different operational environments. Several criteria will be defined for use with the iMM technique to select the best algorithm for a particular service in a given operational environment. Such criteria will include the geographic characteristics of the operational environment (such as land-use, road network density, and building height information) and others (such as complexity, and cost) if required. The exploitation of this proposed research would be in two levels: (1) the algorithms, (2) the actual navigation system which incorporates the algorithms and the navigation sensors. The expectation is that the cost associated with the actual navigation system will be relatively low (at the level of 500 per unit). This is expected to fall as the price of navigation sensor chips and MEMS technology-based sensors reduce over time.