Stephenson, Scott
(2016)
Automotive applications of high precision GNSS.
PhD thesis, University of Nottingham.
Abstract
This thesis aims to show that Global Navigation Satellite Systems (GNSS) positioning can play a significant role in the positioning systems of future automotive applications. This is through the adoption of state-of-the-art GNSS positioning technology and techniques, and the exploitation of the rapidly developing vehicle-to-vehicle concept. The merging together of these two developments creates greater performance than can be achieved separately. The original contribution of this thesis comes from this combination: Through the introduction of the Pseudo-VRS concept. Pseudo-VRS uses the princples of Network Real Time Kinematic (N-RTK) positioning to share GNSS information between vehicles, which enables absolute vehicle positioning. Pseudo-VRS is shown to improve the performance of high precision GNSS positioning for road vehicles, through the increased availability of GNSS correction messages and the rapid resolution of the N-RTK fixed solution.
Positioning systems in the automotive sector are dominated by satellite-based solutions provided by GNSS. This has been the case since May 2001, when the United States Department of Defense switched off Selective Availability, enabling significantly improved positioning performance for civilian users.
The average person most frequently encounters GNSS when using electronic personal navigation devices. The Sat Nav or GPS Navigator is ubiquitous in modern societies, where versions can be found on nomadic devices such as smartphones and dedicated personal navigation devices, or built in to the dashboards of vehicles. Such devices have been hugely successful due to their intrinsic ability to provide position information anywhere in the world with an accuracy of approximately 10 metres, which has proved ideal for general navigation applications.
There are a few well known limitations of GNSS positioning, including anecdotal evidence of incorrect navigation advice for personal navigation devices, but these are minor compared to the overall positioning performance. Through steady development of GNSS positioning devices, including the integration of other low cost sensors (for instance, wheel speed or odometer sensors in vehicles), and the development of robust map matching algorithms, the performance of these devices for navigation applications is truly incredible.
However, when tested for advanced automotive applications, the performance of GNSS positioning devices is found to be inadequate. In particular, in the most advanced fields of research such as autonomous vehicle technology, GNSS positioning devices are relegated to a secondary role, or often not used at all. They are replaced by terrestrial sensors that provide greater situational awareness, such as radar and lidar. This is due to the high performance demand of such applications, including high positioning accuracy (sub-decimetre), high availability and continuity of solutions (100%), and high integrity of the position information. Low-cost GNSS receivers generally do not meet such requirements.
This could be considered an enormous oversight, as modern GNSS positioning technology and techniques have significantly improved satellite-based positioning performance. Other non-GNSS techniques also have their limitations that GNSS devices can minimise or eliminate. For instance, systems that rely on situational awareness require accurate digital maps of their surroundings as a reference. GNSS positioning can help to gather this data, provide an input, and act as a fail-safe in the event of digital map errors. It is apparent that in order to deliver advanced automotive applications - such as semi- or fully-autonomous vehicles - there must be an element of absolute positioning capability. Positioning systems will work alongside situational awareness systems to enable the autonomous vehicles to navigate through the real world. A strong candidate for the positioning system is GNSS positioning.
This thesis builds on work already started by researchers at the University of Nottingham, to show that N-RTK positioning is one such technique. N-RTK can provide sub-decimetre accuracy absolute positioning solutions, with high availability, continuity, and integrity.
A key component of N-RTK is the availability of real-time GNSS correction data. This is typically delivered to the GNSS receiver via mobile internet (for a roving receiver). This can be a significant limitation, as it relies on the performance of the mobile communications network, which can suffer from performance degradation during dynamic operation. Mobile communications systems are expected to improve significantly over the next few years, as consumers demand faster download speeds and wider availability. Mobile communications coverage already covers a high percentage of the population, but this does not translate into a high percentage of a country's geography. Pockets of poor coverage, often referred to as notspots, are widespread. Many of these notspots include the transportation infrastructure.
The vehicle-to-vehicle concept has made significant forward steps in the last few years. Traditionally promoted as a key component of future automotive safety applications, it is now driven primarily by increased demand for in-vehicle infotainment. The concept, which shares similarities with the Internet of Things and Mobile Ad-hoc Networks, relies on communication between road vehicles and other road agents (such as pedestrians and road infrastructure). N-RTK positioning can take advantage of this communication link to minimise its own communications-related limitations. Sharing GNSS information between local GNSS receivers enables better performance of GNSS positioning, based on the principles of differential GNSS and N-RTK positioning techniques. This advanced concept is introduced and tested in this thesis.
The Pseudo VRS concept follows the protocols and format of sharing GNSS data used in N-RTK positioning. The technique utilises the latest GNSS receiver design, including multiple frequency measurements and high quality antennas.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
