Localization with vehicular WLAN based on null frame round trip time measurement

Abstract

A significant part of Intelligent Transport Systems (ITS) is the precise location determination of users in the transport network. Global Navigation Satellite Systems (GNSS) are commonly used, but they are likely to suffer from persistent outages when driving through tunnels. Impairment of the propagation path between satellite and receiver due to interference with other RF signals or blockage due to shadowing will also corrupt the performance. In this thesis, an easy to deploy, low-cost localization application for existing roadside units (RSU) is presented. The goal is to augment the position estimation obtained from a GNS system. The proposed approach is set up on the acknowledgment procedure of IEEE 802.11 based systems, which are commonly used in ITS. Here, a data packet is positively acknowledged with an acknowledgment (ack) frame. The round trip time (RTT) between sending a frame and receiving the acknowledgment frame can be exploited to calculate a pseudorange. Using two spatially separated RSUs, and trilateration of the RTT readings, an 802.11 device can be localized. Before starting with the implementation, the used radio module was extensively tested to identify possible dependencies regarding among others, the reliance on temperature, modulation and rate, a correlation between the RTT and transmit power, and the effect of multipath propagation. In the following, MATLAB was used to simulate the application before implementing it into C code. Subsequently, the ITS-G5 protocol stack, which is deployed on the RSUs, was modified to extract the RTT readings that are further processed in an evaluation unit. Hereby, care was taken to keep the impact of the localization application as small as possible since the main functionality should not be disturbed.The implementation was tested outdoors in three measurement campaigns. While doing so, different RSU positions, antenna setups, and transmission parameters were used. After each measurement series, the implementation was further improved. It showed that multipath propagation has a significant influence, whereby an accurate and robust estimation is only possible in a virtually ideal environment, making it impractical for the use in real-world conditions. However, combining the presented method with received signal strength measurements and using an additional RSU could substantially improve the estimation accuracy.