UWB radiolocation technology: Applications in relative positioning algorithms for autonomous aerial vehicles

This report presents methods and technologies used to investigate the use of commercially available Ultra-Wide Band (UWB) radiolocation technology in the positioning of autonomous vehicles. UWB radio has unique transmission characteristics which make pico-second resolution timing possible, a requirement for centimetre accuracy in positioning applications. Firstly, an investigation into the performance capabilities of UWB positioning infrastructure was undertaken, followed by the development of relative positioning algorithms utilising a relative and decentralised control approach. These algorithms are based on feedback provided by radiolocation data in combination with sensor fusion elements, including inertial (accelerometers), optical (optical flow sensors) and absolute ranging sensors such as Light Detection and Ranging (LIDAR). Performance of the UWB infrastructure was determined utilising static and dynamic measurement techniques to verify overall positioning accuracy and statistical characteristics, with respect to two different ranging algorithms, Time Difference of Arrival (TDOA) and Two Way Ranging (TWR). Data collected from an array of static collection points has shown, in the recommended infrastructure configuration, that commercially available UWB technology has a positional accuracy with an average offset from the real coordinate system of 7.30-15.28%. Logic-based tests using statistical inputs indicate the TDOA algorithm succeeds in more accurate positioning in 79.63%, 58.33% and 85.19% of test cases across twenty-seven points in three vertical planes respectively. Data collected from autonomous flight-path algorithms has provided practical insight into applications of the technology and its dynamic capabilities. Results demonstrate that in a high radio interference environment, performance of the UWB positioning system and drone achieve a 0.11% and 1.23% error on average from the programmed flight path gradient for TDOA and TWR respectively. The results of this investigation suggest that the technology is ready and capable of being implemented on real-world systems, with the recommendation that the technology be coupled with sensor fusion elements in relative positioning applications