School of
Information Technology and Electrical Engineering

Speaker: Mr Size (Adrian) Xiao
Seminar Date: Mon, 30/10/2017 - 14:00
Venue: 78-621
Host: Prof Neil Bergmann

Seminar Type:  PhD Thesis Review


Autonomous vehicles like drones and driverless cars have been developing rapidly in recent years and are widely used in various industries such as logistics, remote data collection and personal entertainment. In a typical Guidance, Navigation and Control (GNC) hierarchy, the path planning function is normally at the top level and involves determining the waypoints to the final goal. For maximum flexibility, it is important that all data processing should be made not only in real time, but also on-board. This is challenging, especially for small/micro vehicles when the path planning algorithm becomes computationally complex. In this research, in order to achieve better algorithm implementation efficiency, existing path planning algorithms are accelerated by dedicated parallel hardware architectures which can be implemented on reconfigurable devices like FPGAs. The RRT (Rapidly-exploring Random Tree) style algorithms are adopted as a case study. By considering algorithmic parallelism, RRT and RRT* hardware architectures are proposed and these show significant power-efficiency improvement compared with corresponding software implementations. The proposed architectures are generic for path planning problems with different dimensionality, and can be adapted for different sizes and styles of reconfigurable computing devices.   


Size (Adrian) Xiao received his Bachelor of Engineering with major of Detection, Guidance and Control from Northwestern Polytechnical University in China, and MPhil in pico-satellite platform design from University of Sydney, Australia. Currently, he is studying toward a PhD degree in path planning hardware acceleration at UQ under the supervision of Prof. Neil Bergmann, Dr. Adam Postula and Dr. Matthew D'Souza.