Safety and Optimisation of Autonomous Vehicles

It is increasingly clear that a structured approach using artificial intelligence (AI) components operating within a framework of conventional logic will deliver autonomous vehicles (AVs), which have the potential to benefit the world by increasing traffic efficiency, reducing pollution and eliminating up to 90% of traffic accidents. The installation of 5G network systems from 2019 will accelerate the development of AVs and make them quickly becoming a reality in the future. Many governments believe that this will happen with 10 years. However, there are still many challenges. One of them is the safety and reliability issue of autonomous vehicles, which is still in its infancy. There were 124 “disengagement” incidents in a Google driverless car test in 2016. 8% were caused by reckless behaviour of other drivers on the road, whereas the rest were caused by software issues and unwanted behaviour from the vehicle. The test results in 2018 were much better, but autonomous vehicles are still not safe enough for public roads. This demonstrates that AVs are far from achieving full disengagements. The existing AI systems are not yet mature enough to handle the full spectrum of unforeseen circumstances on the road. More research should be done on improving safety and reliability of AVs before AVs become commercialisation.

This project aims to investigate and apply effective techniques and software (such as Artificial Intelligence, Optimisation and Reliability Methods) to improve the safety and reliability of autonomous vehicles.

The research topics could be one of the following research (not limited):

· Safety and Reliability. To study the factors of safety and reliability on autonomous vehicles and propose related simulation models.

· Object Detection. To develop more accuracy and efficiency of object detections. Artificial intelligence will be applied to model situation awareness around AVs.

· Optimisation. To develop fast decision-making system, including path planning.

· Data processing. To develop fast data processing algorithms for data collected from sensors.

· Large scale of optimisation of connected vehicles.

This research will be carried out in the Automotive Research Centre

Entry Requirements:

The appointed PhD candidate is expected to have:

· A strong and solid background in computer science, optimisation or reliability.

· Good oral and written communication skills.

· The ability to work collaboratively as part of a team.