Skip to main content

Artificial intelligence training of autonomous vehicles

The aim of this research topic is to develop a 3D urban multiplayer game environment in which artificial training of autonomous vehicles can be performed to navigate successfully through hazardous urban environment by recognizing, anticipating and avoiding hazardous events from their various radar, lidar, camera, ultrasound and GNSS sensors of their world.

This multiplayer game will be populated by cyclists on trainers cycling through this virtual environment, by driver on car control trainers learning how to drive in this virtual environment and by walkers and runners on treadmills getting fit in this virtual environment. Since current legislation forbids autonomous vehicles on public roads this virtual environment will act as a perfect environment for autonomous vehicles to teach their artificial intelligence systems about the potential hazards that may occur in this environment and develop potential strategies for navigating through this environment, This will require a virtual autonomous vehicle to be developed with all the imaging, lidar and proximity sensors and steering, break, gear and accelerator/brake/clutch pedals actuators to be modelled. So an extensive research is required not only on the modelling techniques for creating this multi-player gaming environment but also on the sensors and actuators for the autonomous vehicle’s local artificial intelligence system to make local decisions and the communications systems required for transmitting the relevant data to a centralized artificial intelligence systems in the cloud to make global decisions. In this research project it is likely that the artificial intelligence systems will be developed in Python because it is gaining a reputation for have very advanced artificial intelligence libraries that can be used to develop the functionality to perform these tasks.

In this research project, it is likely that the main application for developing this multi-player gaming environment will be Unity3D software. Unity is a very powerful 3D engine for making large games and various 3D environment effects.

This project can be used in other ways, such as education and gaming. Driving is dangerous for novices. If you don't practice more, it is easy to have an accident, and this project can meet the training needs of the novice driver. In a virtual scene the car simulator can be used to learn how to drive the virtual car. Continuous practice in virtual environments can be used to improve the proficiency of driving a car which can greatly reduce the occurrence of traffic accidents. For games, the Unity engine can make the scene more realistic and enhance the gaming experience.

Urban World 1
Urban World 2
Autonomous Driving in Urban World 1
Autonomous Driving in Urban World 2

Meet the Principal Investigator(s) for the project

Partnering with confidence

UKRI Research England

Organisations interested in our research can partner with us with confidence backed by an external and independent benchmark: The Knowledge Exchange Framework. Read more.

Project last modified 21/06/2021