Safe, Ethical and Efficient Autonomous Vehicle Navigation Algorithms

Lead Research Organisation: Aston University
Department Name: College of Engineering and Physical Sci

Abstract

The UK autonomous transport industry is a rapidly growing sector focused on the development, testing, and deployment of self-driving vehicles. Public trust in self-driving vehicles is a critical factor that can significantly impact the adoption and widespread acceptance of autonomous technology. There are concerns with safety, where reported accidents can erode public trust in autonomous technology. There are also issues with the lack of understanding of the technology and the willingness to hand over control to a machine. The lack of understanding of autonomous technology can lead to complete mistrust, and potentially the dismissal of the technology. The UK automotive industry also contributes to a large share of the annual CO2 emissions. There is also increasing pressure from regulatory bodies for decarbonisation of the automotive sector. A possible solution to these challenges comes from the development of navigation algorithms that factor in safety, social actions, and ethics (morally accepted algorithms). Using a model-based approach (such as model predictive control) with intelligent navigation algorithm design, an approach will be developed to improve the safety, operate morally acceptable algorithms, and improve the efficiency of autonomous vehicles. In the development of intelligent navigation algorithm approaches (that factor in safety, social actions, and ethics), feedforward technology will be considered (e.g., external road cameras, drone cameras and signage). Further considerations will be given to the transition of technologies, e.g., from human-driven vehicles to autonomous vehicles, and the various powertrains (e.g., combustion, hybrid, electric and hydrogen). It is our vision to develop algorithms that build public trust in self-driving vehicles through improved safety, reduced emissions, accepted/ethical approaches that are transparent. The developed navigation algorithms will be tested using data from key Birmingham roads (i.e., junctions, roundabouts, and highways), with the projected benefits disseminated to improve public trust in the technology.

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/T518128/1 01/10/2020 30/09/2025
2885906 Studentship EP/T518128/1 01/10/2023 31/03/2027 Joshua D'Souza