Developing deep learning techniques for semi-automatic detection of road assets in Buckinghamshire

The digital transformation is rapidly changing from the way we travel on our roads and how we provide and maintain our infrastructure. The Department for Transport is launching a competition to help stimulate further transformation and to enhance Britain’s position as a leading transport innovator. Our aim is to help encourage highway authorities to make road transport increasingly smart through the use of data and connectivity to improve safety, efficiency, and emissions.

The competition will provide £500,00 for local authority projects costing between £30,000 £100,000 which will:

  • demonstrate the capability of connected vehicle data;
  • improve the quality of road condition and asset management data;
  • provide the business case for more widespread deployment across a number of highway authorities;
  • enable the development of smart asset strategies based on harvested intelligence; and
  • help support innovation within the private sector supply chain.

Details of our entry

Amongst a number of core issues that relate to Connected Automotive Vehicle Systems (CAV's) is a need for accurate real-world data from which a CAV's inherent spatial location and proximity components can leverage an accurate real-world context. Broadly speaking, these components typically involve Deep-Learning algorithms, drawn from machine learning and neural network technologies, from which a CAV can determine its spatial proximity as accurately as possible. While this technology is showing many positive and promising developments for CAV technology, its effectiveness is significantly determined by its access to appropriate training data that provides a high level of accurately defined real-world contexts.

View full details of our entry

Print entire guide

Last updated: 12 March 2019

Was this page helpful?

Very poor
Neither good nor poor
Very good