AI for old dumb sensors
Road traffic is increasing across the UK year-on-year (DfT,2022), creating an increasingly complicated interface between vehicles and pedestrians. Crossing points in particular are a hotspot for accidents; 1-in-5 pedestrians report having an accident or near-miss at a crossing (Admiral,2022). As the government and local councils push to get more people out of cars and onto bicycles or foot, a landscape of increasing driver numbers and congestion creates a striking need for tools to support road user flow optimisation while minimising accidents.
Connected Smart City Infrastructure will play a major role in road management, with Local Highway Authorities (LHAs) actively seeking new methods for understanding and improving driver, cyclist and pedestrian behaviour at crossings. Many crossing sites have useful sensors already installed, but these generate only low-quality, low-resolution information (from passive infra-red, radar or basic optical sensors) to achieve site-specific functionality. In their current form, the information the generate is unable to be used to optimise the crossing or to integrate with Smart City Infrastructure and facilitate the shift towards safer, smarter crossings.
To overcome this limitation, Starling Technologies (alongside other market leaders) created new smart sensors and analytic software services using computer vision and machine learning (ML) to enable dynamic responses to a range of crossing-use factors, including numbers, types of users and speed. Using this information allows authorities to keep road users safer while improving travel times, tensioning driver, pedestrian and cyclist requirements. This technology has proven effectiveness; however installation remains expensive and time-consuming due to the hardware requirement and must be installed as a replacement to existing crossing-management installations.
Starling therefore proposes the creation of a new, software-first networking technology. This will use machine learning to augment the existing sensors system information to capture and extract data from crossings, enabling the mass-integration of existing crossing management systems within Smart City Infrastructure projects and improving the effectiveness of each site, through smart and dynamic optimisation.
Potential applications go far beyond public road crossings, with on-going testing of Starling technology on private sites and large retail warehouses. We are exploring further use-cases in private roads (e.g. business hubs and retail parks), indoor roads (e.g. major event spaces or stadia) and industrial sites.
This is a game-changing innovation, unique in the market, that will help make roads safer for everyone, improve traffic flow in cities and establish a major UK competitor in the global smart infrastructure market.