Ford's RoadSafe technology uses an machine learning algorithm that combines and analyses anonymised data from various sources. Data sources include connected vehicles, road sensors and traffic news. The aim is to identify where there may be an increased likelihood of traffic accidents. This information can then be displayed on a map that identifies risks and could also be used in the future to warn drivers of dangerous hotspots.
The digital tool is currently being tested in London and the UK county of Oxfordshire. More than 200 vehicles are taking part in the trials. "In every city, there are stretches of road with an increased likelihood of accidents, whether due to a badly placed road sign, an unrepaired pothole or outdated routing. Now we can pinpoint such problem areas so that motorists can be forewarned and authorities can take action," explains Jon Scott, Project Lead, City Insights, Ford Mobility, Europe.
The data gathered enabled the development team to create a special map that identifies road sections of particular concern. This map is composed of different layers of data, including known past accidents and a risk prediction algorithm for each road segment. Colours are used to assess the likelihood of accidents, with red representing the highest risk and yellow the lowest.
To collect the data, the vehicles participating in the trials record relevant driving events such as braking, steering and accelerating. Roadside sensors, meanwhile, track the movements of various road users. The sensors use machine learning algorithms to detect near misses and analyse movement patterns of vulnerable road users such as cyclists and pedestrians.
The combination of vehicle and sensor data can help detect a variety of hazards. These include places where vehicles pass too close to cyclists or inconveniently located bus stops that cause congestion, as well as poorly designed roundabouts and intersections that cause