Making Road Safety more inclusive thanks to AI
Zurich, September 15th 2032. Hannah is driving with her colleagues from the city center to her company’s office outside of the city. Or rather, Hannah is riding, an autonomous shuttle is driving her and her colleagues — full self-driving vehicles have finally become ubiquitous as the technology advances. The shuttle slows down midway down the block at Rämistrasse: there is no intersection nor traffic or crossing pedestrian ahead, however the autonomous shuttle slows down as if it was foreseeing some invisible danger. A few seconds later, a cyclist whips by on an electric bicycle. The shuttle’s ADAS (Advanced Driver Assistance Systems) had incorporated the latest version of Vianova’s Road Safety API (Application Programming Interface) and was able to alter its course in a mixed traffic zone.
2032 seems far away, but the technology described above is already live since its inception by Vianova in early 2022: today it’s being tested and fine tuned, tomorrow it will hopefully become mainstream for millions of connected vehicles across the globe. It all started through an innovation project with several leading car manufacturers which provided millions of anonymized data points to help identify dangerous road corridors and hard braking events across several major European cities. Combining this data with world class analysis of active travel data from vulnerable road users (pedestrians, cyclists, and electric micro-mobility users) , Vianova’s algorithms are able to locate streets in these cities where engineering and design interventions could improve safety outcomes for all road users.
Emerging trends around micro-mobility and autonomous driving have huge potential to decarbonize urban environments, but coexistence with more traditional modes of transport can prove difficult and risky, especially in the early phase. There is increasing competition for the same space, with new modes and uses taking up space that was allocated to cars only a few years ago. Governments have ambitious goals to achieve Vision Zero (zero killed or seriously injured road users) , while at the same time encouraging more active modes of transportation (i.e. bikes, e-scooters, e-bikes, etc.). The confluence of these objectives can create tensions and add to risk. Indeed, 41% of road traffic fatalities in urban areas involve a car and another mode of transportation.
It’s no secret that modern vehicles (i.e. passenger cars, delivery vans, e-scooters, etc.) are sensors on wheels, making them potential tools for generating a very interesting tool for harvesting large amount of real-time, geo-localized data in cities: as a matter of fact, 100% of all new vehicles sold will be connected in 2024. Connectivity is progressively bringing multiple new use cases for car data in cities: hyper precise traffic flows, real-time parking information, street-level change detection, real-time accident warning, collision risk detection (i.e. car vs pedestrian vs bicycle), road pavement conditions, last-mile delivery, etc. On the other hand this data can now be plugged into near real-time digital twins of cities (i.e. HD maps, smart traffic, inventory of city assets and curb, etc.) allowing more and more accurate identification and resolution of safety risks as cities grow denser and mobility complexifies.
The nexus of these two areas is very interesting to the work we do at Vianova. We understand how to visualize and interpret large sets of connected vehicle data — what lessons could we learn from focusing on private cars instead of shared bikes and scooters? And what are the areas of overlap between vision zero and carbon zero? This line of questioning led us to begin to retool our core product, Cityscope, with the features that will better enable it to be a source of road safety intelligence for cities in the future.
The technology developed by Vianova is powered by FCD (Floating Car Data) and is using Machine Learning algorithms in order to:
These features, encapsulated into Vianova’s Road Safety API and dashboard, can support cities and OEMs in achieving multiple use cases:
Thanks to this technology, the cities of Basel and Zürich were able to gather powerful insights from various datasets in order to qualify 10 risk zones (e.g. Werdstrasse, Langstrasse, Flurstrasse, Manessestrasse, etc.). Building off of both historic crash data (i.e. open data from Switzerland), the big data sets from connected cars and micro-mobility devices (i.e. speed, position, braking pressure) and street segment data (i.e. Open Street Maps for road segments and speed limits) the tool identified and rated areas of concern and future evaluation by engineers, including a few unexpected locations. The methodology we used to come up with a reliable risk scoring as an output was to compute the average value from 4 main risks related to cars:
- Most dangerous road over the last 10 years (ranked by number of accidents per segment) ;
- Most over-speeding road by car (ranked by number of over-speeding flows) ;
- Most intense braking road by cars (ranked by the number of braking pressure events > 10) ;
- Most median over-speeding road by car ;
All scores have been normalised by the number of flows, to avoid that bigger is the flow the higher is the risk.
The cities of Basel and Zürich are now working with Vianova to calibrate the model and test its robustness before applying it to the greater cities’ areas. Vianova is working to expand on this Road Safety API initiative and support the safety of road users across the continent through better data insights. This is also our contribution to something much bigger and we believe, for the greater good: the UN and the European Commission have recently launched a number of coordinated initiatives such as AI for Road Safety and “VisionZero” European Framework in order to harness Artificial Intelligence to close the digital and road safety divide.
The initial use cases for this technology have been implemented with municipal governments and car manufacturers through the provision of a Road Safety decision tool, however we are already thinking of other potential powerful use cases that bring safety insights to more people — fleet managers could use the data to create safer routing for vehicles ; infrastructure planners could make more strategic resource allocations of limited right-of-way. Part of Vianova’s mission is to develop the solutions that make cities even safer, especially as more people are getting around on two feet or two wheels. At Vianova, we believe in creating the good data to make cities greener (and safer) — this tool is a great new way to help our users do both.
About Vianova
Vianova is the trusted mobility intelligence platform for cities and mobility providers to achieve CarbonZero and VisionZero. Our data platform helps transport providers and cities, better integrate and manage shared, connected, electric and autonomous transport solutions in the urban space, enabling better use of city infrastructure, and promoting safer and more sustainable mobility. Vianova has offices in Paris, Zürich and London.
If you have any comment that you would like to share with Vianova, please send your comments to hello@vianova.io. If you would like to learn more about what it is like to work at Vianova, and join our talented team, visit our job-board or send directly your application to jobs@vianova.io.
