After a friend in Los Angeles was killed by a motorist after being distracted by his smartphone, Alexander Janink, an Australian, created a camera that could take pictures inside vehicles, thanks to an infrared flash. An artificial intelligence then analyzes the photos: 95% are immediately destroyed, 5% – related to the potential field of telephone use while driving – are sent to the authorities: an operator examines them and decides whether to start the process.
The technology has been installed under the Accenture brand, in two states of Australia, India, Belgium, Spain … Road accidents recorded compared to other states and territories in Australia: 22% improvement over the previous two years, “said Alexander Janink.
The winning trio
A sensor installed on the side of the road or in the car, along with data from the outside, artificial intelligence to analyze everything … Data, connectivity, AI: This is the winning trio provided by car manufacturers, equipment manufacturers and the government, to improve road safety. According to the WHO, 1.27 million people die in road accidents every year. However, while waiting for a hypothetical generalization of a fully autonomous vehicle, there will always be a driver who needs to be controlled, informed, or predicted to behave. When Adas (Advanced Driver Assistance System) is not going to replace him …
Suppression or expectation? The camera of the Australian company RedFlex makes it possible to approve drivers driving red lights. But also to try to undo the consequences of their crime. An AI analyzes slowness and acceleration more than 45 meters before a turn, and by the end of the road a driver is about to cross – he may decide to turn on the red light – almost green.
Another example of dynamic traffic adaptation: American company Fleer’s cameras detect the thermal signatures of vehicles, pedestrians and cyclists. The traffic lights are then controlled depending on the crowd: to protect the bikes they have more time for the car to be red, if they have more.
All of this information, collected by sensors located outside of motorized vehicles, can be shared with vehicles equipped with so-called V2X connections (from vehicle to vehicle: from vehicle, infrastructure, cloud, other users, etc.). Eventually, cars, buses and even trucks will be able to automatically adjust their behavior in a certain area. It is geofencing, geofencing.
“In downtown Stockholm, we are conducting an experiment with several carrier vans: if our sensors detect too many pedestrians, the information is sent to these vehicles which automatically slows them down,” said Johannes Berg, senior adviser to the Swedish Transport Agency. .
The speed can be adjusted automatically according to the weather forecast: “We are developing algorithms in temperature, rainfall intensity, 200 meters section to predict the condition of the road and the risks that may result from it” Explained.
But V2X works differently. “Still in downtown Stockholm, another test concerns concrete mixers that communicate their weight: if they exceed the permitted limits, their speed is automatically adjusted so as not to damage roads or bridges and increase the safety of other users,” Johannes Berg added. Does.
Calculate the direction of the vehicle
Another use of the connection: this time in the vehicle. UTAC, a private group specializing in automated testing and equivalence, has installed cars and motorcycles with V2X antennas and inertia units so that the location of each machine can be determined within 10 centimeters.
Attached to a camera, our AI algorithms can detect a dosing driver. And we are continuing our research to identify drunk drivers.
Gabby Gizderveld, Marketing Director, Smart Eye
The computer in each vehicle continuously calculates the prediction of the speed of the vehicle but also of all the motorcycles revolving within a radius of 800 meters. “Tests, for example, in the case of a car turning left in front of a motorcycle, show 80% avoidance if the driver is warned of danger, 100% if it is a vehicle equipped with an automatic braking system. Considering the situation and still 100% if Drivers of both cars and motorcycles are warned, “said Theo Charbonneau, UTAC’s project manager.
“Extensive efficiency will be achieved when ‘external’ information, from infrastructure, from the cloud, from other vehicles, can be combined with the material collected in the passenger compartment,” estimates Ianic Frisoni, project manager at Continental Automotive France. The driver of a vehicle that has suffered a heart attack may warn other vehicles while driving safely along the road.
It’s not the internal sensors that are missing … “Since the 1990s, Valeo has built 1.5 billion sensors, and many more should be built in the next five years,” said Antoine Laffe, director of research driving support at Valeo. And a new generation of devices is already available. “Our ultra-wideband radar waves travel through the seat, a device for measuring heart rate and respiration, which helps identify a child who is stressed or sleeping driver, a forgotten child in the back seat,” said Sam Young, Canadian managing director and co-founder. Company Xandar Kardian.
“Attached to a camera, our AI algorithms can detect a drowning driver. And we are continuing our research to identify drunk drivers,” said Gabby Gizderveld, marketing director of the Swedish company Smart Eye. Can even warn the police … technically, anything is possible …
Ever the best protected drivers and users
· From March 2018, European law requires new car models to be equipped with eCall technology, automatic calls and ground transmission in the event of an accident.
July 2022: Newer models must have an ISA (Intelligent Speed Assistant) system that, when exceeding the speed limit, emits an audible signal, a vibration, or gently pushes the accelerator pedal backwards or automatically slows down.
July 2022: Newer models must have a socket, among other things, that allows a car ignition interlock breathalyzer to be easily installed; Driver drowsiness and attention loss warning device; In contrast a barrier detection system, etc.
Euro NCAP (European New Car Assessment Program), an independent Brussels-based organization that classifies vehicles based on crash test resistance, has been awarding extra points to vehicles equipped with CPD (Child Presence) since 2022. Identification) The system is able to alert the owner or emergency services if a small child is left behind.
Risk of acceptance
From 2018 to 2020, Gustav-Eiffel University’s Laboratory of Applied Psychology and Ergonomics, which specializes in the study of urban and urbanization processes, conducted research on the acceptability of a clock attached to nearby infrastructure and vehicles that vibrates when pedestrians are in danger while crossing the road. “Experiments with the Life-Size Street Crossing Simulator have reduced the risky decisions, but on the part of certain users, there has been a phenomenon of psychological reactions: they have done the opposite of what was expected of them. “Some drivers are wary of using driving aids,” said Aurelie Domes, research director at Renault, confirmed by Jean-Paul Dry, director of Renault’s electric vehicle, Segment C (medium vehicle) and a Megane E-Tech with 26 driving assistance functions. Introduce 100% Electric ৷ Solution? Slowly introduce these new technologies to improve road safety, which will give them time to demonstrate their effectiveness.