Can AI End Car Crashes?
Over the last two decades, we have seen a huge shift in vehicle safety. Many passenger vehicles now come standard with rear-view cameras, park distance control systems, electronic brakeforce distribution systems, lane assist, and impact proximity sensors. Those simple elements that many of us take for granted are a part of the advancement in artificial intelligence (AI) that strives to keep us safer on the roads.
Road safety is more important now than ever before. While road accident fatalities have long been a cause for concern, with the World Health Organization (1) estimating 1.4 million annual deaths due to fatal road accidents, the recent spate of fatalities has sparked debate. The debate at hand: Can AI end car crashes?
What Caused the Recent Spike in Road Accidents?
Some have proposed that the mental strain of the pandemic is one of the key factors that has spurred on this spate of fatal road accidents. Pair this with the fact that people are becoming exceedingly distracted on the roads and not adhering to basic safety precautions – like wearing seatbelts – and it’s a recipe for disaster. This was noted by the US Department of Transportation which reported the highest bi-annual spike in road fatalities during the period of January-June of 2021.
The Future of Driving
AI, as we’ve already learned, has been implemented in vehicle safety implements for quite some time now, but we are now seeing a surge in interest around the topic. With smartphone apps that can track your driving and offer you rewards and other perks for good driving coming to the fore, experts are weighing in on just how much AI can advance our road safety.
While some of these experts have their sights set on self-driving cars and vehicle-to-vehicle communication avenues as a means to avoid fatal collisions, the European Union has other ideas in mind. Currently set to roll out across the EU in July of this year, the Intelligent Speed Assistance System (ISA) will actively monitor driver speeds and behaviors using a sophisticated but low-cost system of onboard cameras and sensors. As part of a drive to curb road fatalities across the region, the ISA system will be mandatory for all new cars that are registered for road use in the EU.
Meanwhile, Across the Pond in Australia
Acusensus is an Australia-based company that is taking the lead in this sphere. In 2019, their product, Heads-Up, took New South Wales by storm and was soon rolled out across the country and even abroad. The system’s onboard imagery feature records driver behavior which is then assessed to gauge the probability of traffic offenses. Within 24 months of deploying Heads-Up, New South Wales reported a more than 20% drop in road fatalities and a whopping 80% drop in mobile phone use while driving.
Now, the company’s “Intelligent Eyes” are being used in combination with machine learning to draw attention to erratic driver behavior. This behavior, more often than not, occurs beneath the dashboard and out of view of fellow drivers and traffic officers. The Intelligent Eyes come in the form of highly intuitive roadside cameras that monitor drivers’ body language and behaviors. They can also assess with a high degree of accuracy whether or not a driver is actively involved in driving behavior that poses a threat or risk to themselves and other road users. According to Acusensus’ VP of sales for North America, Mark Etzbach, the Intelligent Eyes can gauge:
- Driver distraction;
- Occupant restraint;
- Vehicle speed.
Using a brief flash that is invisible to the human eye, Acusensus’ Intelligent Eyes can get a clear view of driver behavior through their windshield.
What Inspired the Birth of Intelligent Eyes?
Sadly, the co-founder of Acusensus, Alexander Jannink, lost a friend to reckless driving in 2013. The man, a fellow software engineer, was mowed down while biking that year and did not survive the collision. The driver was later found to be detrimentally impaired as well as distracted while behind the wheel. This spurred Jannink to find ways to enforce driver responsibility and safety.
Heads-Up has now evolved into Heads-Up Real Time which is set to be rolled out across the United States in the coming year. As the name suggests, the images and data captured using Intelligent Eyes will be relayed to nearby traffic officials in real time, helping them intervene in what could turn into calamitous car crashes before they’ve even occurred.
The Governor’s Highway Safety Association’s Senior Director of External Affairs, Pam Shadel has hailed Heads-Up as a life-saving measure, stating that there is incredible potential for the system to drastically reduce road fatalities.
Being Watched Makes Us Behave Better
It’s been reported that regions where traffic enforcers are clearly visible showcase far lower rates of road fatalities. The simple truth of the matter is that we behave better when we are under the watchful eye of authorities. We are less prone to taking risks on the road because we are afraid of receiving a traffic violation. While we should be more focused on staying safe than evading punishment, it would appear that this level of self-responsibility is not the case for some. Unfortunately, those few put us all at risk on the road.
Intelligent Eyes make it so that there is technically a traffic officer on every street in a given state. This urges drivers to act responsibly and to take their safety as well as the safety of others very seriously. Now, because the system also records data, it is able to create categories and groups that indicate which roads are so-called “hotspots” for traffic infractions, prompting law enforcement to increase visibility in that area. It can also give city planning departments the insight that they need to address infrastructure in the area either in the form of signage, speedbumps, and traffic lights amongst others.
The Data Speaks for Itself
The collection of data in and around accident hotspots is providing welcomed relief to city planning departments and traffic officials as they’re able to now pre-empt which areas across certain states pose a risk as high accident zones. With Missouri recording some of the highest incidences of road fatalities in the country, Acusensus in conjunction with Mrs. Shadel’s office recorded a day-long valuation of one of the highest risk roads in the area. They noted that more than half of all drivers in that particular corridor were speeding while the use of mobile phones amongst drivers in the area was more than twice the national average. Impressed with the data produced by such experiments, neighboring Indiana is also planning on adopting a data-fed system of this nature to direct traffic officer deployment to hotspot roads.
From Paris to Barcelona and beyond, similar AI is being used to map traffic infractions. While, in some countries in the EU, smart cameras are being mandated for use within new vehicles, other countries have taken to rigging public transportation vehicles with cameras. The recordings from these cameras are then assessed to see where the potential for fatalities with both vehicles and pedestrians alike is the highest.
However, these are no ordinary cameras and the use of AI helps officials make sense of the immense amount of data that streams in. For those who have decried the use of cameras, citing an infringement on their right to privacy, the videos are almost always deleted once the data has been processed – unless otherwise needed to prove certain grievous infractions. This, however, is not a comfort for some experts as they believe the level of intrusion into our privacy in order to keep us safe can spiral out of control if not kept in check.
The next step in ensuring that this data doesn’t just end up in an encrypted vault is for there to be data sharing amongst governments. Marketplaces would need to come to the fore to drive this sharing initiative and to allow for the procurement of quality data.
The ultimate goal for many nations around the world is something known as Vision Zero – the idea to bring road fatalities and grave injuries around the world to zero by using a series of passenger, driver, and pedestrian safety measures that form a safety net or protective barrier. In essence, if one measure fails, there are many other safety layers in place to ensure that a fatality never occurs. Safety data capturing and sharing is a huge step in the direction of the initiative that was addressed in Sweden in the 90s.
A Long Way to Go
There is no denying that AI of this nature can assist in the detection and prevention of dangerous driver behaviors, but we are still a long way from having AI that is capable of assessing pedestrian intent via body language and relaying that to a driver or traffic officer in real time to prevent an accident. As the driver is still the one behind the wheel, their discipline and concentration on the road are still the optimal preventative measures for fatal car crashes.
At the end of the day, the consensus seems to be that the human element of driving cannot be removed from the equation altogether at this point, nor should it be.
The short answer to the question we posed earlier is “yes!” AI can end fatalities on the road, but probably not on its own.