Would self-driving cars solve the road safety problem?

Roadly, Inc.
5 min readJul 12, 2022


Indeed, would they?

A well known and unfortunate fact is that human error accounts for 90% road accidents. So there you go, remove the driver out of the driving wheel, and the problem is solved for the most part! But if we dig deeper into the subject, things become more complicated pretty soon.

What makes the “90%” statement misleading is the fact that any road accident is caused by a multitude of factors, systematic and random. A tragic sequence of events that caused irretrievable consequences.

Let’s draw an imaginary scenario. After a long and busy day Sally is returning home in her old but well maintained Toyota Corolla made in the early 2000s. She feels a bit sleepy, but she believes that should not be a problem for an experienced driver like her. Due to the road maintenance, she takes an alternative route — not the one she is used to. It is already dark and somewhat foggy outside.

At the same time Jake decids to take an evening promenade to the grocery store nearby, it is pretty close , and there is no need to put much thought into it. He went outside in his jeans and dark leather jacket. The grocery store is right across the road. Although the traffic is mild at this time and seemingly safe to cross the road at a random spot, there are no crosswalks nearby. Unfortunately it took too long for Sally to notice the pedestrian on her way and react, and she hit Jake.

Later the traffic police inspector concluded the following reasons for the accident:

  • A bit drowsy driver after long work day;
  • Dark road — late evening with poor lighting on the road;
  • Slippery road — it took longer for the car to stop;
  • J-walking (crossing the road at an inappropriate place);
  • Dark and not reflective clothes of Jake;
  • Tall bushes on the roadside that created an obstacle for Sally and Jake to see each other well in advance;
  • Lack of active safety system in the car that could make the car automatically stop before hitting the object.

Removing 1 or 2 factors from the equation (i.e. minimizing their impact somehow) may have prevented that particular accident but would it significantly improve the road safety on a larger scale?

In general, what kind of solutions could eliminate these negative factors to minimize the number of accidents?

Road and street design. It is biased to place full accountability on the pedestrian or the driver in this case who, despite making some general and situational errors, did not intentionally break the traffic law. Ideally, road design should “forgive” mistakes that people make, and therefore reduce the “human factor” influence to the minimum. When transportation engineers collect enough statistics of road accidents in specific spot, they often can identify whether there is a dangerous pattern that leads to a higher number of accidents. There are some examples of road design flaws that could be corrected or even prevented by experienced traffic safety engineers.For instance, a road could seem attractive for speeding because of its width and straightness. However, it could also have a large pedestrian traffic crossing it because of the density of residential buildings in the area. Another example could be that there might be obstacles in the area that limit the view of the driver and do not allow the driver to see the pedestrian until the very last moment. If there is a case of systematic J-walking, it often suggests that the number of existing crosswalks is insufficient or their locations are not optimal for the pedestrian traffic flow.

Legislation and speed enforcement. Both of these categories are closely related to the road design. There is a common consensus that the car speed correlates with the chance of an accident. However, due to historic legacy or lack of political will, the speed limits are often unreasonably high or not well enforced and therefore not being followed.

Road condition. We mentioned that it was too dark for Sally to see the danger on time and prevent the accident. That could be caused by lack of lighting infrastructure or its insufficient maintenance. Unsurprisingly the slipperiness of the road could drastically increase the chance of having a road accident. A less known fact is that the road slipperiness may be caused by so-called “asphalt bleeding” or other road defects. That could be minimized by timely road inspection and maintenance. Our core mission at Roadly is to provide municipalities and transportation agencies with the solution that enables low-cost and scalable inspection of road assets.. I will share more details on our solution in my next articles.

Education and social ads. Emphasize the danger of drowsy distracted and intoxicated driving. Educate the drivers to assess their physical condition, level of tiredness and stress and its impact on their driving capabilities. Address and communicate the consequences and danger of J-walking and the importance of wearing retro-reflective materials at night when there is less visibility on the roads.

And finally the driving automation. Automatic Emergency Braking systems are effective tools to prevent road crashes or minimize the casualties if a crash is unavoidable. AEB is an active safety system and a Level 0 ADAS feature. It is considered to be the foundation of the safety-related functionalities of driving automation.

The list of potential solutions for road safety doesn’t end there. It could include encouraging alternative means of transportation (bikes, public transport), using map data to choose safer routes etc.

With all that being said, it becomes obvious that road safety is a complex problem and there is no single “silver bullet” solution to fix it overnight. For instance, the Automatic Emergency Braking system of Tesla (well-known and beloved leader in self-driving) is apparently having difficulties at night conditions due to the absence of lidar. The best way to make the roads safer is to apply the systematic approach for road safety and not simply wait when self-driving cars come and completely solve the car accident problem. Moreover, having crumbling road infrastructure, immature and inconsistent traffic rules, chaotic driving culture and other unsolved problems would risk the very fact of successful deployment of autonomous cars in your region.

If you work at a department of transportation or generally you want to know more about Roadly, contact us to see how we can help you solve the challenge of timely road inspection.

About Roadly: With Roadly cities and transportation agencies can conduct Roadway Condition Surveys, ensure the most up-to-date asset and pavement data, build PCI, monitor the condition of the road and get notified of on-road distresses such as potholes, cracks, utility patching, sealants, etc. at pixel-level granularity.



Roadly, Inc.

Automate asset mapping using computer vision