The autonomous car is not as wonderful as it seems: there are more than 150 serious problems still to be solved

The second Conference on Technology and Road Safety – organised by Seguridad Vital (RTVE), FESVIAL, and Carglass – revealed a great dose of reality about the rapid evolution that autonomous mobility has undergone in recent years.

During the event, the president of FESVIAL and professor of road safety, Luis Montoro, sought to curb the high expectations surrounding the autonomous car. According to the professor, many issues still need to be resolved before level 5 autonomous driving can be introduced onto public roads, and even then, it might not be profitable even if it were technically possible.

According to an article in Technology Review, there are as many as 150 serious unresolved problems “We live in a dangerous era, with many inventions and little reflection. Autonomous cars cannot be considered solely in technological terms, because these technologies affect the environment, people, the economy and society. Before reaching level 5 autonomous driving, much remains to be done technologically, but above all legally, personally and socially,” states Luis Montoro.

In an article published by the magazine Technology Review – belonging to MIT (Massachusetts Institute of Technology) – up to 150 serious problems that the autonomous car must face are listed, for which there is still no concrete solution. During his presentation, the president of FESVIAL listed some of these issues: “What insurance will it have? At what age will it be possible to use it?, Who will be responsible in an accident?, Who will decide the moral dilemma in a complicated situation? What training will we have? Will millions of professional drivers disappear?, Will there be traffic police and fines? Who will oversee the maintenance of the laser, radar, camera and satellite systems in autonomous cars?” the professor of road safety pointed out.

Besides the issues that will need to be resolved in the future, there are many factors that currently limit the safety of autonomous mobility. “Today, serious failures are recorded in night driving, or in rain, snow or fog. Current autonomous systems need time to analyse and react to some of the thousands of situations they may encounter; sometimes up to 7 seconds to think and react. There are also problems recognising road signs and markings if they are not perfectly maintained. There are serious limitations in recognising pedestrians and cyclists, for example if a pedestrian crouches at a zebra crossing…” declares the president of FESVIAL.

Once the problems are solved, would the autonomous car become profitable for manufacturers? In addition to the above, Montoro listed a series of possible risks that could hinder the use of this technology: “There is a risk of hacking and interference between cars. The car’s own systems may also pose a threat to safety. Another serious problem is real-time communication with the satellites that guide the car; if there is a brief disconnection, an accident may occur. Or if a camera gets dirty, for example from bird droppings, the car may face serious problems. With road markings it has already been shown that a laser pointer and a light reflection are enough to deceive autonomous cars,” notes the professor of road safety.

The magazine Technology Review states that it is currently impossible to program vehicles that predict what other cars or pedestrians are going to do. For that artificial intelligence would be needed. “According to many experts, level 5 autonomous cars must be proactive, creative and anticipatory, something impossible to program today,” asserts Luis Montoro.

To top it all off, the professor of road safety questioned whether the 100% autonomous vehicle would ever be profitable. “There is still a lot to be done to reach level 5, although nobody doubts it will be achieved. That said, these are very complex, expensive, vulnerable technologies that will require such a level of maintenance that we will have to consider whether implementation is viable and profitable, even if it is technically possible,” concluded Luis Montoro.