Autonomous vehicles: state of the art, problems and prospects

Numerous economic players, both in the automotive industry and digital economy players such as Google, are actively working on the development of autonomous vehicles.
The implementation of such vehicles, however, poses a number of technical, viability and safety issues.

First of all, it is necessary to take into consideration that the generalization of autonomous vehicles will not happen instantly, but more or less gradually. This phase is probably the most delicate since it will coexist in the same place, the same road vehicles that are autonomous but also vehicles that are not and can not communicate with other vehicles. This situation is even more important in urban areas where vehicles must share the road with other users: pedestrians, cyclists.

The transition to autonomous vehicles also requires that vehicles be able to perceive their immediate surroundings and thus go through the networking of vehicles in order to share certain data (position, speed, etc.) with other vehicles. . This involves several constraints. The first is the need to put in place a standardized language that allows a vehicle to communicate with others, even though they do not come from the same manufacturer. How could vehicles communicate with each other if they do not speak the same language? The second is related to the security, confidentiality and reliability of the data. Indeed, the networking of vehicles communicating with each other requires an infrastructure and therefore to pass data between the vehicles. Since these vehicles are mobile, this data will require the establishment of network infrastructures allowing the transmission of wireless data, such as GSM, 4G, etc., and thus raise questions concerning data security and confidentiality. so that: 1) the data is not accessible, modifiable by a third party; 2) the authenticity and confidentiality of the data is ensured. Could a malicious person get into the system and have access to the positions, data of all vehicles circulating around the world? This problem also arises when we know that the main actors working on the development of autonomous vehicles are players in the digital economy, who could see an interest in collecting, analyzing and possibly reselling the data and thus in knowing the slightest movements of 7 billion people in the world. This will require the introduction of specific legislation. But will all the countries agree to put in place such legislation when it could allow governments, especially in a context of terrorist risk, to know all the movements of all their inhabitants, or even all motorists of the planet.

Finally, the question of road infrastructures arises, since they will have to adapt to autonomisation so that the data (speed limits, for example) are reliably transmitted to the vehicles.