Self-driving cars: Will they conquer the world ?!

Robotic cars are one of the applications of mobile robotics, which is why the technologies used to develop them are the same technologies that researchers have been developing for decades, with an important difference: the acceptable margin of error in the case of robotic cars is close to zero. This is not currently supported by technologies developed for research purposes.

How do we deal with this matter? A logical start is a review of the main assumptions on which all this research is based, and there we may find what helps us. One of the main assumptions is that the environment in which the robot (or vehicle) moves does not contain any amount of intelligence or reactivity, so the robot must do everything on its own by relying on its own sensing capabilities.

For example, if we assume that the robot we are talking about is a car approaching an intersection, what we would expect of the car is that it will be able to recognize a traffic light, stop with a red light, and travel with a green light. Even for a mission as simple (on the face of it), it is in fact impossible to build a system that can recognize a traffic light under all conditions with 100% efficiency. The chances are that the red light will always be green, or the green light will be red.

Now, what if we assume the environment is smart?

Continuing with the previous example, if we assume that the car is moving in a smart environment, there will be no need to build a system to recognize the traffic signal from the ground up, because the traffic light itself will tell the car what to do. The scenario of the interaction between the traffic light and the car would be as follows: a car enters the range of a traffic light, and communication between them begins. The car informs the traffic light of its current speed, so the traffic light calculates the remaining time for the light to turn from green to red, and accordingly it calculates the recommended deceleration speed so that the car can stop at the appropriate time, and sends the information to the car. The car receives this information, and based on traffic conditions, surrounding cars decide what to do.

The previous example illustrates how many of the difficult technical challenges associated with developing mobile robots turn into relatively simple problems simply by assuming that the environment is smart enough to “help” the robot (or car) move through it.

Since we don't have that environment yet, it has to be built. Modernizing city infrastructure in line with the requirements of robotic cars and converting it into IoT-enabled infrastructure is the first step towards transforming robotic cars into everyday reality. The importance of this is as important as building railroads to trains - would trains have become a daily means of transportation without the railways?

The second challenge is a legislative challenge. Even if we modernize urban infrastructure and develop exemplary technologies for robotic cars, following current traffic laws will lead to accidents. Traffic laws themselves are what will lead to accidents!

The current traffic laws were developed from the ground up as recommendations to humans and not as orders to a machine, and so the language in which these laws were formulated is human ambiguity, in addition to containing logical contradictions. While humans can deal with ambiguous language (and they can ignore laws in the first place), this is not the case with machines.

Without reformulating traffic laws on a mathematical basis devoid of ambiguity - as is the case with programming languages, for example - it will be impossible to stand on a single standard for what those laws are, because the field will be open for each developer of robotic car technologies to interpret them differently, which will be represented in the end An obstacle towards their popularity.

So modernizing the infrastructure, reformulating traffic laws. As we can see, these matters require direct government intervention, because their scope exceeds the capabilities and powers of the private sector. Governments that move first to act in these directions will later be proud of being the first to turn robotic cars into an everyday reality.