Digitalizing the social contract makes roads safer

When Mobileye started designing safety concepts for autonomous vehicles (AVs), the first thing we did was to study the concepts and mechanisms of human road safety. We needed to build a framework that fully complies with human road safety systems so that autonomous vehicles can travel on the road with human-driven cars. In addition, we also needed to design a safer system to increase society's acceptance of autonomous vehicles.

In developing this system, we found that this framework, originally developed to solve the challenges of autonomous vehicles, can now also significantly improve road safety through its application in advanced driver assistance systems (ADAS). Currently, most road safety rules are informal, difficult to enforce, and customary, and this solution has digitized these road safety rules.

Loopholes in current traffic rules

The foundation of the existing road safety system is traffic rules, which are generally implemented by giving drivers clear and unambiguous instructions through road and roadside signals and signs such as traffic lights, stop signs, lane dividers, etc.

Nevertheless, traffic rules are still an uncertain system. Even if all traffic participants strictly abide by the rules, there is still a risk of traffic accidents. This is because if the traffic rules are too certain and require traffic lights at every intersection (not roundabouts), it would be necessary to draw solid lines on every lane, which would not only be costly but also lead to a significant reduction in traffic flow.

Although broken lines and yield signs can improve traffic efficiency, they also bring potential conflicts. When road conflicts occur, for example, when changing lanes or at a four-way stop, road users must negotiate with each other. If the negotiation is completely unregulated, then the outcome will depend entirely on the attitudes of traffic participants towards time utility and risk aversion. This is when the social contract comes into play.

The social contract for careful driving aims to fill the safety gap caused by the uncertainty of traffic rules. It can minimize the chance of time-sensitive conflicts and regulate the mutual negotiation between road users in some efficient ways, such as instructing traffic participants to keep a safe distance from the vehicle in front, drive carefully when visibility is limited, and give up the right of way when others claim it. This is a social contract, in the sense that we all support this unwritten set of rules because it is in everyone's interest.

Replacing traffic rules with social contracts can compensate for the consequences of violating traffic rules to a certain extent. For example, according to the social contract, if a vehicle in the opposite lane passes directly in front of a traffic participant, the latter can cross the solid line - as long as it does not lead to another violation of the social contract.

Although the social contract for careful driving plays a vital role in humanity's road safety system, it is also flawed. It is broad, has no clear definition of what is safe and appropriate, and its correct application depends on people's judgment at the time. As a result, errors in judgment are the main cause of accidents. The social contract is also almost impossible to enforce because identifying violations requires a detailed analysis of traffic conditions.

Digitizing the social contract in autonomous vehicles

Humans must understand this ambiguous and unmeasurable system while driving, but the decisions of autonomous vehicles must be clear and quantitative, so we need a more understandable explanation of the contract. This is exactly the premise of Mobileye's Responsibility Sensitive Safety Model (RSS): a clear, concise, parameterized, effective, real-time and traceable digital explanation of the social contract.

RSS has other important contributions to road safety. First, it is a mathematically validated social contract. If all traffic participants deploy RSS, then, assuming all other vehicle-related factors are working properly, the vehicle will not cause an accident due to poor decision-making. Second, it is completely explicit and quantitative, so it can help investigators understand how well different traffic participants abide by the digital social contract after an accident.

From people to self-driving cars, and back to people

What began as a responsibility for self-driving cars to understand our human road safety systems has now evolved into an undisputed opportunity to greatly improve them.

Now that we have a fully measurable, explainable, and executable safety model, we wondered: Why wait until self-driving cars are a reality to experience the road safety benefits of this new reality? Why not find a way to let human drivers benefit from RSS, the digital version of the social contract, now?

To achieve this goal, we designed the Vision Zero driver assistance system, which is specifically designed to meet mass market deployment and demand. The system uses preventive technology to help people avoid emergency maneuvers. Based on the principle of cautious driving, the system is able to provide preventive micro-interventions using a set of surround cameras and RSS framework. In addition, the system will benefit from the precise foresight of negotiation points ahead provided by many road users, as well as insights into dynamic road use patterns and road network security vulnerabilities.

This approach is very different from the "Vision Zero" strategies currently used in many scenarios. Those strategies focus on the "road slimming" approach widely adopted by cities around the world, choosing to deepen traffic rules with static and pre-set rules such as speed limits, speed bumps and physical barriers. These road restrictions make traffic rules more invasive, greatly affecting traffic efficiency and with questionable effectiveness.

The digitization of the social contract can help improve road safety and has great potential to increase traffic flow. With appropriate regulatory support, these digital principles for careful driving could eventually become formal, enforceable and binding contracts, thereby reducing the weaknesses of the current informal social contract.

The RSS framework is a digital solution to the social contract, which effectively solves the inherent flaws of the social contract. It also avoids the limitations of road slimming. Although RSS was originally designed for autonomous vehicles, we can now apply it to ADAS solutions and start to work from this moment. I believe this is the next revolution in ADAS. After going full circle, we are back to the concept of people.