I don't know when, people gave the "tram problem" that tortured the conscience of countless people to the driverless car. So in various industry forums and conferences, we can see that many people have faced the problem of “who is going to hit†under special circumstances. On the surface, this seems to be the "people worrying about the sky", but the deeper reasons behind it cannot be ignored: on the one hand, the value of autonomous driving technology and the development trend that will eventually replace human drivers are generally recognized. On the other hand, due to the widespread existence of information asymmetry, everyone is basically ignorant of new technologies, let alone trust. As a wide-ranging participant in social transportation, self-driving cars must also gain the trust of people. Therefore, the relative transparency of this new technology in decision-making and execution must also be put on the agenda. In fact, many companies are working hard and exploring this aspect. A few days ago, Yiou Auto interviewed Jack Weast, senior chief engineer and chief system architect of Intel's driverless solutions. In an hour-long conversation, he launched the responsibility-sensitive security model RSS (Responsibility SensiTIve Safety), which was launched by Intel's auto-driving technology company Mobileye, which is described by many media including E-Car. The mode of operation of the model. In the exchange, Jack Weast defined RSS as the “Asimov's Three Laws†for self-driving cars. For the foreseeable future, a security model like this will play an important role in the development and promotion of autonomous vehicles. Before we elaborate on RSS, let's talk back to the "tram problem." This ethical issue was coined by the philosopher Philippa Foot in his 1967 paper Abortion Issues and Dogma Dual Influence, originally used for utilitarianism. The latter's point of view is that most of the ethical decisions are based on the principle of “providing the greatest benefit to the most peopleâ€, that is, “the value of human life cannot be measured by quantity aloneâ€. In other words, the problem is built on the assumptions from start to finish. In fact, unless someone on the rails is determined to commit suicide, no one will see the tram coming and not avoiding. (The current tram problem has produced countless variants, but the bottom layer is similar logic) People have been afflicted by this problem for a long time since this puzzle was raised for more than 50 years. Therefore, after the auto-driving technology with artificial intelligence has entered the reality, everyone has also used this problem to test new technologies. However, using a topic within a philosophical category to test autonomous vehicles in the real world is obviously not of practical value. If this problem has a little value in the development of autonomous driving technology, then it is: Autopilot technology should establish a reasonable premise for itself to avoid the embarrassing situation of dilemma. When people talk about the advantages of self-driving cars, "safety" must be listed first. However, as long as there are other road users in the road besides artificial intelligence, 100% security cannot exist. But what exactly is the level of "safety" of autonomous vehicles, and the outside world is not clearly defined. In this regard, RSS gives two clear goals for the safety of self-driving cars: 1. Autonomously driving a car may only be involved in an accident and will not become the responsible party of the accident; 2. The self-driving car must respond correctly to the accident. According to statistics, the current incidence of casualties caused by human-powered vehicles is 10^-6 times/hour. And if you want to make the society widely accept new technology, it is more reasonable to reduce the accident rate of self-driving cars by three orders of magnitude to 10^-6 times/hour. But if this is done through drive tests, the autopilot system needs to accumulate more than 3 billion miles of test data. At the same time, once the software is upgraded, the corresponding test is likely to need to be repeated. And if you simply rely on stacked sensors and supercomputers, then autonomous vehicles can only be mass-produced because of the high cost. Therefore, autonomous vehicles need to be from the bottom of the planning process to avoid becoming the responsible party of the accident. In the face of an accident, the self-driving car should also be given the right response. 4 common sense, do not let auto-driving cars become the responsible party for the accident At the specific implementation level, RSS defines four “common senses†to prevent autonomous vehicles from becoming responsible for traffic accidents. In fact, these common senses have covered most road accident scenarios: Rear-end collision is not the responsibility of the preceding vehicle. The rear vehicle needs to maintain a safe distance. The rear-end collision caused by sudden parallel connection is in the front of the vehicle. It cannot be used because the "right of the road" of the self-driving vehicle is high. Avoid accidents; pay attention to road blind spots and avoid accidents caused by being blocked by other vehicles. At the same time, RSS also stipulates that the response mode of autonomous vehicles in the face of unexpected situations should be consistent with humans. At the same time, vehicles can not stop when they can not meet the situation, but also ensure the efficiency of road traffic. For these four common senses, Mobileye is defined in detail in RSS. First, the R&D staff proposed a formula for calculating the minimum safe distance. These parameters and constants also need to be dynamically adjusted according to different factors such as road conditions, weather, and driving environment. In fact, this time Intel cooperated with Tsinghua University and the Chinese Academy of Sciences to determine the parameters and constants that autonomous vehicles need to use when driving in China according to the particularity of China's road environment. (Parameter: Vf front speed, Vr rear speed, reaction time Ï, minimum brake acceleration αmin, brake, maximum brake acceleration αmax, brake and maximum acceleration αmax, accel) Second, define the issue of the right to land. To put it simply, the judgment of RSS on the right and the right of the road is similar to what we humans learned in the driving school, that is, “the rear car makes the front carâ€, the “auxiliary road makes the main roadâ€, “turns straight to go straightâ€, “the parallel line allows the rear car†". However, RSS also emphasizes that autonomous vehicles should not only strictly obey the right of way, but also judge the behavior of other vehicles and avoid traffic participants who violate their road rights. After all, even if it can be exempted from accidents, it is king to sever the accident from the root cause. Finally, in the face of pedestrians and road blind spots, RSS requires pedestrians to have higher road rights than cars in complex roads such as residential areas. In addition, in the face of scenes such as occlusion, RSS stipulates that the speed of the vehicle should be lower than that of pedestrians. Therefore, even if a collision occurs, it will not cause serious casualties. Looking back at the RSS model, we can see that this technology is designed to solve the problem of the relationship between self-driving cars and other traffic participants. That is how to protect other people, vehicles, and how to protect themselves. At the end of the exchange, Jack Weast told E-Car, the RSS-responsible security model, and the self-driving car, just like Asimov's three laws for conscious robots. "RSS is actually the basic rule for autonomous vehicles. ". When the famous science fiction writer Isaac Asimov proposed the "three laws of robots", he might not think that his ideas in the book would be given priority in the automatic driving scene. However, the rules that are only recognized by one company are at most corporate norms. With the development and landing of autonomous vehicles on a global scale, all parties including governments, car companies, technology companies and chip manufacturers need to join the development and formulation of the basic rules of autonomous vehicles. After all, a production car needs to be able to operate on a global scale, as is autonomous vehicles. Pv Connector,Mc4 Connector,Solar Panel Connectors,Solar Connector Sowell Electric CO., LTD. , https://www.sowellsolar.com