XI Международная студенческая научная конференция Студенческий научный форум - 2019

We live in the age of information technology which requires constant changes and innovations in any sphere of society. The automotive industry is not an exception, it has undergone rapid changes over the past decade. For example, England and France have decided to abandon the internal combustion engine due to environmental problems and by 2040 it is planned to introduce a ban on the sale of new cars with gasoline and diesel engines.

It is impossible not to notice that every day in the world there appear more and more cars equipped with various systems of assistance to the driver. A forecast has been made about investment in the development of unmanned vehicles since the beginning of 2050. But such developments require an accurate mechanism of unmanned vehicle. An accident involving a Tesla car in 2016 shocked people and shattered their confidence in the drone mechanism. The cause of this accident was a system error, it could not recognize the truck because of the high ground clearance between the wheels of the truck and the trailer, which led to a collision and fatal accident. From this we can conclude that not all unmanned vehicle systems work as needed and require improvements and tests. From January to November 2017 in the Tyumen region collisions made 40% of road accidents. The cost of repair of a vehicle after the “insurance shot” (according to the magazine Autoreview) excess on average 100,000 rubles.

The use of forward collision warning system (FCWS) can significantly reduce the number of road accidents, increase the capacity of roads and increase the average speed of the vehicle without the risk of an accident.

The European organization Euro NCAP has published the results of tests of automatic braking systems before an obstacle (Autonomous Emergency Braking, AEB), which from January this year will become a mandatory part of the crash test program. The first new type of tests was exercised on eight cars: Mercedes-Benz E-Class, Volvo V40 and XC60, Mitsubishi Outlander, Fiat 500L, Ford Focus, Honda Civic and Volkswagen Golf. Forward collision avoidance systems in Euro NCAP are divided into two categories: systems running on urban (AEB City) and high speeds (Inter-Urban AEB). During the tests the machines are "pushed" twice: when the driver reacts to the electronics warnings and brakes independently, and then completely relying on the operation of the systems without touching the brake pedal. According to the results of the tests in each of the categories assessment is made from zero to three points. Then, depending on the total score, the cars get an overall score: "Good", "Enough" or "Bad".

"Urban" automatic braking systems are tested in front of a moving obstacle at speeds of up to 50 kilometers per hour (Mercedes-Benz E-Class, Volvo V40 and XC60, Mitsubishi Outlander, Fiat 500L and Ford Focus). The best of them was "Mercedes", which earned a maximum score of three points. Systems from the category of "country" were tested at speeds of up to 80 kilometers per hour (Mercedes-Benz E-Class, Volvo V40, Volkswagen Golf, Mitsubishi Outlander and Honda Civic). Here the best was also a German sedan, which earned 2.7 points. The worst result showed the Civic, which received only 0.44 points for the system's Collision Mitigation Brake System. Previously, similar tests were conducted by the American insurance Institute for road safety (IIHS). Of 74 models which took part in the tests, only seven were awarded the highest score for the operation of automatic braking systems: Cadillac ATS and SRX, Mercedes-Benz C-Class, Subaru Legacy and Outback, as well as Volvo S60 and XC60.

The system reports potentially dangerous situations on the road, reducing the likelihood of getting into an accident. One of the means of technical vision is lidar. Lidar is a laser radar that uses ultrasonic pulses to detect objects in front of the car, as well as their speed and distance to them. Since the lidar works best at close range, the camera is mounted in the windscreen and the radar in the front bumper, which work together at high speed as part of a collision warning system with full braking capability. The radar can detect objects a few hundred meters in front of the car, but it cannot determine the type of object. Here the camera is to identify the object and determine whether it is a problem or it can be ignored. Advanced systems (like Subaru's EyeSight) use two black-and-white cameras that work as eyes to triangulate speed and distance to the car ahead of you. They scan the situation every 0.1 seconds, looking for a contrast between the background and any vertical surface. The software is programmed to recognize several types of images, such as the back of a car, a motorcycle, a bicycle and even a pedestrian.

An electronic control unit (ECU) is used for the proper operation of the system, which checks its performance and processes the data coming from the sensors. It analyzes the distance to the potentially dangerous object in front and, in the event of a dangerous approach, notifies the driver of a possible collision. The main scientific value of this algorithm is the block of the decision-making system. You have to take into account many facts for this block to work properly, otherwise the consequences will be very serious, or it will simply interfere with the driver when driving.

The practical significance of the research lies in the further possibility of its application in practice in an improved form, correcting the shortcomings and modernizing the principle of the system. That in turn will remove the possibility of system errors, thereby reducing the accident rate on urban and suburban roads.