Because the sensor has its own limitations in the design, a single sensor can not meet the precise perception of various working conditions, if you want the vehicle to run smoothly in various environments, you need to use multi-sensor fusion technology, which is also the key technology of the large category of environmental perception technology.
According to data provided by the World Health Organization in May 2017, about 1.25 million people die from road traffic accidents every year, which equates to 3,500 people dying in traffic accidents worldwide every day. Media reporters learned from the National Work Conference on Safety Production that there were 60,000 traffic accidents in China in 2016, and the number of deaths reached 41,000.
Autonomous driving, infographic
Car accidents are fierce, and this sentence is vividly reflected in these data. In order to improve driving safety, there are two general directions at present, one is to strengthen traffic control and force drivers to drive safely with high-pressure policies; the other is to get cars out of human operations, which is also what global car companies and technology companies are currently doing.
Getting the car out of human hands is, in technical terms, autonomous driving or driverless driving. As a kind of intelligent car that realizes unmanned driving through computer systems, self-driving cars rely on artificial intelligence, visual computing, radar, monitoring devices and global positioning systems to work together to allow computers to automatically and safely operate motor vehicles without any human active operation.
The key technologies of autonomous driving can be divided into four parts: environmental perception, behavioral decision-making, path planning and motion control.
Perception technology
As the first step of environmental perception, it is the collection and processing of environmental information and in-vehicle information, which is the basis and premise for intelligent vehicles to drive autonomously. Obtaining ambient information involves road boundary detection, vehicle detection, pedestrian detection and other technologies, that is, sensor technology, and the sensors used are generally laser rangefinders, video cameras, vehicle radar, speed and acceleration sensors and so on. Of course, this part is also the most expensive part of a smart vehicle.
But the perception technology is not to install a million radar, engage in a few high-definition cameras can be, because the sensor in the design of the time has its own limitations, a single sensor can not meet the precise perception of various working conditions, want the vehicle to run smoothly in a variety of environments, you need to use multi-sensor fusion technology, the technology is also the key technology of this large category of environmental perception technology, the current domestic and foreign major gaps are also concentrated in the multi-sensor fusion.
Decision-making techniques
After completing the perception part, the next thing that needs to be done is to make decisions and judgments based on the information obtained by the perception system, determine the appropriate working model, and formulate the corresponding control strategy. This part functions similarly to giving the vehicle the corresponding task. For example, in lane keeping, lane departure warning, distance keeping, obstacle warning and other systems, it is necessary to predict the state of the car and other vehicles, lanes, pedestrians, etc. in the future for a period of time, advanced decision theory includes fuzzy reasoning, reinforcement learning, neural network and Bayesian network technology.
Path planning
Intelligent vehicles have driving tasks, the path planning of intelligent vehicles is to carry out environmental information perception and determine the location of the vehicle in the environment, according to a certain search algorithm, to find out a passable path, and then to achieve the autonomous navigation of intelligent vehicles.
The method of path planning can be divided into two categories according to the completeness of the intelligent vehicle's working environment information:
A global route planning method based on complete environmental information; for example, if there are many roads from Shanghai to Beijing, the planning office is a global planning as a driving route. Such as raster method, viewable method, topology method, free space method, neural network method and other static path planning algorithms.
For example, there will be other vehicles or obstacles on the route from Shanghai to Beijing that is globally planned, and if you want to avoid these obstacles or vehicles, you need to turn to adjust the lane, which is local path planning. The methods of local path planning include: artificial potential field method, vector domain histogram method, virtual force field method, genetic algorithm and other dynamic path planning algorithms.
Motion control
After planning the driving path, the next step is to control the vehicle to follow the desired trajectory, which is what the motion control part needs to complete.
Motion control includes lateral control and longitudinal control, simply put, horizontal control is steering control, longitudinal control is speed control, now more research is lateral control, the methods used mainly include synovial control, fuzzy control, neural network control, optimal control, adaptive control and pure tracking control.
In layman's terms, the lateral control gives a speed, through the control of steering to achieve the purpose of the vehicle along the predetermined trajectory; and the longitudinal control purpose is to meet the speed requirements of the vehicle in the process of driving, and sometimes it is necessary to cooperate with the lateral control to meet the vehicle in the trajectory tracking at the same time, but also to meet the purpose of safety, stability and comfort. Because the vehicle is a particularly complex system, horizontal, longitudinal and vertical have coupling relationships, it is necessary to carry out horizontal, longitudinal, and even horizontal, vertical and vertical collaborative control of intelligent vehicles. Due to the complexity of its coupling relationship, the collaborative control technology of intelligent vehicle motion control is also the technical difficulty of this part.
If car companies and technology companies can make these four technologies work perfectly, then cars can drive on their own, and transportation will usher in the era of autonomous driving.
Reproduced from the network, the views in the text are only for sharing and exchange, do not represent the position of this number, such as copyright and other issues, please inform, we will deal with it in a timely manner.