Author: Zen and the Art of Computer Programming
With the widespread popularity of mobile terminal devices such as smartphones and tablets, smart cars are attracting more and more attention. As a means of transportation that can bring humanity, the research and development of smart cars has always been a hot topic. In recent years, with the continuous emergence of smart cars, related news reports have emerged one after another. I believe that many readers have heard about the research and development of smart cars, and many manufacturers are developing their own smart car products.
As an important force in the smart car industry, artificial intelligence (AI) is an advanced technology used for task decision analysis, which can help cars understand the environment, identify objects and judge behaviors more intelligently. At present, most smart car systems perform decision-making analysis based on data collected by sensors, and the "car + cloud" model based on the cloud makes the overall system architecture of the car more complex and rich in functions. This article will introduce some experiences in the design and implementation of artificial intelligence system architecture for intelligent vehicles from the perspective of the overall architecture through pictures and texts.
2. Explanation of basic concepts and terms
2.1 Definition and characteristics of intelligent vehicles
Generally speaking, a smart car system consists of hardware components, control systems, sensing systems, information processing systems, navigation systems, safety systems, etc. Its system architecture can be divided into three types: fuzzy integration, precision integration, and distributed. In the fuzzy integrated system, the vehicle actively detects the surrounding environment, uses remote sensing, radar and other sensors to identify obstacles, and performs feedback control based on the distance and relative position of the obstacles to complete the autonomous driving task. The precision integrated system is also called the Perceptual Navigation System (PNAS), including navigation, map, control, information transmission and storage modules. It mainly completes vehicle position estimation, map construction, path planning, vehicle motion control and information interaction. In a distributed system, various subsystems are connected through network communication to form a complete vehicle system.
In particular, the definition of “intelligence” also includes wearable personal assistant robots and autonomous driving systems with self-learning capabilities. In addition, smart cars have many other features, such as on-demand charging, unlimited battery life, long-term parking, super-smart batteries, and more. Although these characteristics