1.1 Background of Research on PathRecognition System of Intelligent Smart Trash Vehicle

Thesmart unmanned vehicle is an integrated system integrating environmentalawareness, planning and decision-making, and multi-level assisted driving. Itintegrates computers, modern sensing, information fusion, wirelesscommunications, artificial intelligence, and automatic control. It is typical High-techcomplex. The intelligent unmanned vehicle can liberate the driver's hands andrealize the automatic driving of the vehicle, which saves people's time costand provides the society with a large amount of surplus labor.

Withthe advancement of science and technology and the substantial increase in thelevel of national material consumption, the production of various types ofgarbage in China has grown rapidly, and environmental issues have becomeincreasingly severe. The status quo of China's waste sorting and processingsystem is that it does not have a set of perfect models in the front-endgarbage classification and distribution, and it can solve the problem ofdifficulty in classification and collection and difficulty in classificationcollection, resulting in the later period. When carrying out garbage disposal,the workload of sorting is huge and the waste of resources is serious. Therational application of intelligent unmanned vehicle technology in the field ofwaste sorting and trans-shipment can provide a new solution for waste sortingand collection. The application of the pure electric intelligent garbagesorting transfer car can guide people to develop the living habits of garbagesorting, improve the efficiency of garbage sorting and treatment, and reducethe secondary environmental pollution of rubbish.

Asa key technology of intelligent unmanned vehicles, image-based path recognitiontechnology is an indispensable part of smart driving. The key to intelligentvehicle lane recognition is to ensure the reliability and real-time performanceof the recognition. To achieve this, in addition to recognizing the oppositeside of reliability and real-time performance, it is more important to realizethat they complement each other, that is, high reliability in a local area caneffectively eliminate interference factors, thereby reducing subsequent data.The amount of processing facilitates the improvement of the overall real-timeperformance; the overall high real-time performance can reduce the amount ofdata processing as a whole, thereby ensuring local reliability.

Thisproject intends to develop a low-cost, reliable visual path recognition systemfor unmanned intelligent garbage transfer vehicles to realize the function of pathrecognition and obstacle avoidance in the complex and changeable communityenvironment of unmanned intelligent garbage transfer vehicles. Ensure thereliability and real-time performance of the path identification system.

1.2Analysis of Current Status of Foreign Route Identification

Dueto the market demand, all countries in the world are making great efforts todevelop smart vehicles. As an important part of smart vehicles, manyresearchers from Europe, Japan, and other countries have studied the lanedeparture warning system earlier, and have accordingly made many importantachievements such as The AURORA system designed by Carnegie Mellon University,the Auto Vue system jointly developed by Iteris and Daimler Chrysler, andGoogle’s driverless smart car.

1AURORA system

In1997, Carnegie Mellon University successfully developed the AURORA system. Thereare many differences in the system, which is embodied in its components. Thesystem consists of three main parts: a color camera, a digitizer, and aportable Sun Sparc workstation. The color camera is mainly used to collect theimage of the lane marking line. The position of the color camera is placed onboth sides of the vehicle. The image captured by the camera is the lane markingline in the area of 1.5m-1.6m on both sides ofthe vehicle; the image of the lane marking line taken by the camera is thentaken. Transmitted to Sun Sparc workstations, Sun Sparc workstations processimages in real time. The system can detect the lane marking line in real timeand carry out tracking and early warning.

2Auto Vue System