| Summary: | 碩士 === 國立雲林科技大學 === 電子工程系 === 105 === As technology progressed rapidly in recent years, driving safety and automation have been gradually valued by the public and became the main development focus of automakers. However, whether it is before the process of official production or during the process of developing an algorithm in the lab, the proposed technology is required to simulate and verify using a great number of videos with various scenes. Although there are many products of advanced driver assistance system (ADAS), there is no complete development environment for ADAS developers to use when developing these driving safety system. Therefore, the purpose of this paper is to establish an effective, automatic and reliable ADAS development environment, and provide developers a simulation and verification platform which approximates the actual driving environment.
The development environment for advanced driver assistance system which this paper proposes includes five points. First: system collects environmental information of various scenes to offer a complete development environment. Second: system automatically analyzes a great amount of driving information for developers to use and evaluate and system enables to choose different environmental parameters to simulate according to the requirements. Third: the system enables to adjust the frame rate which algorithm simulates according to the actual driving situation, and automatically analyzes and verifies the system accuracy of each function. Fourth: the system offers users a web platform to view all kinds of statistics after analyzing and simulating. Finally: the system saves the time of implementing manual simulation and data analysis, and improves the development speed of algorithms.
In addition, the paper also develops millimeter-wave radar fusion technology and applies 24GHz millimeter-wave radar at low-cost which domestic manufacturers research, develop and design themselves. The detection distance of this system is 46 meters and the system integrates received information after signal processing with image information of forward collision warning system (FCWS) which was developed by the laboratory. By the interaction of millimeter-wave radar with image, the system corrects the condition of generating errors when image calculating distance and also improves the shortcomings of not knowing directionality and radar unable to recognize targets because of the low development costs which millimeter-wave radar used in this paper.
|
|---|
