Optimization of Autonomous Driving Vehicle Environment Perception and Decision Algorithm Based on Vehicle mounted Sensor Data
Download as PDF
DOI: 10.25236/icceme.2024.044
Corresponding Author
Yipin Fang
Abstract
In recent years, with the rapid development of artificial intelligence (AI) technology, AutoPilot System (APS) technology has become a research hotspot in the automotive field and is leading the wave of innovation in the automotive industry. In APS, environmental perception is the core component of APS, providing key information for subsequent vehicle control and decision-making. However, the real road traffic environment is complex and ever-changing, with significant differences in targets such as vehicles and pedestrians, and frequent dynamic changes. At the same time, visual sensors, as one of the main means of environmental perception, are easily affected by external factors such as lighting conditions and weather conditions, leading to a decrease in imaging quality under harsh conditions, which poses a huge challenge to environmental perception technology. Based on this, this article proposes an APS vehicle environment perception and decision-making algorithm based on onboard sensor data. This algorithm fully utilizes deep learning (DL) technology and achieves high-precision perception and understanding of the vehicle's surrounding environment by training neural network models. The simulation results show that the algorithm proposed in this paper has excellent performance and robustness in environmental perception and decision-making, providing strong support for the further development and commercial application of autonomous driving technology.
Keywords
Vehicle mounted sensor data; Autonomous driving vehicles; Environmental perception; Decision algorithm