Technical Analysis of Autonomous Driving in Complex Road Conditions Based on Multimodal Perception Fusion

Abstract

: With the continuous development of autonomous driving technology, single-modality perception has been unable to meet current development needs. Therefore, multi-modality perception fusion has become an important technology to ensure the safe driving of vehicles. This paper mainly analyzes the defects of single-modality perception and the feasibility of multi-modality perception in complex environments. In addition, this paper discusses the fusion methods and novel algorithms based on the data layer, feature layer, and decision-making layer of the auto drive system in recent years, and raises some questions. Utilizing these studies, this paper explains how multi-modality perception fusion overcomes the impact of complex environments on vehicle driving. Finally, this paper summarizes the current technological bottlenecks and limitations of autonomous driving and proposes future development directions and challenges.