LANEGUARD: SSLA-BASED REAL-TIME TRAFFIC SIGN AND LANE DETECTION SYSTEM

Abstract

Autonomous driving systems rely heavily on accurate and real-time perception of road environments, including lane markings and traffic signs, to ensure safety and navigation efficiency. Traditional detection methods often struggle in complex traffic scenarios, adverse weather conditions, and varying lighting environments. This study introduces LaneGuard, a real-time traffic sign and lane detection system leveraging the SSLA (Spatial–Semantic Layer Attention) framework to enhance detection accuracy and robustness. The system integrates convolutional neural networks with attention mechanisms to focus on critical features, enabling precise identification of lane boundaries and traffic signs under diverse conditions. Experimental evaluations on benchmark autonomous driving datasets demonstrate that LaneGuard achieves superior accuracy, lower false detection rates, and faster processing times compared to conventional methods. The results highlight the potential of SSLA-based detection systems in improving autonomous vehicle perception, safety, and overall driving performance.