Advancing Real-Time Object Detection for Autonomous Vehicles with YOLOv11

The safe and efficient operation of autonomous vehicles relies heavily on robust and real-time object detection systems. This paper presents YOLOv11, the latest advancement in the YOLO (You Only Look Once) series, optimized for detecting critical objects in autonomous driving scenarios, including vehicles, pedestrians, and road signs. YOLOv11 incorporates novel architectural enhancements, such as the C3k2 block for efficient feature extraction and the Cross-Stage Partial with Spatial Attention (C2PSA) module for enhanced spatial attention, enabling state-of-the-art accuracy and real-time performance. Evaluations conducted on the KITTI dataset, augmented with advanced data augmentation and multi-scale training techniques, demonstrate substantial improvements in precision, recall, and mean Average Precision (mAP) compared to its predecessors. Furthermore, the integration of YOLOv11 into autonomous systems leverages sensor fusion to deliver comprehensive perception capabilities, establishing its potential for real-world deployment. This study highlights YOLOv11’s contributions to advancing autonomous vehicle technology and provides valuable insights for developing robust, scalable, and efficient object detection frameworks for future applications.