Improving Mean Average Precision (mAP) of Camera and Radar Fusion Network for Object Detection Using Radar Augmentation

With the recent availability of automotive data such as the Nuscenes dataset, sensor fusion for object detection has gained popularity in the field of autonomous driving. One of the most common sensor combinations is camera and radar. While vision has been studied for many years, the recent availability of radar data has added novelty to the field of object detection. This study explores the concept of radar data augmentation using methods inspired by commonly used image augmentation techniques. The model was trained on the Nuscenes mini dataset and a mean average precision (mAP) of 0.45 was achieved, 20.98% greater than the baseline results of the network.