Object detection system plays a pivotal role in modern driver assistance. The integration of sensor technologies is crucial for advancing perception systems’ reliability, particularly in autonomous vehicles’ Advanced Driver Assistance Systems (ADAS) and surveillance applications. We define reliability as the model’s ability to consistently detect obstacles, regardless of weather or lighting conditions. Additionally, we assess accuracy by measuring the model’s precision in correctly identifying and localizing objects in its surroundings. This paper presents a comparative analysis of two object detection models: a RADAR and camera fusion RADAR Multiple Perspective Convolutional Neural Network (RAMP-CNN) model and the widely used camera-only You Only Look Once v8 (YOLOv8) model. The research investigates the effectiveness of the RAMP CNN architecture, specifically designed for processing both RADAR data and image data, for object detection in comparison to the established YOLOv8 model, which operates only on standard image data.