Reliable detection and classification of road users, including pedestrians, cars, bicycles, strollers, wheelchairs and scooters, are critical challenges for perception sensors in autonomous vehicles. These targets vary in size, shape, materials and motion, with pedestrians being particularly complex due to clothing, posture and body dynamics. Environmental factors and changing orientations alter sensor responses, affecting reflectivity and signal signatures. By analyzing features such as micro-Doppler effects, reflectivity and aspect-angle dependence across radar, lidar and camera systems, this study aims to improve detection accuracy and enable the development of highly realistic simulations to better model these diverse targets.

• Techniques for improving detection and realistic simulations of road targets
• Impact of body shape, size, gender and clothing on detection
• 360° RCS of some common road targets