"* Preconditions: Radar detection and tracking algorithm (including ML based postprocessing/classification) together with video detection + recognition algorithm shall achieve 99% accuracy and speed measurement error within 2kph/2% (bigger of) under poor lighting situation at night and cloudy environment (at most Y% light Intensity) * Input conditions / steps: images from simulator/real-world with the reduced lighting including corresponding radar data * Expected result: The system achieves to detect and count the object in the radar data (based on point cloud), or video images (based on LP detection), or their combination correctly, meeting the expected accuracy."