SWEET: A Realistic Multiwavelength 3D Simulator for Automotive Perceptive Sensors in Foggy Conditions

Abstract

Improving the reliability of automotive perceptive sensors in degraded weather conditions, including fog, is an important issue for road safety and the development of automated driving. Cerema has designed the PAVIN platform reproducing fog and rain conditions to evaluate optical automotive sensor performance under these conditions. In order to increase the variety of scenarios and technologies under test, the use of digital simulation becomes a major asset. The purpose of this paper is to revive the debate around the realism of the various models underlying the numerical methods. The simulation of the radiative transfer equation by Monte Carlo methods and by simplified noise models is examined. The results of this paper show some gaps in foggy scenes between the ray-tracing method, which is considered to be the most realistic, and simple models for contrast evaluation, which can have a particularly strong impact on obstacle detection algorithms.

A Comprehensive Exploration of Fidelity Quantification in Computer-Generated Images

“…SWEET [ 24 ] (Simulating Weather for Intelligent Transportation Systems) is a Monte Carlo simulator developed by the French road planning institute CEREMA. It serves as a research-oriented and physically based simulator designed for internal use.…”

Qualification of the PAVIN Fog and Rain Platform and Its Digital Twin for the Evaluation of a Pedestrian Detector in Fog

“…Since pedestrians have different clothing, genders, and shapes, a random selection of Np pedestrians from each accessory size sub-list has been repeated 100 times with Np = [2,5,10,15,20,25,33,42,50]. The number of pedestrians per sub-list is presented in Table 1.…”

“…For this, one solution is to use numerical simulation. Many numerical simulators dedicated to autonomous vehicles exist [15][16][17][18]. Most offer variants regarding pedestrians, environments, or weather, but only a few are calibrated [15] against real-world conditions, to our knowledge.…”

“…This model, called the Koschmieder model [28], consists of reproducing the visual effect of fog on fog-free images by applying a reduction of contrast. It is commonly used in today's simulators [16][17][18], but some studies show that it has important limitations [15]. This type of validation remains rare in the literature [15].…”

“…It is commonly used in today's simulators [16][17][18], but some studies show that it has important limitations [15]. This type of validation remains rare in the literature [15]. The Koschmieder model is used here in the hardware-in-the-loop mode, using the images of the camera without fog, to add fog to it.…”