There is a strong demand for high fidelity sensor models which are capable of simulating realistic automotive sensor perception of Radar, LiDAR and camera sensors in real time, in order to validate advanced driving assistance functions like lane change assist (LCA), automated emergency breaking (AEB), or even path planning virtually. For central data fusion the sensor models need to deliver realistic, artificial sensor raw data. In especially, optical sensors are heavily influenced by precipitation, fog and sun irradiance. However, most LiDAR models lack the capability of replicating the impact of specific weather characteristics. Furthermore, there is – in contrast to numerous publicly available LiDAR datasets – a strong lack of datasets which are annotated with quantitative weather data such as the precipitation rate or meteorological visibility in order to develop and validate such models. Hence, within this work, an automated infrastructure is setup to measure time-correlated LiDAR and weather data to develop and calibrate weather models. The effects of varying precipitation rates on an automotive Flash LiDAR system is demonstrated based on in-field measurements and a respective modeling methodology is developed. Based on the in-field measurement data, raw data LiDAR models can be developed which augment virtual LiDAR data obtained from raytracing capable driving simulation suits as well as real data, recorded under ideal weather conditions.

With current advances in automated driving, optical sensors like cameras and LiDARs are playing an increasingly important role in modern driver assistance systems. However, these sensors face challenges from adverse weather effects like fog and precipitation, which significantly degrade the sensor performance due to scattering effects in its optical path. Consequently, major efforts are being made to understand, model, and mitigate these effects. In this work, the reverse research question is investigated, demonstrating that these measurement effects can be exploited to predict occurring weather conditions by using state-of-the-art deep learning mechanisms. In order to do so, a variety of models have been developed and trained on a recorded multiseason dataset and benchmarked with respect to performance, model size, and required computational resources, showing that especially modern vision transformers achieve remarkable results in distinguishing up to 15 precipitation classes with an accuracy of 84.41% and predicting the corresponding precipitation rate with a mean absolute error of less than 0.47 mm/h, solely based on measurement noise. Therefore, this research may contribute to a cost-effective solution for characterizing precipitation with a commercial Flash LiDAR sensor, which can be implemented as a lightweight vehicle software feature to issue advanced driver warnings, adapt driving dynamics, or serve as a data quality measure for adaptive data preprocessing and fusion.