Simulating Realistic LiDAR Data Under Adverse Weather for Autonomous Vehicles: A Physics-Informed Learning Approach

Vivek Anand; Bharat Lohani; Rakesh Mishra; Gaurav Pandey

cs.ROApr 1, 2026

Simulating Realistic LiDAR Data Under Adverse Weather for Autonomous Vehicles: A Physics-Informed Learning Approach

Authors: Vivek Anand, Bharat Lohani, Rakesh Mishra, Gaurav Pandey

Organizations: Geoinformatics Group, Department of Civil Engineering, Indian Institute of Technology Kanpur, India · Department of Geodesy and Geomatics Engineering, University of New Brunswick, NB, Canada · Department of Engineering Technology & Industrial Distribution (ETID) of Texas A&M University

Abstract

Accurate LiDAR simulation is crucial for autonomous driving, especially under adverse weather conditions. Existing methods struggle to capture the complex interactions between LiDAR signals and atmospheric phenomena, leading to unrealistic representations. This paper presents a physics-informed learning framework (PICWGAN) for generating realistic LiDAR data under adverse weather conditions. By integrating physicsdriven constraints for modeling signal attenuation and geometryconsistent degradations into a physics-informed learning pipeline, the proposed method reduces the sim-to-real gap. Evaluations on real-world datasets (CADC for snow, Boreas for rain) and the VoxelScape dataset show that our approach closely mimics realworld intensity patterns. Quantitative metrics, including MSE, SSIM, KL divergence, and Wasserstein distance, demonstrate statistically consistent intensity distributions. Additionally, models trained on data enhanced by our framework outperform baselines in downstream 3D object detection, achieving performance comparable to models trained on real-world data. These results highlight the effectiveness of the proposed approach in improving the realism of LiDAR data and enabling robust perception under adverse weather conditions.

Figures & tables

Fig. 1 : Simulation-to-Real Gap: LiDAR point cloud with intensity information: a) Simulated Domain—VoxelScape data augmented with snow using a physics-based method [ 11 ] ; b) Real Domain—CADC data collected in snowy conditions. Objects marked with A (vegetation), B (metallic pole), C (car), and D (road surface). The significant difference in intensity distribution arises because the physics-based simulation fails to capture intensity attenuation due to adverse weather conditions like snow.

Fig. 2 : Physical Modalities:(a) The range is calculated using the Euclidean distance between the LiDAR sensor and the point, while the incidence angle is determined using the dot product of the LiDAR ray direction vector and surface normal vectors. (b) Material reflectance is obtained by mapping semantic labels in the LiDAR data to spectral reflectance values from NASA’s ECOSTRESS Library, based on identified materials.

Fig. 3 : PICWGAN Training framework: LiDAR point clouds, projected onto a spherical surface, are input into our proposed PICWGAN architecture to generate LiDAR intensity under adverse weather conditions.

Fig. 4 : Architecture: Generator and Discriminator architecture.

Fig. 5 : Qualitative Analysis - 3D: Visual assessment of the back-projected 3D point cloud with intensity information by our PICWGAN architecture trained on the snow weather CADC dataset. (a) Physics-based simulated snow-weather LiDAR point cloud from the VoxelScape dataset; (b) Same point cloud (from a) after domain adaptation by our PICWGAN; (c) Ground truth LiDAR point cloud from the CADC dataset used for training the PICWGAN.

Fig. 6 : Qualitative Analysis - 3D: Visual assessment of the back-projected 3D point cloud with intensity information by our PICWGAN architecture trained on the rain weather Boreas dataset. (a) Physics-based simulated rain-weather LiDAR point cloud from the VoxelScape dataset; (b) Same point cloud (from a) after domain adaptation by our PICWGAN; (c) Ground truth LiDAR point cloud from the Boreas dataset used for training the PICWGAN.

Fig. 7 : Probability Distribution Function for Same-Domain LiDAR Intensity Transfer on a) Snow Weather CADC dataset b) Rain Weather Boreas dataset between the Ground Truth and our PICWGAN Generated Intensity: PI-CGAN( CADC $\rightarrow$ CADC ) and PI-CGAN( Boreas $\rightarrow$ Boreas ).

Fig. 8 : Error Histogram of the prediction errors between the Ground Truth and our PICWGAN Generated Intensity for Same-Domain LiDAR Intensity Transfer on a) Snow Weather CADC dataset and b) Rain Weather Boreas dataset

Condition	Method	Data Combination	Total Frames
Snow	Baseline	VS-Phy	4000
		VS-Phy + CADC	2000+2000
		CADC	4000
	Proposed	VS-PICWGAN	4000
	Proposed	VS-PICWGAN + CADC	2000+2000
Rain	Baseline	VS-Phy	4000

TABLE I : Training Data Combinations for Snow and Rain Conditions

Class	Data (Snow)	Snow				Data (Rain)	Rain
		3D (IoU - 0.7)	3D (IoU - 0.5)	BEV (IoU - 0.7)	BEV (IoU - 0.5)		3D (IoU - 0.7)	3D (IoU - 0.5)	BEV (IoU - 0.7)	BEV (IoU - 0.5)
Car	VS-Phy	4.08	10.58	4.42	12.42	VS-Phy	18.89	33.67	20.45	36.34
	VS-PICWGAN	5.98	14.98	6.60	16.60	VS-PICWGAN	24.23	37.89	26.78	40.78
	VS-Phy + CADC	30.35	38.35	31.68	44.68	VS-Phy + KITTI-Rain	43.67	62.34	46.23	64.38
	VS-PICWGAN + CADC	37.73	42.73	38.04	47.04	VS-PICWGAN + KITTI-Rain	46.89	65.78	49.43	68.34
	CADC [4000 Frames]	45.18	50.98	47.60	53.60	KITTI-Rain [4000 Frames]	48.90	69.78	51.34	72.34

TABLE II : Performance comparison of PC-RCNN model trained on different dataset combinations with and without Physics-Informed Cycle-Consistent Weather GAN-generated (VS-PICWGAN) intensities for snow and rain, evaluated using BEV and 3D IoU at 0.5 and 0.7 thresholds, where higher IoU values indicate stricter matching criteria.

Explore similar work

ReaLiTy and LADS: A Unified Framework and Dataset Suite for LiDAR Adaptation Across Sensors and Adverse Weather Conditions
Apr 11, 2026Vivek Anand, Bharat Lohani, Rakesh Mishra +1Autonomous Driving Datasets Autonomous Driving Simulation
Geoinformatics Group, Department of Civil Engineering, Indian Institute of Technology Kanpur, India · Department of Geodesy and Geomatics Engineering, University of New Brunswick, NB, Canada · Department of Engineering Technology & Industrial Distribution (ETID) of Texas A&M University
LIORNet: Self-Supervised LiDAR Snow Removal Framework for Autonomous Driving under Adverse Weather Conditions
Mar 20, 2026Ji-il Park, Inwook ShimAutonomous Driving Perception Point Cloud Learning
Ministry of National Defense, 22, Itaewon-ro, Yongsan-gu, Seoul, Republic of Korea. · Department of Smart Mobility Engineering, Inha University, Republic of Korea.
Open-Weather Robust 3D Detection via Dual-Critic Diffusion Alignment
Jul 2, 2026Shuyao Li, Chuanxing Geng, Heyang Sun +2Lidar-Based 3D Object Detection Domain Generalization
Nanjing University of Aeronautics and Astronautics, Nanjing, China · MIIT Key Laboratory of Pattern Analysis and Machine Intelligence, Nanjing, China

Apr 11, 2026J/K move · Enter open · S save

ReaLiTy and LADS: A Unified Framework and Dataset Suite for LiDAR Adaptation Across Sensors and Adverse Weather Conditions

Vivek Anand, Bharat Lohani, Rakesh Mishra, Gaurav Pandey

Geoinformatics Group, Department of Civil Engineering, Indian Institute of Technology Kanpur, India · Department of Geodesy and Geomatics Engineering, University of New Brunswick, NB, Canada · Department of Engineering Technology & Industrial Distribution (ETID) of Texas A&M University

Autonomous Driving Datasets Autonomous Driving Simulation Unsupervised Domain Adaptation Point Cloud Learning

Open paper AskPDF arXiv

Reliable LiDAR perception requires robustness across sensors, environments, and adverse weather. However, existing datasets rarely provide physically consistent observations of the same scene under varying sensor configurations and weather conditions, limiting systematic analysis of domain shifts. This work presents ReaLiTy, a unified physics-informed framework that transforms LiDAR data to match target sensor specifications and weather conditions. The framework integrates physically grounded cues with a learning-based module to generate realistic intensity patterns, while a physics-based weather model introduces consistent geometric and radiometric degradations. Building on this framework, we introduce the LiDAR Adaptation Dataset Suite (LADS), a collection of physically consistent, transformation-ready point clouds with one-to-one correspondence to original datasets. Experiments demonstrate improved cross-domain consistency and realistic weather effects. ReaLiTy and LADS provide a reproducible foundation for studying LiDAR adaptation and simulation-driven perception in intelligent transportation systems.

Abstract

Explore similar work

ReaLiTy and LADS: A Unified Framework and Dataset Suite for LiDAR Adaptation Across Sensors and Adverse Weather Conditions

LIORNet: Self-Supervised LiDAR Snow Removal Framework for Autonomous Driving under Adverse Weather Conditions

Open-Weather Robust 3D Detection via Dual-Critic Diffusion Alignment