Towards Efficient Roadside LiDAR Deployment: A Fast Surrogate Metric Based on Entropy-Guided Visibility

TL;DR

This work tackles the challenge of efficiently deploying roadside LiDAR by introducing the Entropy-Guided Visibility Score (EGVS), a fast surrogate metric rooted in information theory. EGVS leverages a Traffic Probabilistic Occupancy Grid (TPOG) and ray tracing over empty frames to quantify how much information LiDAR beams contribute to detecting objects within a region of interest, accounting for both static occlusions and traffic dynamics. The authors validate EGVS against ground-truth 3D object detection performance (AP) using the AWSIM simulator with PointPillars, showing a strong, monotonic correlation across BEV and 3D views and enabling rapid evaluation (seconds per configuration). The approach reduces reliance on labor-intensive labeling and full perception models, supporting scalable optimization of LiDAR placements and paving the way for gradient-based search and real-world deployment in urban infrastructures.

Abstract

The deployment of roadside LiDAR sensors plays a crucial role in the development of Cooperative Intelligent Transport Systems (C-ITS). However, the high cost of LiDAR sensors necessitates efficient placement strategies to maximize detection performance. Traditional roadside LiDAR deployment methods rely on expert insight, making them time-consuming. Automating this process, however, demands extensive computation, as it requires not only visibility evaluation but also assessing detection performance across different LiDAR placements. To address this challenge, we propose a fast surrogate metric, the Entropy-Guided Visibility Score (EGVS), based on information gain to evaluate object detection performance in roadside LiDAR configurations. EGVS leverages Traffic Probabilistic Occupancy Grids (TPOG) to prioritize critical areas and employs entropy-based calculations to quantify the information captured by LiDAR beams. This eliminates the need for direct detection performance evaluation, which typically requires extensive labeling and computational resources. By integrating EGVS into the optimization process, we significantly accelerate the search for optimal LiDAR configurations. Experimental results using the AWSIM simulator demonstrate that EGVS strongly correlates with Average Precision (AP) scores and effectively predicts object detection performance. This approach offers a computationally efficient solution for roadside LiDAR deployment, facilitating scalable smart infrastructure development.

Figures (6)

• Figure 1: Overview of the proposed surrogate metric and its performance evaluation. Random traffic data collection is needed only once for each scenario for different roadside LiDAR placements inside the ROI.
• Figure 2: The testing environment in the Nishishinjuku map, which is based on the real landscape in Shinjuku Ward, Tokyo. The red box denotes the region of interest (ROI) with the dimensions of $50\,\mathrm{m} \times 100\,\mathrm{m} \times 5\,\mathrm{m}$.
• Figure 3: The traffic probabilistic occupancy grid (TPOG) from a top-down view. The occupancy probability of each grid is summed through the Z-axis for display.
• Figure 4: These two images demonstrate a case where occlusion reduces the surrogate score. Light-colored strips in Figure \ref{['fig:ray_tracing']} represent occlusions in the LiDAR view. Entropy in these voxels does not contribute to the surrogate metric score since no laser beams intersect them, leading to reduced detection rates when vehicles occupy these voxels. A 3D view of the LiDAR is shown in Figure \ref{['fig:occlusion']}.
• Figure 5: The view of the LiDAR is occluded by traffic lights, trees, and poles.