Skip to main content
ICTD 2024 Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.

Abstract for presentation (Poster or Podium)

AI in Transportation

Optimizing Traffic Sensor Location Using Information Theory

Monday, June 17, 2024
1:30 PM - 2:45 PM ET
Room: Galleria
Abstract:

Traffic sensors play a crucial role in monitoring and assessing highway traffic at the network level. However, the installation of adequate sensors to ensure full network flow observation in real-world highway networks is often cost-prohibitive, necessitating the need for a strategic selection of optimal sensor locations under budget and resource constraints. This well-researched problem is commonly referred to as the Traffic Sensor Location Problem (TSLP).  In this study, we present a novel approach to address the TSLP by minimizing information loss within a joint feature embedding space. Our method considers both segment-level features and network topology to formulate the optimization problem. To effectively solve this optimization problem, we design and implement an evolutionary algorithm. This algorithm is further enhanced by a reduced solution space formed through physics-guided random walk, making it highly efficient in identifying optimal sensor locations for large-scale highway networks. We evaluate our proposed framework using the highway network for Savannah in Georgia, demonstrating its effectiveness in achieving optimal sensor placement. The results highlight the significant potential of our approach in enhancing traffic monitoring and management in real-world highway networks.

Learning Objectives:

Attendees can expect to learn the following from this session:
  • Learn how information theory is employed for formulating the sensor location optimization problem.
  • Learn modern methods for network topology embedding.
  • Learn how Genetic Algorithm can be adapted to solve the sensor location optimization problem with large real-world highway networks.