Discovering How People Come and Go: Inferring Queue Dynamics and System-Level Time Series Data

Description:

This technology is a two-stage data-driven methodology that analyzes queue dynamics by leveraging time series data to infer arrival and service patterns. It utilizes a non-parametric screening stage followed by parametric estimation, enabling accurate analysis without needing detailed event-level data. Validated through simulations and real-world testing in scenarios such as parking garage occupancy, the framework offers a scalable and privacy-preserving solution for queue analysis.

Key Advantages:

• Operates effectively with minimal and aggregated time series data.
• Preserves privacy by avoiding invasive data collection methods.
• Low-cost and non-invasive compared to traditional queue monitoring techniques.
• Validated through simulation and real-world application.
• Adaptable to various types of queue-based systems.

Problems Solved:

• Inability to analyze queue dynamics without detailed event-level data.
• High costs and privacy concerns associated with invasive data collection.
• Lack of scalable solutions for real-time queue monitoring in diverse industries.
• Challenges in applying traditional methods to incomplete or aggregated datasets.

Market Applications:

• Smart parking systems for efficient space management.
• Logistics and supply chain operations optimization.
• Facility and service operations monitoring.
• Enterprise analytics for customer flow and service performance.
• Development of software toolkits for broader adoption and customization.