SBIR Phase I: Applying Machine Learning to Mitigate Disruptions in Novel Relay Trucking Model

Connect Dynamics, Inc., Bentonville AR

Investigators

Abstract

The broader impacts of this SBIR Phase I project include a more sustainable and profitable trucking industry, enhanced conditions for the trucking workforce, improved public road safety, reduced emissions, and a path to future advancements using alternative energy and autonomous technologies. This project centers on a patented technology that pools shipments and matches tractors and drivers with trailers and cargo in a relay fashion that ensures equipment and cargo keep moving while drivers return home daily. Research has established the theoretical viability of applying relay models in long-haul trucking to solve equipment and human capacity issues. Using relay trucking could potentially double asset utilization, cut delivery times in half, and decrease the high cost of truck driver turnover. To date, no company has created a technology to implement scalable relays in American trucking. Customer discovery and research suggests this is due to the complexity of predicting and mitigating real-world trucking disruption events (e.g. traffic accidents, equipment breakdown etc.). The project aims to assess the feasibility of developing a machine learning (ML) based predictive analytics tool to make our relay technology resilient to disruptions in the driver scheduling and truck route optimization problem. A successful project would extend a relay trucking model beyond largely theoretical studies to build the first real-world application in the U.S. It aligns with NSF’s mission by transforming the trucking industry while creating better jobs for new and existing truckers and reducing carbon emissions. We estimate that resolving the inefficiencies resulting in the loss of $110 billion per year. This project aims to prove the feasibility of using machine learning (ML) techniques to build a predictive tool that integrates with our relay scheduling algorithms to effectively estimate the likelihood of occurrence of disruption events and provide actionable intelligence for deployment of mitigation strategies. A successful outcome will create resilient relay algorithms necessary to commercialize the first scalable relay trucking operations in the U.S. While the use of ML has been explored in predicting weather and traffic events in the recent past, no off-the-shelf ML tools are known to exist that can be utilized for this patented relay technology. This project leverages real-world proprietary operational data from large trucking companies, combined with public data sources and artificially populated datasets representing drivers and equipment informed by work with local partners at relay nodes. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.