Huixin Liu

Research

Subject

Simulating Knowledge and Experience-Based Driver Behaviours for LGVs Based on Agent-Based Model

First and second supervisors

Dr Sarah Wise
Prof Bani Anvari, Department of Civil, Environmental and Geomatic Engineering, Faculty of Engineering Sciences

Abstract

Transport is responsible for a third of carbon emissions in the UK, with freight producing one third of this. Within freight transport, Light Goods Vehicles (LGVs) are one of the biggest sources of emissions, not only because of numbers but also because of the nature of delivery. Moreover, the needs of registered LGV drivers have grown rapidly since 2005. To cut transport emissions by 2050, decarbonising LGVs emissions is both urgent and potentially influenceable. However, current decarbonising LGVs research only focuses on freight vehicles themselves, lacking consideration of the role of driver behaviour. Drivers’ experience levels can be a very important factor influencing the delivery efficiency and, consequently, decarbonizing. This research aims to explore the difference of in/experienced LGVs drivers’ behaviour in the last-mile delivery process.

Using a case study based in London, this research will use a bottom-up simulation technique (agent-based modelling, or ABM) to recreate drivers’ behaviour in LGVs delivery. A Unity virtual simulation environment will set up in London based on Open Street Map (OSM) to help collecting the data. Delivery company drivers will be invited to use the driver simulator to conduct the experiment in at the PEARL Laboratory, UCL. All the participants will be drawn from delivery companies. A framework will be built to simulate the agents (drivers) behaviour based on the experiment results, exploring the LGV delivery process relative to the differences of drivers experience level. I anticipate comparing travel distance and time, vehicle idling time, carbon emissions, driving experience, education level, the impact of different consumer neighbourhoods, and vehicle types. To improve delivery efficiency, an efficient delivery framework will be set up based on the simulation results, which will provide support for delivery companies to train inexperienced drivers to improve delivery efficiency and help decarbonise freight.

Biography

Huixin studied in Logistics Management during her bachelor’s degree. She changed her major to Design Science in her first Master’s degree at Beijing Jiaotong University, and achieved her MRes degree in Spatial Data Science and Visualisation at CASA in 2021. Huixin began her PhD study at CASA in Sep 2022.

Publications

Huixin Liu and Sarah Wise (2023) ‘Agent-Based Modelling and Disease: Demonstrating the Role of Human Remains in Epidemic Outbreaks (Short Paper)’, International Conference Geographic Information Science [Preprint]. Available at: https://doi.org/10.4230/lipics.giscience.2023.48.