Shihan’s research

Estimating freight movement activity in urban areas is crucial for freight transport planning, modelling, and network congestion management. Reliable freight origin-destination (OD) information is vital for strategic transport models used in future infrastructure investments and evaluating the logistics implications of economic development and city expansion.

Shihan’s research aims to develop methods that integrate multiple data sources, including video and traffic counts, to estimate freight OD activities. The focus is on leveraging machine learning and video analytics to automate the post-processing of captured data.

By developing new insights into freight journeys, we can enhance our understanding of travel patterns, improve network management, and identify potential freight routes for safe and efficient deliveries.

In her research, Shihan will conduct a field trial using a multi-camera system strategically deployed across a section of established freight routes within Melbourne to test the feasibility of the proposed methods.

What questions will this research answer?

The core question of Shihan’s research is how to improve the feasibility and effectiveness regarding the integration of multiple data sources to estimate freight OD activities. Specifically, the research seeks to answer:

1. How can machine learning and video analytics be used to automate the post-processing of captured data for more accurate freight OD estimation?
2. What is the optimal framework for integrating and fusing data from various sources to enhance the accuracy and reliability of OD information?
3. How can these methods be applied to improve freight transport planning, modelling, and network congestion management in urban areas of Australia?

Shihan produced a poster to illustrate her research. Download a copy of the poster at: Development and evaluation of video analytics solutions for freight origin-destination estimation

How far along is Shihan in her PhD?

In the first year of Shihan’s PhD, she primarily focused on conducting a systematic literature review on computer vision and machine learning-based freight transport applications. This provided her with a comprehensive understanding of how cutting-edge technologies have been utilised in the freight transport area. As the installation of roadside surveillance cameras increased and data availability improved, she saw the potential to apply video analytics to freight surveillance, a topic that was scarcely discussed.

Additionally, the research team has prepared and submitted a conference paper (currently under review) that proposes a vision data augmentation method based on Generative Adversarial Networks (GANs) to enhance freight detection with low-quality images. Simultaneously, Shihan is working on a survey paper that summarises the findings of the literature review.

What obstacles did Shihan overcome? How did she do it?

Although Shihan has experience in computer vision, she initially felt confused and unfamiliar with the project background of freight transportation and the concept of origin-destination (OD) estimation. With the guidance and suggestions from her supervisors, she refined her research direction from vehicle re-identification to computer vision applications in road freight transportation. This shift provided her with a comprehensive understanding of road freight transport and the existing computer vision-based applications in this area, ensuring that her research is highly relevant to the project.

In addition to research challenges, Shihan found that pursuing a PhD is more like a lifestyle. Building a PhD-adapted routine is essential, especially for maintaining a work-family balance. She developed valuable skills in effective time management, health management, and using advanced digital tools to improve productivity, all of which are beneficial for lifelong success.

What does Shihan see as her next move upon completing her PhD?

Upon completing her PhD, Shihan plans to engage in ongoing collaborations, presentations, and other outreach activities to share her research outcomes and promote their potential applications. She is always enthusiastic about bridging academic outcomes to real-life scenarios.

A word from PhD supervisor, Professor Hussein Dia

Why is this PhD important to investigate?

This work is important to advance our understanding of freight movement activity in urban areas, which is an important consideration in freight transport planning, modelling and network congestion management. Developing new insights into freight journeys can enhance our understanding of travel patterns, network management, and identification of potential freight routes for safe and efficient deliveries.

Current approaches to freight movement estimation, and also general traffic origin-destination estimation, relies on manual and semi-automated methods that are costly and time consuming to complete. The key approaches today for freight movement determination includes roadside manual OD surveys (which are costly to conduct and require considerable human and financial resources), or using technologies such as cameras, automatic number plate recognition, Bluetooth scanners, weigh-in-motion sensors, inductive loop detectors as well as in-vehicle sensors, mobile phones and GPS for vehicle tracking. All of these methods require significant post-processing of collected data and have been reported to have low estimation accuracy. However, out of all these methods, infrastructure-side solutions have been found to be the most reliable and cost-effective methods.

By developing fully a fully automated system that leverages the latest advances in AI, this work will investigate and develop cost-effective machine vision solutions with the potential to be scaled to large-scale transport networks. In particular, the approach adopted in this PhD study will be focused on infrastructure-side solutions, including a network of low-cost cameras to be installed at strategic locations within the study area, supplemented by other sensor data. To demonstrate the feasibility of the approach, this PhD will undertake a field trial and PoC study on a confined network in Melbourne.

What are the major challenges to overcome in the field?

The major challenges to be overcome include development of AI-based machine vision solutions that can operate with a high degree of accuracy under all conditions (e.g. day, night and adverse weather conditions).

Another challenge will be to develop a data fusion framework for manipulating and integrating data from different sources (e.g. cameras, pavement detectors and road-side sensors) and develop robust methodologies for their use in origin-destination estimations

Where might this work lead in the (near and far) future?

Successful completion of this work will result in cost-effective technology solutions that can automate the collection of freight movement data and make it easier and faster to draw insights about freight activity in urban areas.

Widespread deployment of these solution can provide road agencies and other stakeholders with valuable insights about the origins and destinations of freight trips, routes used, average travel times and distances travelled in addition to other data that can help understand the road safety aspects of freight operations and how these can be mitigated in future transport policies.

Contact Shihan

If you’d like to contact Shihan about her research, contact can be made via her LinkedIn profile.