cCDI 2022 Seed Funding Applications

IoT can have environments with devices, users, and applications from different application domains. These environments tend to be heterogeneous, as each application can have different purposes. The efficient, secure, and privacy-preserving storage and sharing of context information is essential for critical infrastructure. For instance, context sharing enables IoT-enabled systems to have a common view of the context information, which facilitates interoperability, management, and optimisation of processing efforts. This project aligns with the cCDI’s mission statement and involves close collaboration with Cisco. It lays the foundation for a long-term project with the industry partner that involves developing practical solutions enabling the secure, efficient, and smart adoption of IoT systems in Australia’s critical infrastructure.

Grain breeding is a key component of the basic infrastructure of the grains industry and the whole agriculture industry. AI/machine learning (ML) driven digitalized breeding process, including experimental design, sampling, and analytics, is promising to make breeding more accurate, robust, and efficient. We will conduct research on the fundamental and specific ML and computer vision technologies that can work in dynamic agriculture environments with diverse streaming data for breeding. Specifically, we will study the causal ML and continual learning techniques in the context of this application. This project will result in high-quality interdisciplinary research works and establish the impact on the digital infrastructure in the agriculture domain for cCDI. It will also lead to potential external funding.

Prevention of cyber-attacks on critical infrastructure through highly robust operating system technology based on the locally developed seL4 microkernel, world’s most secure operating system kernel. The technology is open-source infrastructure that applies to a wide range of critical systems and will lead to strong industry engagement.

The objective of insect surveillance is that the knowledge of the density and location of target species allows the rapid and localised control of these insects. However, existing solutions do not scale due to operational costs and reliance on human experts. Automatic insect surveillance has the potential to revolutionise agriculture, providing real-time heat maps of the insect pest densities on the farmer's smartphone. However, such a vision will require much more than Al­ equipped traps, as such devices must operate reliably in field conditions with minimal existing digital infrastructure. Such infrastructure development is essential to the success of automatic pest surveillance technology.

Accurately estimating urban energy consumption is critical for electricity networks to meet the consumer demand. Urban energy consumption is dominated by occupant behaviours. Yet, occupancy information is not always available due to privacy concerns. This project explores how human mobility information could be used as proxy for building occupancy data in energy modelling. Adverse events directly affect human mobility (E.g., climate changes, pandemics). Shifts in mobility could be exploited to infer extreme consumption patterns. This ensures resilience of energy distribution infrastructure to abnormal events. Hence, the project outcomes directly impact the proper functioning of the energy industry.

IEEE 2030.5 protocol is used for communication between DER end devices and the utility level energy management system. IEEE 2030.5 server code is not yet publicly available. This project aims to provide the open-source implementation of IEEE2030.5 server and client codes in the RUST language. Other researchers will be able to use this code for their testing and development of DER related projects. The project targets the impact theme to develop future capabilities to support industry in improving critical digital infrastructure. This work is complementary and can benefit Australia and the global community.

Device drivers are the leading source of operating system vulnerabilities, accounting for 39% of CVE publicly disclosed vulnerabilities in Linux over the last five years. Thus, they are a prime target for cyber-attacks on critical infrastructures: Formal verification can eliminate such vulnerabilities, yet previous attempts at device driver verification have failed to scale beyond toy examples. The purpose of this project is to use an approach based on a novel verification-friendly systems programming language, dubbed Pancake, to make formal verification scale to real-world device drivers. This will enable the development of embedded systems with higher assurance against cyberattacks.

The project addressed the need for high-precision simulations and visualization of light-based loT solutions during their design and development phase. Light-loT provides efficient sensing and communications for many critical infrastructures, such as factories, tunnels, nuclear power plants, underground mines, etc., where radio frequency is often not reliable due to interference from other radio instrumentations and congestion in the limited and narrow radio frequency bands.
On the other hand, lighting is already available for illumination purposes, which can be exploited by Light-loT, bringing value added benefits for critical digital infrastructure. Finally, in many critical digital infrastructures, manual battery replacements of thousands of loT devices are challenging and costly. Light-based loT can harvest energy from the light energy to self-power themselves without requiring battery replacements. Investigators will develop strategies that can obtain correct values of light using machine learning techniques.

The UNSW Blockchain Interest Group is an interdisciplinary group of UNSW staff and students working on blockchain systems and their applications. We have recently submitted a proposal (requesting $1.4M) to the NSW Government Tech Central Research Infrastructure fund to establish a NSW Blockchain Research and Innovation Hub which would provide a focal point for a network of research, government and industry partners to conduct research and education and foster commercialisation in blockchain in NSW. We anticipate building towards applications for longer term grant and centre funding by way of Linkage, ITRH or CRC-P schemes. The present proposal is for preliminary work towards establishment of the Hub and subsequent grant applications by deepening engagement with the partners. It integrates blockchain related proposals submitted in the call for cCDI Seed Funding.

There are no reliable techniques to detect structural failures with sensors on a moving vehicle. There are many challenges to this, such as accounting for bridge-vehicle interaction, calibration and configuration of the sensors, small amount of training data and limited computational resources. This project aims to develop a reliable technique to overcome these challenges. The cCDI is interested in this project because we intend to use machine learning to develop the software tools for signal processing on a moving vehicle. Our techniques are adaptable to other critical digital infrastructures.