The aim of this project is to develop a transformative robust and scalable autonomous landing system for drones. This is the critical missing technology needed to unleash exponential growth in a potentially enormous drone delivery industry by enabling a multitude of applications to deliver goods and supplies via drones to a wide range of destinations in Australia and the world in a timely, flexible and accurate manner. Such an autonomous landing solution would revolutionise drone technology, and propel Australia to the forefront of technology innovation. This project would benefit not only large scale delivery by drone in urban and suburban areas of Australia but also long distance delivery via drone to remote areas of Australia.
This project aims to enhance the reliability and safety of emerging self-driving vehicles, through a framework that supports the validation and verification of autonomous driving systems. This project expects to generate new knowledge in areas of software engineering, intelligent transport, and machine learning, using a multi-disciplinary research combining expertise from various fields. Expected outcomes of this project are a family of new context-aware techniques to verify and validate complex behaviours in autonomous driving. This should provide significant benefits, such as safe autonomous driving systems and the improved journey experience and security for road users.
In this project, Macquarie University will collaborate with UTS and COEUSTECH to develop an innovative safety-preserving ecosystem for autonomous driving. This system will not only be adopted by COEUSTECH customers (automotive companies) to secure their latest autonomous driving models, but also be commercialised as a toolset that can be plugged into existing autonomous vehicles to detect and prevent malicious attacks on autonomous driving models. The project will lead to two innovations: in theory design an attack detection and prevention ecosystem for autonomous driving and in application implement a safety analysis toolset for industry-scale autonomous systems.
In this project, we aim to develop testing and validation technologies to enhance the trustworthiness of machine learning models trained on edge devices. As a new form of AI models training, we first will use new software development patterns to effectively integrate data from multiple modalities for training AI-Human Interaction models. As a new form of AI testing, we plan to adopt data-driven metamorphic relations to thoroughly test AI-Human Interaction models to detect models' safety issues. As a new form of runtime verification for AI models, we will develop an online-learning anomaly detection architecture that can detect anomalies in the AI-Human interaction models at runtime. As a case study, we will create a variety of AI models for driver fatigue detection using semi-autonomous or autonomous vehicles.
We successfully applied our solutions in Advanced Driver Assistance System. We developed a fatigue driving detection system deployed in the vehicle. Graphene-based sensors are attached on the steering wheel. By analysing ECG and EMG signals collected from the sensors, the system knows whether the driver is concentrated enough on driving.
In 2018, we cooperated with Alibaba establishing a large-scale e-commerce decision-making intelligence model. The model was applied in the largest shopping day in China, 11 Nov. The model helped Alibaba reduce about 40% waste of coupons and vouchers, meanwhile, increasing final sales ratio by 25%.