Periodic Reporting for period 1 - INITIATE (Intelligent and Sustainable Aerial-Terrestrial IoT Networks)
Reporting period: 2022-01-01 to 2023-12-31
This project aims to address the communication and energy issues of the Internet-of- Things (IoT) when they are deployed for environmental disaster management. Through establishing an international and cross-sector consortium, the project tackles the limitations and stringent Quality-of-Service requirements that currently hinder IoT networks. The main objectives include designing an agile aerial-terrestrial network architecture with software-defined networking (SDN) and ultra-reliable network management, developing technologies for efficient wireless power and information transfer, creating lightweight AI models and online machine learning methods, and implementing smart network management strategies. Through these efforts, INITIATE will design a robust IoT network with advanced self-organization and healing capabilities for environmental monitoring. This system will bring in significant benefits for the society by enabling sustainable and robust environment observations, enabling real-time situation awareness for effective disaster management. The overall project will help reduce disaster impact and thus protect environment and save lives, contributing to the welfare of people in the EU and worldwide.
This project has completed the research and dissemination work planned for the first period. The research work has led to fruitful outcomes, dissemination events, and impactful results. Specifically,
WP1 focused on the project management and impact. The consortium has completed two annual plenary meetings and organized two focused workshops, as well as delivered numerous conference presentations, keynote speeches and exploitation events.
WP2 completed two use case studies focusing on landslide monitoring and oil pipeline monitoring, respectively. For monitoring landslide disasters and ensuring the safety of oil pipelines, IoT networks are highly effective and valuable for their abilities to provide comprehensive, efficient, and real-time data collection, transmission, processing, and analysis. WP2 designed a software-defined aerial-terrestrial network architecture and completed the definition of the functional components.
WP3 reviewed the energy needs in IoT networks for environmental monitoring, identified the key limitations and QoS requirements, and developed an effective machine learning algorithm to achieve dynamic power splitting, aimed at optimizing the power ratio for energy harvesting and traffic scheduling.
WP4 designed a lightweight AI model and online learning methods, including a streamlined object detector based on Convolutional Neural Networks (CNN) for embedded systems.
WP5 completed an exploratory analysis on IoT trust management techniques for security and reliability of interconnected devices. An effective approach for network management and anomaly detection was further developed for the IoT networks.
WP1 focused on the project management and impact. The consortium has completed two annual plenary meetings and organized two focused workshops, as well as delivered numerous conference presentations, keynote speeches and exploitation events.
WP2 completed two use case studies focusing on landslide monitoring and oil pipeline monitoring, respectively. For monitoring landslide disasters and ensuring the safety of oil pipelines, IoT networks are highly effective and valuable for their abilities to provide comprehensive, efficient, and real-time data collection, transmission, processing, and analysis. WP2 designed a software-defined aerial-terrestrial network architecture and completed the definition of the functional components.
WP3 reviewed the energy needs in IoT networks for environmental monitoring, identified the key limitations and QoS requirements, and developed an effective machine learning algorithm to achieve dynamic power splitting, aimed at optimizing the power ratio for energy harvesting and traffic scheduling.
WP4 designed a lightweight AI model and online learning methods, including a streamlined object detector based on Convolutional Neural Networks (CNN) for embedded systems.
WP5 completed an exploratory analysis on IoT trust management techniques for security and reliability of interconnected devices. An effective approach for network management and anomaly detection was further developed for the IoT networks.
This project aims to design an integrated aerial-terrestrial IoT network architecture that seamlessly integrates aerial systems and ground IoT networks through an innovative software-defined network (SDN) framework, where network resources are optimised in real time. To achieve this goal, the project will advance the state-of-the-arts in the following aspects:
1) Designing an efficient SDN aerial-terrestrial network architecture to optimise the performance of IoT networks.
2) Developing an efficient simultaneous power and information transfer approach.
3) Developing efficient compression, pruning and simplification methods to decrease the complexity and size of AI models while maintaining their performance.
4) Proposing a new approach for smart network management and anomaly detection.
1) Designing an efficient SDN aerial-terrestrial network architecture to optimise the performance of IoT networks.
2) Developing an efficient simultaneous power and information transfer approach.
3) Developing efficient compression, pruning and simplification methods to decrease the complexity and size of AI models while maintaining their performance.
4) Proposing a new approach for smart network management and anomaly detection.