A Machine Learning-based Intelligent Middleware Platform for a Private Edge Cloud

This project focused on designing an intelligent middleware architecture for managing heterogeneous compute and network resources in a private edge cloud composed of computers, mobile devices, sensors, and robots. The project developed and evaluated several components, including machine-learning-based resource allocation, multi-network management, link-quality and link-lifetime prediction, and computation offloading. The work aimed to improve performance, reduce latency and energy consumption, and support IoT and beyond-5G applications.

Private Mobile Edge Cloud for 5G Network Applications

This research explored localized edge cloud environments created from mobile and stationary devices already available within homes, offices, campuses, and other physical environments. The objective was to reduce dependence on remote cloud data centers while using nearby computing and networking resources to support latency-sensitive and data-intensive applications. The work investigated private edge-cloud architecture, computation offloading, task placement, network-aware resource management, energy consumption, and communication latency.

Smart Ageing Service based on Web of Objects

This interdisciplinary project investigated connected technologies designed to support independent living and improve quality of life for older adults. Research topics included representation, monitoring, discovery, and coordination of connected objects within intelligent environments.

EmoSpaces

EmoSpaces investigated how emotion and sentiment could be incorporated as contextual information within intelligent IoT services and adaptive environments.

Korean Satellite Based Augmentation System

Korean Satellite Based Augmentation System was a large-scale research project funded by the Ministry of Land, Transport, and Maritime Affairs Korea. It aimed to augment the Global Navigation Satellite System by broadcasting additional signals from geostationary satellites and providing differential correction messages and integrity data for the GNSS satellites. Satellite-based augmentation system uses a network of wide-area reference stations, wide-area master station, ground earth station, and geostationary satellites. Wide area reference stations are widely dispersed GNSS data collection sites that monitor and process satellite data to determine satellite orbit and clock drift plus delays caused by the atmosphere and ionosphere. This information is then transmitted to a wide-area master station through a terrestrial communication network. A wide area master station creates correction messages and broadcasts them through geostationary satellites. The user segment receives the correction messages and applies them to improve position accuracy. The project was carried out by a consortium of seven universities and research institutes. The team at the Electronics and Telecommunications Research Institute was involved in the design and development of protocols, data processing software, and network monitoring tools for wide-area reference stations and user segments.

---

Mobile Ad hoc Computational Cloud

This research investigated a mobile distributed computing infrastructure that enables mobile nodes to share computing resources either with existing cloud-based systems or directly with nearby devices through short-range wireless technologies.

The project examined the following system design and resource management challenges.

Task Migration

Task migration can improve resource utilization, balance workloads, and prevent application failure when devices move, disconnect, or become overloaded. However, migration decisions are difficult in mobile environments because communication conditions and migration costs can change rapidly, particularly for data-intensive tasks. This research examined migration strategies that account for task-completion time, communication overhead, data transfer requirements, and the availability of candidate devices.

Power Management

Mobile ad hoc clouds rely on battery-powered devices, making energy management essential for extending both node and application lifetimes. Energy is consumed by computation, memory use, and wireless communication, with data transfer becoming particularly costly when interdependent tasks exchange large volumes of data. Building on our earlier mobile-grid work, we developed an energy-efficient resource-allocation approach that considered both communication cost and transmission power. Interdependent, data-intensive tasks were assigned to nodes able to communicate at lower transmission power, reducing the energy consumed during data exchange.

Mobility Management

Node mobility can cause task failure, increase communication costs, and make previously collected resource information inaccurate. Movement outside the available network area may disconnect a node entirely, while movement within the network can change route quality and increase data-transfer time. The research examined task migration and reallocation as mechanisms for responding to mobility. It also considered the trade-off between continuous resource monitoring, which introduces communication overhead, and on-demand resource queries, which can delay scheduling and migration decisions.

Mobile Cloud Architecture

Centralized architectures can make informed resource-allocation decisions using a global view of the system, but they may suffer from limited scalability and a single point of failure. Fully decentralized architectures improve resilience and scalability, but their decisions may be less efficient because individual nodes have only partial system information. The research examined hybrid architectural approaches that balance resource-allocation quality, scalability, communication overhead, and resilience across different mobile ad hoc cloud environments.

Transport and Routing Protocols

Mobile ad hoc computational clouds require communication protocols that account for application deadlines, energy consumption, changing network topology, and unstable wireless links. The research investigated adaptive transport and routing approaches that used mechanisms such as multi-hop communication, transmission-power control, alternate paths, and coordination with resource- and mobility-management components. The objective was to improve communication reliability and energy efficiency without violating application requirements.

Wireless Communication Technologies

Mobile devices may support multiple communication technologies, including Wi-Fi Direct, Wi-Fi, Bluetooth, and cellular interfaces. These technologies differ in bandwidth, range, energy consumption, connection setup, and reliability. The research examined how communication interfaces could be selected, switched, or combined to support efficient and reliable data transfer in mobile ad hoc computational clouds.

---

Mobile Ad hoc Computational Grid

This project investigated how mobile devices can share computing resources in an ad hoc environment to support resource-intensive applications such as intelligent mobile video surveillance. Unlike conventional clusters and cloud platforms, mobile ad hoc computational grids operate over shared and unreliable wireless links and must account for node mobility, limited battery capacity, and the absence of fixed infrastructure.

We developed resource-management methods tailored to these conditions. The research addressed energy consumption, mobility, task dependencies, communication costs, node reliability, and the effects of transport, routing, and medium-access-control protocols on distributed-application performance. Many of these challenges are also relevant to mobile peer-to-peer systems, vehicular networks, and contemporary edge-computing environments.

Energy Efficient Resource Allocation

Energy efficient resource allocation is a critical design criterion due to the limited battery life of nodes. The primary sources of energy consumption include CPU processing, memory usage, and data transmission within the network. Key factors contributing to transmission energy consumption are the transmission power required to send data and communication costs induced by data transfers between tasks. In this research, we developed several resource allocation strategies aimed at reducing energy consumption caused by data transmission between data-intensive dependent tasks. The core idea was to allocate data-intensive dependent tasks to nodes accessible at minimum transmission power. Additionally, we proposed a hybrid architecture that enables effective allocation while reducing the processing burden on individual nodes and communication costs associated with the exchange of control information.

Effective and Robust Resource Allocation

The performance of parallel applications depends heavily on selecting reliable nodes, which is particularly challenging in mobile ad hoc computational grids due to node mobility. To address this challenge, we developed a Markov-based mobility model that used historical mobility patterns. Furthermore, we identified open issues and challenges in developing a robust and effective resource allocation scheme and proposed a set of novel resource allocation policies that consider both the characteristics of applications and the dynamic, distributed nature of mobile ad hoc computational grids. We also analyzed the performance of the transport, routing, and medium access control layers and identified key issues that degrade the performance of parallel applications and ad hoc computational grids.

Energy Efficient and Robust Resource Allocation

This research addressed the challenges of power consumption and node mobility in mobile ad hoc computational grids, aiming to reduce data transfer times between dependent tasks and the energy consumed during data transmission. It also investigated the impact of lower-layer protocols on data transfer times and energy consumption. Extensive simulation results demonstrated that the proposed policies significantly reduced energy consumption, although their performance in terms of application completion time depends on node deployment and application configuration.

Wi-Fi Direct-based Mobile Ad hoc Network

Existing mobile ad hoc network technologies offer limited bandwidth and are therefore unsuitable for data-intensive applications such as automated video surveillance. Wi-Fi Direct allows mobile devices to communicate directly with each other without relying on a wireless access point. Compared to existing technologies, it provides higher data rates, which are sufficient for many mobile data-intensive applications. However, Wi-Fi Direct has two main limitations: it does not support communication between two client devices within a group, and it lacks support for multi-hop communication—a critical requirement for several IoT and mobile cyber-physical system applications. This research developed an energy-efficient and robust Wi-Fi Direct-based multi-hop mobile ad hoc network.

---

RFID and WSN Integration

This research investigated architectures combining radio-frequency identification and wireless sensor networks to provide both object identification and contextual sensing. The work addressed redundant RFID data, communication overhead, in-network filtering, and energy-efficient data processing.

Continuous Object Detection and Tracking

This project investigated energy-efficient methods for detecting and tracking continuously distributed phenomena, such as fires, hazardous materials, and environmental events, using wireless sensor networks. The work addressed changes in object size, shape, location, diffusion, and fragmentation over time.

Virtual Gateway

This research investigated fault-tolerant gateway mechanisms for maintaining connectivity between mobile ad hoc networks and fixed communication infrastructure. The virtual-gateway approach was designed to improve continuity of service when gateway devices moved, disconnected, or failed.