A5046445187- Advanced Image and Video Retrieval Techniques
- Cellular Automata and Applications
- Ferroelectric and Negative Capacitance Devices
- Advanced Memory and Neural Computing
- Image Retrieval and Classification Techniques
- IoT and Edge/Fog Computing
- Advanced Battery Technologies Research
- Mobile Crowdsensing and Crowdsourcing
- Spacecraft Design and Technology
- Advanced Neural Network Applications
- Water Systems and Optimization
- Context-Aware Activity Recognition Systems
- Network Time Synchronization Technologies
- Stochastic Gradient Optimization Techniques
- Indoor and Outdoor Localization Technologies
- Energy Harvesting in Wireless Networks
- Energy Efficient Wireless Sensor Networks
Aalborg University
2022-2024
The traditional role of the network layer is transfer packet replicas from source to destination through intermediate nodes. We present a generative that uses Generative AI (GenAI) at or edge nodes and analyze its impact on required data rates in network. conduct case study where GenAI-aided generate images prompts consist substantially compressed latent representations. results flow analyses under image quality constraints show can achieve an improvement more than 100% terms rate.
This paper introduces EcoPull, a sustainable Internet of Things (IoT) framework empowered by tiny machine learning (TinyML) models for fetching images from wireless visual sensor networks. Two types learnable TinyML are installed in the IoT devices: i) behavior model and ii) an image compressor model. The first filters out irrelevant current task, reducing unnecessary transmission resource competition among devices. second allows devices to communicate with receiver via latent...
Building Management Systems (BMSs) are transitioning from utilising wired installations to wireless Internet of Things (IoT) sensors and actuators. This shift introduces the requirement robust localisation methods which can link installed correct Control Units (CTUs) will facilitate continued communication. In order lessen installation burden on technicians, process should be made more complicated by method. We propose an automated version fingerprinting-based method estimates location with...
This letter advocates the use of a Tiny Machine Learning (TinyML) model for energy-efficient semantic data retrieval from Internet Things (IoT) devices. In our framework, edge server (ES) transmits task-related TinyML before starting collection so that IoT devices can send only semantically relevant data. However, receiving ML and its processing at consumes additional energy. We consider specific instance image investigate gain brought by proposed scheme in terms energy efficiency,...
The traditional role of the network layer is transfer packet replicas from source to destination through intermediate nodes. We present a generative that uses Generative AI (GenAI) at or edge nodes and analyze its impact on required data rates in network. conduct case study where GenAI-aided generate images prompts consist substantially compressed latent representations. results flow analyses under image quality constraints show can achieve an improvement more than 100% terms rate.
In many distributed sensing use-cases, sensor nodes need to allocate a time-stamp sensed data. When combining data from different sensors, it is important have synchronized timestamps. using wireless networks (WSNs) without dedicated hardware such as GPS clock, keeping the sensors synchronised via synchronization protocols will increase communication overhead and resulting clock accuracy depend on latency jitter of end-to-end paths. order become independent transmission medium, this paper...