Social-aware Federated Learning: Challenges and Opportunities in Collaborative Data Training

Federated learning (FL) is a promising privacy-preserving solution to build powerful AI models. In many FL scenarios, such as healthcare or smart city monitoring, the user's devices may lack the required capabilities to collect suitable data, which limits their contributions to the global model. We contribute social-aware federated learning...

Cyclists in Motion: from data collection to behavioural models

As the title suggests, cyclists are the main topic of this dissertation<br/>and more specifically, their behaviour while they are ‘in motion’. The<br/>term ‘in motion’ is used in the title to represent microscopic operational<br/>cycling behaviour, which is the behaviour of cyclists, treated as individuals<br/>(microscopic level), while they are...

Taking Advantage of Data Generated by Products: Trends, Opportunities and Challenges

Now that all kinds of products are increasingly getting connected to the Internet, it is expected that it will become easier to collect data on how they are actually used during the middle-of-life stage of their product lifecycles. At the same time, a growing number of data analytics technologies offers opportunities to transform this data into...

Network Lifetime Analysis of Data Collection Protocols

When a sensor network is deployed, we fundamentally care about three main outcomes: to obtain as much data as possible (high delivery rate), to obtain data as fast as possible (low latency), and to obtain data for as long as possible (long lifetime). This last metric, called network lifetime, is of great importance and has been widely...

Do we really need A Priori Link Quality Estimation?

Traditionally, link quality estimation (LQE) has been viewed as an a priori step in sensor network routing protocols because it filters out unreliable links before data transmission. Recent results, however, show that protocols can perform well without a priori LQE. Because getting rid of LQE seems rather counter-intuitive, the aim of this work...