March 2025 - Nuno Costa is an Associate Professor at the Department of Computer Science Engineering, at the Superior School of Technology and Management of the Polytechnic Institute of Leiria in Portugal. In this interview, he describes his research and teaching experiences related to IoT systems. He explains how he came across oneM2M, and projects he is supporting to push the frontier of interoperable IoT.
Q: Would you begin with by introducing yourself and your activities in the IoT arena?
NC: My background is in computer science and systems, which provides a solid foundation for my professional activities and interest in IoT systems over time. I first held a software engineering position in a tech company called Critical Software before pursuing a career in academia. I completed an MSc in Informatics Engineering and then a PhD in Informatics. Now, I am a Professor at the Polytechnic Institute of Leiria where I teach undergraduate and graduate students, and work on research projects.
As part of the Computer Science and Communication Research Centre (CIIC), I belong to a Smart IoT Ecosystem Research Group. Within that, my research focuses on IoT and interoperability with an emphasis on smart objects and ubiquitous computing. Our research includes several funded projects with regional and national industries with the aim of enhancing competitiveness. We also have research links and PhD supervisions to organizations in Brazil, Ecuador, and Spain among others.
Q: What do you mean when you use the ‘smart objects’ term?
NC: IoT is a broad concept that is applicable to many sectors and use cases. My work focuses on a niche within the IoT landscape. With smart objects, the focus is on making everyday objects smart. These might be light bulbs, tables, doors, fridges, windows, etc. When these smart objects are well orchestrated the so-called ‘Smart Environments’ could be a reality.
Q: What aspects of smart objects do your projects focus on? Are they related to connectivity, to analytics that trigger actions, or to something else?
NC: My research focuses on emerging concepts, aiming to stay ahead of the commercial market. One way to frame this is through the four IoT phases proposed by Chin et al. The first phase involved devices and connectivity, now well-established with numerous market solutions. The second introduced remote-control applications, followed by the third, which focused on human interfaces. We are focused in the fourth phase, where seamless integration, AI, and reasoning become crucial as IoT systems grow in scale and complexity, opening an opportunity to innovate and provide better solutions for industry and society.
Consider a smart lightbulb: once plugged in, it should function with little to no human intervention. It must learn when to turn on and off by analysing user habits and collaborating with nearby smart objects. Additionally, it should offer an intuitive interface, allowing users to adjust its behaviour seamlessly—for instance, by clapping to turn it on or off, effectively "teaching" it over time.
Q: Is what you describe similar to what the Matter protocol and suppliers from the smart home ecosystem are aiming for?
NC: To some extent, yes. Matter is designed to enhance interoperability and compatibility between various manufacturers, aiming to make everyday devices work seamlessly together. At its core, Matter was created for cloudless environments; in my view, it fills a similar niche in the market that the AllJoyn software framework once addressed.
The focus of our group extends beyond the local domain to include infrastructure, such as cloud and edge computing, which is a key element of the oneM2M framework. However, we are also applying oneM2M to cloud-less and infrastructure-less networking to address final user spaces. We also want to go a step beyond by taking advantage of oneM2M’s semantic discovery and semantic interoperability capabilities.
Q: What is the connection between smart objects and your second research interest in ubiquitous computing?
NC: I draw inspiration from Mark Weiser, who introduced the concept of Ubiquitous Computing in 1988. Ubiquitous Computing can be present in any device, at any location, and in any format—not just limited to computers or smartphones. The key idea is that technology fades into the background of our lives, eliminating the need for explicit interaction between humans and the surrounding environment (smart objects).
Imagine a scenario where a person arrives home. Their front door and lock recognize them and grant access without the need for a physical key or passcode. Once inside, various appliances are activated automatically. In a family home, the activated devices could differ depending on whether a parent, mother, father, or child enters. The primary goal is to enhance safety, efficiency, and comfort in everyday life. Now, imagine the impact of this in scenarios involving the elderly, people with disabilities, or people requiring care.
Q: That sounds quite futuristic. How will researchers and industry get there?
NC: Fulfilling the vision of ubiquitous computing presents numerous challenges. It requires extensive distributed computing power, along with efficient reasoning and decision-making on behalf of users.
Additionally, factors such as mobility, trust, privacy, and safety must be carefully addressed. Limited resources, energy constraints, and the environment heterogeneity further complicate seamless interoperability and automation, demanding innovative solutions to ensure reliability and efficiency.
Apart from these challenges, potential solutions share common threads, such as the need for interoperability across heterogeneous devices and applications. Addressing issues like mobility, trust, privacy, safety, limited resources, and energy efficiency requires a step-by-step approach through intermediate research projects, gradually refining technologies to achieve seamless and reliable ubiquitous computing.
Q: So, what are some of the projects you are working on?
NC: There are two projects that I would like to mention because they involve real world problems comprising heterogeneous elements in an IoT system. These and other research projects are coordinated by my colleague Prof. António Pereira. The first is called Digital Twin Boids fire prevention System (DBoidS) and tackles three key forest fire stages: prediction/prevention, detection, and firefighting support. It utilizes cooperative UAVs, Digital Twins, AI, AR, and Systems Interoperability. Digital Twins digitally represent forests and UAVs, enabling real-time and predictive insights into fire hazards, human/animal presence, and risk conditions over time.
The other project is called Digital Twins Heterogeneous Unmanned Vehicles Ocean Preservation System (DUVOPS) and seeks ocean conservation through two approaches: Preventive Mode, predicting spill-prone areas using digital models to guide surveillance, and Reactive Mode, deploying resources efficiently based on reported incidents from European authorities or other sources.
Both projects use fleets of heterogeneous unmanned vehicles (UVs). However, DUVOPS integrates aquatic, aerial, and terrestrial UVs, adding complexity to interoperability. Our fleet includes home-built, open-source, proprietary UVs, and even virtual replicas, each differing in hardware, software architectures, control systems, and communication protocols. The glue to make all these heterogeneous parts exchange data and services, transparently and seamlessly, is oneM2M.
Q: What is the scale of these projects?
NC: There are a few ways to answer this. We designed our system to manage multiple dozens of UAVs. Since we are working with different agencies for these systems, we have also constructed a multi-tenant system that supports real and virtual UAVs, where virtual UAVs play an important role on simulations in the digital twins context.
For the DUVOPS project, from the information we are going to collect, we can alert our partners in the Portuguese Navy, for example, to monitor specific zones where a problem might arise. Everything operates in (soft) real time. So, on the system dashboard, users can click on an Unmanned vehicle to see its multi modal data as it is being acquired.
Q: You mentioned using oneM2M a couple of times. How did you learn about?
NC: It was a combination of academic needs and project requirements. In our Bachelor's in Computer Science Engineering, the Systems Integration and Interoperability course covers integration principles, focusing mainly on syntactic interoperability. In the Master’s in Computer Science Engineering – Mobile Computing, we extend this to semantic interoperability and we needed a practical framework for lab exercises. After extensive research, around 2019, we found oneM2M.
Seeing other research organizations using it reinforced our confidence. As highlighted in our projects, real-world scenarios often involve interoperability across heterogeneous environments—precisely where oneM2M excels.
Q: What has been your experience with the standard?
NC: Mastering oneM2M takes time, as the standard initially appears vast and complex. However, it is manageable. In our undergraduate program, ongoing projects focus on data gathering and remote control. Additionally, oneM2M’s openness enables more advanced projects, typically undertaken in the Final Project course, which runs for several months. Here, students interpret and implement the oneM2M standard, with a scoped focus on four key elements: the Common Services Entity (CSE), Application Entities (AEs), Containers, and Content Instances. Understanding these elements provides a solid foundation for building basic IoT systems while reinforcing object-oriented programming principles, as oneM2M resources map perfectly to Classes and Class extensions. Students work with C, C++, or Node.js, comparing performance metrics across implementation approaches.
In our MSc in Computer Science – Mobile Computing, oneM2M is also integrated into the Mobility in Computational Systems course. Beyond traditional data gathering and remote-control applications, students use oneM2M to develop smart objects over infrastructure-less networks, applying a P2P communication pattern. These projects emphasize integration, interoperability, and cooperation in spontaneous networking scenarios. They are typically tailored for smart home applications, including Virtual Butlers, smart objects as tangible interfaces (e.g., an augmented glass of water), indoor location tracking, user monitoring, etc. This is also a way to promote oneM2M within the industry as students enter the job market, as well as through consumers who become aware of products implementing the standard.
Q: We have covered a lot of topics, so what are your key messages and advice to any organizations tackling projects similar to yours?
NC: Standardization is crucial in IoT, but multiple standards create interoperability challenges. Since no single standard will dominate in the mid-term, it's essential to adopt solutions that can also bridge across different frameworks.
In our case, we focus on oneM2M because it is backed by leading global standards organizations, it is tailored for real-world use cases, it supports horizontal and semantic interoperability, it is commercially adopted, and it is prepared to bridge to other standards.
For future development, oneM2M offers great flexibility with its different device roles, infrastructure domain and field domain, the extensive set of well-defined resources, and protocol bindings. We leverage this to explore unconventional use cases using infrastructure-less and P2P communication patterns, as demonstrated in our paper, 'Smart Environments Based on Peer-to-Peer oneM2M IoT Standard: Preliminary Results,' published in DSAI 24.
