Datamite and PATH2ZERO projects

Grigorios currently also participates in the Datamite project. DATAMITE develops a simple but impactful technical framework that enables European enterprises and public administrations to overcome existing challenges and facilitate the monetisation of their data. The core target consists of helping users to better monetize, govern and enhance the trust of their data by developing a set of key modules: Data Governance, Quality, Security, Sharing & Supporting Tools. Grigorios is supporting activities for the integration of software tools that were developed as part of Datamite, during the final months of the project work.

PATH2ZERO: minor typical interactions with the PATH2ZERO project during its initial stages.

Data hub for the creation of Energy Communities at Local Level and to Advance Research on them

Grigorios is also currently involved in the Data Cellar project, which aims to create a federated energy dataspace that will support the creation, development and management of Local Energy Communities in the EU. Grigorios is involved in various activities of the project, including the specification and requirements of the dataspace design, algorithms for data anonymization and approaches for data aggregation, energy data interoperability issues and others. Regarding data anonymization techniques, methods for k-anonymization were reviewed and developed, e.g. Mondrian multi-dimensional partitioning, and are proposed to be considered for integration in the data space for the case studies. For data aggregation methods, temporal and spatial aggregation has been proposed, combined with training of self-organizing maps to extract useful information from large datasets. Integration with validation cases will also highlight the approaches.

A POsitive Energy CITY Transformation Framework

Grigorios is currently working on POCITYF project, on issues related to participatory systems and smart cities, for historical cities to become greener, smarter and more liveable while respecting their cultural heritage. By implementing and testing Positive Energy District in its cities, POCITYF will support Europe in the race to become the first Carbon Neutral Continent by 2050. Grigorios's current main research activity is about citizen engagement software platforms and citizens' behavioural change towards positive energy districts, in parallel with customization of software platforms for smart city indicators monitoring (KPIs). Further, Grigorios contributes on the smart cities and communities joint vision for the future activities.

Leveraging universal social networking and the IoT for urban-scale participatory systems.

Grigorios experimented on urban-scale participatory systems at INRIA Paris. The aim of the work was to examine whether, and how, the IoT and universal social networking can improve engagement and participation in communities, through digital participation. Relevant technologies were studied, regarding cyber-physical systems that blend the cyber world with physical entities in urban environments, concepts such as the Social IoT for connectivity and interaction of personal devices and users, and technologies such as the Named Data Networking, which can be combined with Universal Social Networking as alternative networking solutions. Further, the study of relevant digital assistants in supporting middleware was considered as a way to boost participation and engagement in collective activities organized by communities, institutions or governmental organizations. Towards the realization of participatory practices, the Social Participation Network was proposed as an initial design of the participating entities in the cyber-physical system, by linking people with things and entities that can be modelled and implemented technically in a supporting social middleware,i.e. where people can choose the communication platform of their preference. Furthermore, the initial design of SPNs paved the way for modelling the relations among the participating entities towards a fully-defined and functional Social Participation Network.

Smart assembly robots with advanced functionalities.

The aim of the SARAFun project was to enable non-expert users to teach an assembly robot how to perform an assembly task in less than a day. It is commonly known that robots until then needed an expert engineer to re-design the actions of a robot when a new task was requested. By using real time data from RGBD cameras and tracking of the teacher's moves in front of a learning robot, the robot could learn how to implement the actions. Grigorios' main work was the key-frame extraction module, i.e. the extraction of a set of key-frames from the labeled real time robot learning data, so that the action can be effectively re-produced. The underlying objects were modelled as 3D ellipsoids through a convex optimization program, and the object interactions in the learning scene as semantic graphs, transforming the learning task into a fully-defined tracking problem. Further, the non-expert user could finally interact with the robot via a common tablet pc platform.

Technologies that were studied for the project include computer vision and robotics. RGBD data were handled as 3D Point clouds and implementation was based on C++ programming language, while the Robot Operating System was considered as the operational middleware. The employed robot for the experimentation was the ABB YuMi robot.

Simulation-based optimization for intermodal transportation networks.

Part of Grigorios' work during the period of employment at the University of Ioannina was the optimization of intermodal transportation networks for Adriatic ports' decision support system. The main research question was how to effectively manage the vehicles and containers that reached the ports of Adriatic sea in order to alleviate from congestion problems and port dysfunction. Based on real data that were modelled in stochastic simulators, methods for simulation-optimization were developed that were employed to solve the research problems. The methods were based on nature-inspired heuristics, to name the most common, the Particle Swarm Optimization method. Modelling techniques were also considered to solve the problem, although the complexity and organization of data required the use of simulation-based solutions.

The numerical optimization algorithms were developed in C++ and connected through SOAP web-services with the external simulators to transfer data and re-evaluate the problem.

Solving operations research problems using nature-inspired heuristics.

During his Master studies at the University of Ioannina, Grigorios participated in research activities at the domain of scientific computing, mainly working on numerical optimization. Multiple problems were solved ranging from theoretical benchmarking to operations research problems. Grigorios master thesis was based on experimenting with nature-inspired metaheuristic algorithms for inventory management and supply chain problems. Further, experimentation with global optimization algorithms, memetic algorithms, and local search methods on benchmarking problems, provided useful results and important research experience. Multiple methods were considered during this period in the field of metaheuristics and global optimization, gradient-based and gradient-free methods. Furthermore, some heuristics were also considered for solving non-continuous problems, e.g. integer or mixed integer problems, under proper algorithm configuration.

Development and experimentation was based on C++ programming language and analysis of results on matlab and python.