Foundational research in artificial intelligence, multiagent systems, game theory, automated planning, robotics and cybersecurity
applied research translated into building large scale systems and technology transfer by contracted research and building startups.
We conduct research on formal modelling and algorithmic foundations of next-generation trip plannings systems. Our algorithms are capable of proposing trip itineraries combining the full spectrum of mobility services available in modern cicties, combining individual and collective, fixed-schedule as well as on-demand means of transport while taking into account individual user preferences and availability of transport services and resources.
We explore how multi-agent systems and game theory can be used to improve maritime security, with particular focus on fighting maritime piracy. Our ultimate objective is to develop an integrated set of algorithmic techniques for maximizing transit security given the limited protection resources available. We achieve this by improving the coordination of the movement of merchant vessels and naval patrols, while taking into account the behavior of pirates.
AgentFly system is fast-time agent-based large-scale air traffic simulation including aircraft, air traffic management, and cognitive models of air traffic controllers. AgentFly is designed for what-if analysis and impacts of current air traffic, new tools, future concepts, and unmanned systems integration to traffic flow, controllers’ workload, and infrastructure. AgentFly is integrated with other air traffic simulators and tools used by Federal Aviation Administration and others.
Within this project we develop agent-based system for robust decentralized command and control of heterogeneous multi-asset team. Focus is put on real-time flight trajectory planning in dynamic 3D environment, coordination and cooperation algorithms and high-level task planning for resource allocation. Features of developed methods and algorithms are being verified by their deployment on various hardware assets, both VTOL and fixed-wing.
The goal of the Stratosphere Project is to use advanced machine learning algorithms to protect NGOs for free. We believe that NGOs are at risk and that we can provide a trusted, advanced and free solution to them. Stratosphere receives network flows and identifies known malicious behaviors based on detection algorithms. We provide free tools to download and a more complex free cloud service in the University where organizations can send their flows to receive alerts. Stratosphere is a complementary solution for your organization because it focuses on outgoing malicious behaviors to detect successfully attacked computers.
The project deals with novel approaches to improve mobility on demand services. Currently we focus on double auction mechanisms for passenger-taxi driver matchmaking. Our task is to develop plausible models of all participants. Large datasets available allow gradual transfer from expert-driven to more efficient fully data-driven models based on machine learning including deep neural networks, feature selection, feature extraction and related methods. The problems solved by this project serve as a testbed for more general algorithms of spatio-temporal data mining as well.
Models and algorithms for making teams of physical and/or digital autonomous agents coordinating their activity and collaborate on solving challenging problems, such as in multi-drone surveillance, connected vehicles and/or adaptive logistics.
Planning and decision making provides general algorithmic solutions to problems of single or repeated action selection in formally modelled environments of various domains. We develop algorithms which combine techniques of dynamic programming, heuristic search and provide efficient and practically usable planning approaches.
Intelligent Data Analysis applies modern AI methods to data mining in order to facilitate and improve knowledge acquisition. Based on semi and fully-automatic approaches, methods of machine learning such as pattern recognition or clustering are typically employed. The process often deals with high-dimensional or complex spatio-temporal data and their visualization.
We focus on online and offline computation of optimal strategies in dynamic domains that evolve in time with finite and infinite horizon. The main research topics include equilibrium computation algorithms, online sampling algorithms, learning in games, and computational complexity of related problems.
We aim to design new scalable algorithms and computational models enabling application and improving capabilities of collective robotic systems to operate in dynamic, unstructured environments with imperfect sensing and perception.
We have a deep knowledge and an extensive experience of algorithmic foundations of intelligent transport systems, ranging from next-generation trip planning, through agent-based transport modeling up to auction-based intelligent transport service marketplaces.
We explore how transport systems can be modelled as genuinely multi-agent systems, composed of autonomous agents with continuous, asynchronous control modules and the ability to interact freely with the environment and other agents. Such a fully agent-based approach – which is not supported by existing platforms – improves the modularity and extensibility of transport simulations and reduces constraints on the type of decision models that can be simulated (e.g. within-the-day replanning). We also explore how agent-based simulation models can be used as testbeds for analysing on-demand mobility and logistics services, including real-time ride sharing, transportation network services or next-generation taxi services.
The industrial cooperation is one of the most important application areas for us. The goal of this activity is to apply state of the art data engineering practices and algorithmization know-how to real world problems encountered within partner enterprises, including production planning, production scheduling and logistics domain assistive tools. This effort fits well into what is currently perceived as the most feasible way of integrating advanced (decentralized) AI algorithms to industry expert driven applications helping the industry to comply with the Industry 4.0 needs.
Our modern world is based on complex applications, networks, social relationships, mobile devices, permanent updates and synergic exponential organizations. The only drawback of our societies are our concerns for security, privacy and anonymity. We need to be secure in a world where every new connection leave us more vulnerable. We believe that science can help our community by leveraging the security risks with powerful and advanced protections.
Need to solve a challenging problem using AI? We can design a prototype solution for you.
Need advice on whether your problem could be solved using AI technology? Which technology and/or vendor to choose?
Something not clear? Do you want to improve your understanding of a particular AI subject?
Need incubate to an early tech idea in R&D environment. Do you want quick R&D mockup?
Want to collaborate in building new IP? Looking for partner for a 3rd party funded research?
Need to find out how to transfer R&D into business? How to build / get funded startup.
