- Description of the Research Group +
The key objectives of the R&D work carried out at the Dynamical Systems and Ocean Robotics Group (DSORG) are twofold: i) to study a number of challenging theoretical problems in the areas of advanced robotic vehicle systems design, navigation, and control and ii) to exploit the theoretical methodologies developed to yield faster, cheaper, and far more efficient systems and tools for ocean exploration and exploitation and critical infrastructure inspection, than those available today. The tools include surface and underwater robots, as well as aerial vehicles working as communication relays or re-directing the operations of marine vehicles upon detection of relevant episodic events. These goals have motivated the definition of a research and development program addressing theoretical and practical engineering topics, as well as issues that are at the crossroads of marine science and technology and are the main focus of some of the cooperative research and development work set forth under Research Lines D and E of the LARSyS. The following lines of action underpin the work carried out at the DSORG:
T1). Linear and nonlinear systems theory: study and development of theoretical tools for the analysis and design of linear and nonlinear control / filtering systems.
T2) Multiple Model Adpative Estimation (MMAE) and Robust Multiple Model Adaptive Control (RMMAC): Development of new methodologies for the design of estimators and robust adaptive controllers for plants with structured and unstructured uncertainty.
T3) Design of Navigation and Positioning Systems for autonomous underwater platforms and groups of marine vehicles interacting with human divers. Study of advanced solutions focusing on the: i) study and practical evaluation of optimal sensor placement algorithms and acoustics-based systems for simultaneous underwater vehicle tracking; iii) development of geophysical-based navigation algorithms for single and cooperative multiple vehicles using terrain and geomagnetic data.
T4) Motion Control of single and multiple vehicles under stringent communication constraints, including those imposed by the harsh conditions observed in the ocean: temporary communication losses, acoustic multiple path effects, "ray bending", and fading. Problems addressed: i) Cooperative control and navigation of groups of autonomous vehicles; ii) Close-range multiple vehicle formation control using acoustic and optical-based range/bearing measurements, iii) Networked control over faulty communication links, iv) Terrain Contour Tracking, v) Visual servoing control.
T5) Development of advanced methods for Cooperative Multiple Vehicle Motion Planning in the presence of stationary and moving obstacles, subjected to energy and temporal constraints, as well as navigation-related constraints.
T6) Development of networked control systems by resorting to the framework of hybrid systems to model several networked control systems that are periodic or event based and using the machinery of Volterra equations, piecewise deterministic processes, dynamic programming, and Lyapunov based tools to address and solve the resulting problems.
P1) Design and development of Autonomous Underwater Vehicles (AUVs), Autonomous Surface Craft (ASCs), and Unmanned Air Vehicles (UAVs) with on-board integration of scientific sensor suites and data acquisition / logging systems.
P2) Development of distributed hardware and software architectures for cooperative navigation and motion control of multiple vehicles, and mission control of heterogeneous platforms
- Main achievements +
- Structure of the Research Group +
The Dynamic Systems and Oceans Robotics Group (DSORg) of ISR is composed of four permanent IST staff members, two post-doctoral fellows who are also integrated members of the research unit, three research engineers, two technicians, and one administrative officer. Currently, the group hosts and supports the research activity of 7 MSc students and 16 PhD students. The group has a web page listing its members, projects, publications, and research activities in general.
The management structure of DSORg is kept light and effective: it consists of a Group Coordinator and a Scientific Board composed of all PhD holders. Periodically, or whenever deemed necessary, the board members are requested to provide their perspectives on relevant strategic decisions regarding scientific avenues to explore, projects to propose, and topics for PhD. theses. The board is also in charge of deciding on the investments to be made in the infrastructures needed to carry out practical experiments in the areas of air and marine robotic systems. Currently, four major supporting infrastructures are in place: i) an electronics/mechanical laboratory at the campus of IST in Lisbon, Alameda, ii) a special laboratory for assembling robotic vehicles, at the installation of IST near Parque das Nações. Lisbon, with close access to the Tagus river, iii) a water tank and associated installations to test robotic vehicles and acoustic equipment at the campus of IST in Oeiras, Tagus Park, and iv) a research vessel (ISeaTec) to carry out experimental work with marine robots and supporting systems in the Tagus river and at sea. Each of the infrastructures is managed by a senior member of the DSORg and a research engineer or technician.
To coordinate the R&D activities of the DSORg, a simple hierarchical organization is adopted: i) PhD holders supervise PhD students and oversee the participation of members of the group in R&D projects as well as dissemination and outreach activities that target the scientific community, the student population, and the public at large, ii) doctoral students carry out research on selected topics of interest to the group and help in the supervision of MSc students, iii) MSc students focus their activity on self-contained theoretical and practical issues that arise in the scope of doctoral theses and research projects undertaken in the group.
The research engineers and the technicians in the group are key for a seamless transition from the laboratory to the real world, to assess the efficacy of algorithms for single and multiple robotic vehicle motion planning, navigation, and control. They are for the most part involved in practical work pertaining to tasks carried out in the scope of national and international projects. Namely, in the design, development, and testing of the electrical, electronic and mechanical systems that go into the making of aerial and marine robots, as well as in the writing of software that is run in the vehicle-resident target computers. They are also responsible for maintaining the vehicles and associated systems ready to be used in real world applications. To a great extent, the software and hardware solutions adopted are the result of studies carried out within the group towards the design of mechanical, electrical, and electronic systems as well as the development of advanced algorithms for single and cooperative (multiple vehicle) navigation, guidance, and control, as well as mission planning and mission control. These activities are in line with the main objectives of the group.
All members of the DSORg are strongly encouraged to participate actively in one or more activities that include student supervision, internal project coordination, consultancy with the industry, and execution of R&D contracts with national and international funding agencies and companies. The latter two activities afford the group the financial leverage needed to maintain and upgrade the infrastructures and equipment.
- Objectives of the Research Group +
We set forth the goal of bringing together dynamical systems theory and systems engineering practice and applying the resulting methods and tools to the design, development, and operation of innovative ocean robotic platforms and associated enabling systems, including those needed to observe the ocean from the air. Strengthening the existing cooperative links with research groups worldwide and participating in national and European projects and Training Networks in areas that require an adequate balance between challenging theoretical and practical issues are also a priority. In this context, the proposed research and development effort will pave the way for an in depth understanding of core theoretical issues that are at the root of advanced marine/air robotic systems design; at the same time, it will afford the researchers involved in Thematic Area of "Ocean Exploration and Exploitation" robotic platforms and supporting systems for the seamless integration of ocean sensor suites, acoustic devices, and signal processing software for the execution of missions with scientific and commercial impact, at sea. Through its aerial segment, the work will also impact on Thematic Area of "Aeronautic and Space Systems". From a theoretical standpoint, the following topics will deserve special consideration:
Cooperative motion planning - study of motion planning systems that take explicitly into account the vehicle dynamics and address inter-vehicle collision avoidance, while meeting a number of criteria that may include simultaneous times of arrival at assigned target points, energy minimization, acoustic communication constraints, and the "maximization" of terrain information along the vehicle paths for terrain-based navigation purposes.
Cooperative Motion Control - design of distributed control laws for cooperative motion control in the presence of limited communications, time-delays, model uncertainty, external disturbances (currents, waves, wind), and noisy measurements. This is particularly significant in the case of underwater vehicles that rely on acoustic communications. The core activity will focus on the development of algorithms for compliant formation control using range and/or bearing angle measurements, thus dispensing with the need to equip all vehicles in a group with overly expensive inertial units. We shall contribute to and exploit the mathematical formalism that lies at the intersection of graph theory (to capture the time-varying characteristics of the underlying communication channels), logic based communications (to decide what information to transmit and when to transmit), and moving horizon estimation (to deal with information losses across the communication channels). In the case of air vehicles, the research will also address the challenging problem of visual servoing control.
Single and Multiple Vehicle Cooperative Navigation - study of algorithms for underwater navigation and target positioning by exploiting new concepts that are currently at the forefront of research. Namely, single and multiple beacon navigation, simultaneous beacon and vehicle localization using SLAM-like techniques, and geophysical based navigation (including terrain-based and geomagnetic). Interesting theoretical issues are the characterization of the observability properties of the navigation and positioning systems developed, assessing the type of performance that can be obtained with these systems, and determining the cardinality of the sets of indistinguishable states to guide the design of multiple model adaptive estimation (MMAE) filters. The research work will also address cooperative navigation, whereby multiple vehicles exchange inertial, terrain-based, and geomagnetic information over an acoustic communication network. The DSORg is extremely well positioned to undertake research in this area, given its access to a fleet of 6 autonomous marine vehicles, a network of 6 acoustic modems/ranging devices, and a magnetic gradiometer.