2015-current: Localization, mapping and path planning for an unmanned ground vehicle (UGV) with the aid of a group of unmanned aerial vehicles (UAVs) using active collaborative vision and multi-robot belief space planning.
Collaboration between unmanned ground and aerial vehicles (UGVs and UAVs) is of paramount importance in numerous applications, such as autonomous navigation, search and rescue, and surveillance scenarios. Aerial and ground robotic platforms can typically provide different sensory information that is often complimentary, and as such, can be used to enhance key capabilities required for reliable autonomous operation. The proposed research will investigate collaboration aspects between a heterogonous group comprising a UGV and small UAVs, focusing on operation in uncertain environments. The latter could be completely unknown, include dynamic scene or objects, or represented by an outdated and imprecise map. To facilitate a reliable autonomous operation, robots will collaboratively perceive the surrounding environment and infer their own states, and plan high-quality actions and paths while taking different sources of uncertainty into account. In this research we will investigate theoretically and experimentally such a multi-robot collaborative framework, with the aim to advance the state of the art in multi-robot autonomy under uncertainty.
This is a joint project of Intelligent Robotic Systems Laboratory (LIRS), Innopolis University, and Autonomous Navigation and Perception Laboratory (ANPL), Technion. LIRS will receive a grant in the amount of $110 thousand (₽7.2 million) from Russian Foundation for Basic Research (RFBR) and the Ministry of Science and Technology of Israel. More details are available here.
2014-current: Research and Development of Software Solutions for Static and Dynamic Based Control of Anthropomorphic Bipedal Robots Locomotion.
Anthropomorphic robots will become an important part of our daily life in the near future, and in order to assist a human such robots require reliable locomotion algorithms. Bipedal walking is one of the basic abilities for an anthropomorphic robot and without it the robot cannot fully operate in a natural for a human environment. Our research is targeted for developing efficient humanlike locomotion for a bipedal robot and a novel human-size Russian robot AR-601M is used for this research. This is a new ₽66 million research project which is supported by ”Scientific and Technological Research and Development Program of Russian Federation for 2014 - 2020 years” (research grant ID RFMEFI60914X0004) and Russian scientific company “Android Technics”.
2013-2014: Trustworthy Robotic Assistants (TRA).
Robotic assistants are being developed to assist with a range of tasks at work and home. The development of such assistants is being held back by the lack of a coherent and credible safety framework. The aim of our project is to bring together formal verification, simulation-based testing and formative user evaluation approaches in order to tackle the holistic analysis of safety in human-robot interactions. This is a new £1.2 million research project which is supported by Engineering and Physical Sciences Research Council. The details about the project could be found at the site of the TRA project.
2012-2013: Research on Adaptive and Modular Multi-terrain Mobility Planner.
The team of Neya Systems (Non Traditional Defense Contractor), Jet Propulsion Lab (FFRDC), and Carnegie Mellon University proposed to develop an adaptive, modular, multi-terrain mobility planner specifically designed to maximize mobility performance by adapting in real time to multiple mission objectives, terrain types and platform configurations as they change during a mission. The effort was focused on extending existing state-of-the-art mobility approaches by enabling autonomous platforms to be driven with the same adaptive techniques and highly reactive driving maneuvers leveraged by professional human operators. Our team developed a Homotopy based high level planner that selects pareto optimal trajectory classes based on user-tuned minimization criteria and implemented the planner in C++ with a friendly graphical user interface.
2007-2011: Research on 3D path planning for a crawler type robot.
“Optimal Path Selection for a Pilot System of Rescue Robot Navigation in
Random Step Environment”.
I proposed a new balance-based algorithm for crawler rescue robot navigation in random step environment. The algorithm was implemented in Matlab and confirmed through exhaustive simulations and experiments with "Kenaf" rescue robot. A complete framework of 3D path planning for a crawler type robot with regard to static balance property was developed and it could be generically used as a research scheme for other types of crawler robots.
This research was a part of NEDO Project for Strategic Development of Advanced Robotics Elemental Technologies, High-Speed Search Robot System in Confined Space sponsored by the Japanese government.
Teleoperation with the proposed Assistant Pilot System framework for Kenaf rescue robot
2011: Research assistant in Yamazumi Project.
My task was to identify and model in Simulink (Matlab) an approximation of a crawler vehicle steering system from a raw experimental data which would enable further design of a steering control system.
Yamazumi-4 Fully Autonomous Wheel Loader
2006-2007: Independent researcher.
Urban Search and Rescue (USAR) and Teleoperation.
I studied modern trends in USAR robotics and learned teleoperation process experimenting with "Acros" rescue robot in random step environment as a part of a rescue team.
"Acros" rescue robot
2002-2006: Research on 2D path planning.
My Master Thesis “Autonomous Robot Navigation” consisted of two independent parts.
First part dealt with mathematical development and implementation in C language of a new Bug-family algorithm for sensor-based navigation.
Second part proposed a spline-based iterative global navigation algorithm using potential field method, implemented in Matlab.
New CBug algorithm operating in office type environment - Simulation (Unix)
Smooth Path Search algorithm - Simulation (Matlab)
2003-2004: Sensor-based robot navigation in a maze project.
I implemented two Bug-family algorithms in C programming language (UNIX) and, after creating a set of mazes, conducted an exhaustive simulation to compare average and worst case behavior of the algorithms.
Experiments were conducted with "Yamabico" platform robot.
2002: Technical assistant in Image-based wafer navigation project.
My task was to support the research from mathematical point of view and to develop an effective data structure for map storage and fast navigation on a wafer.
2002-2006: Object recognition project.
Detection of intrinsic geometric properties of a surface from a polygonal mesh obtained from range data using different computational schemes for local estimation of intrinsic curvature geometric properties. The results on triangular meshes representing tessellations of synthetic geometric models were compared with the analytically computed values of the Gaussian and mean curvatures of the non uniform rational B-spline (NURBS) surfaces from which these meshes originated. Further this work was extended to range images of geometric objects obtained from a 3D laser scanner.
Triangular meshes of synthetic geometric model