SCAMO: Scalable task and motion planning in multi-robot systems using cloud services (Plánování v multi­robotických systémech s využitím cloudových technologií)

I this project we investigate methods for fast task and motion planning using cloud-based services. The project is motivated by the industrial bin-picking scenario. In this scenario, the task of the robots is to unload objects from container. We consider the randomized version of the problem, where positions/orientations of the objects are not known in advance and therefore trajectories for the robots cannot be computed in advance. Besides development of new task and motion planning methods, we employ cloud-based computing.
 
The project is realized with Prof. Dr.-Ing. Jörg Krüger, and Dipl.-Ing. Axel Vick from Faculty V of Mechanical Engineering and Transport Systems, Department of Machine Tools and Factory Management,Industrial Automation Technology Group, Berlin, Germany.
Project supported by MSMT-Mobiltiy and DAAD.
 
Publications:
  1. V Vonasek, A Vick and M Saska. Motion Planning with Motion Primitives for Industrial Bin Picking. In ETFA. 2017. BibTeX

    @inproceedings{vonasek2017mp,
    	title = "Motion Planning with Motion Primitives for Industrial Bin Picking",
    	author = "V. Vonasek and A. Vick and M. Saska",
    	booktitle = "ETFA",
    	year = 2017
    }
    
 
MUSE: Motion planning for Multi UAV Surveillance (Metody plánování pohybu autonomních helikoptér v úlohách dohledu)
 

In the task of cooperative surveillance using Micro Aerial Vehicles (MAVs), MAVs cooperatively observe a given set of Areas of Interest by hovering in a suitable sensing location. In this project, we investigate methods for planning of surveillance missions in GPS denied environments. The MAVs are localized using a relative on-board visual-based locations. The the motion of MAVs is constrained due to requirements of the on-board localization system on mutual visibility of MAVs and limits of the relative distance. Positions of the team members have to be maintained in a given range in order to ensure proper function of the localization system and a safe flight. This requires to find an appropriate distribution of MAVs above the areas of interest and also feasible trajectories from a distant depot to reach these locations while keeping the onboard localization and MAV motion constraints. 

   
Cooperation with Prof. Dr.-Ing. Dr. h.c. Heinz Wörn, Dipl.-Ing., Dipl.-Math., David Oertel and Dipl.-Ing. Sergej Neumann from Institut für Anthropomatik und Robotik (IAR) - Intelligente Prozessautomation und Robotik (IPR), Karlsruhe Institute of Technology, Germany.
Project supported by MSMT-Mobiltiy and DAAD.
 
Examples of trajectories constructed for four MAVs that start from the depot (bottom left) and have to visit to Area of Interest (top right). During the motions, MAVs have to maintain visibility so the on-board relative localization system can work.
     

 

Cooperative surveillance and environment monitoring by heterogeneous teams of unmanned aerial vehicles (Kooperativní dohled a senzorické snímání heterogenními týmy bezpilotních vzdušných a pozemních prostředků)

The main objective of the proposed project is to design, implement and experimentally evaluate a complex heterogeneous system of fixed-wing unmanned aerial vehicles (UAVs), vertical take-off and landing micro aerial vehicles (MAVs) and unmanned ground vehicles (UGVs) designed for cooperative surveillance and environment monitoring. Taking advantage of expertise of Centre national de la recherche scientifique (CNRS) partner in deployment of UAVs in environment monitoring applications and expertise of Multi-robot Systems (MRS) group in design of techniques for MAVs flying in GPS-denied environment, we aim to combine abilities of UAVs to cover large areas under surveillance and MAVs to operate in GPS-denied areas. In addition, both partners have a long-term experience with integration of UGVs into teams of aerial vehicles, which also increases deployability of the system and its operation time. The aim of the project is to design methodologies that provide motion planning for UAV-MAV-UGV teams in a cooperative manner and that enable fusion of low-resolution data gathered during flights of fixed-wings UAVs in higher altitudes together with high-resolution information provided from MAVs and UGVs simultaneously observing areas in which the rough overview from UAVs is insufficient.
 

Cooperation with Simon Lacroix from Centre national de la recherche scientifique (CNRS), France: Simon Lacroix, Antonio Franchi, Juan Cortes 

Publications:
  1. R Pěnička, M Saska, C Reymann and S Lacroix. Reactive Dubins Traveling Salesman Problem for Replanning of Information Gathering by UAVs. In ECMR. 2017. URL BibTeX

    @inproceedings{penicka_ecmr_rdtsp,
    	author = "R. Pěnička and M. Saska and C. Reymann and S. Lacroix",
    	title = "Reactive Dubins Traveling Salesman Problem for Replanning of Information Gathering by UAVs",
    	booktitle = "ECMR",
    	year = 2017,
    	url = "http://mrs.felk.cvut.cz/ecmr17rdtsp"
    }