@InProceedings{Pecka-CVWW2015,
  IS = { zkontrolovano 26 Jun 2015 },
  UPDATE  = { 2015-03-10 },
  author =      {Pecka, Martin and Zimmermann, Karel and Svoboda, Tom{\'a}{\v s} },
  affiliation = {13133-13133-37240},
  title =       {Safe Exploration for Reinforcement Learning in Real Unstructured Environments},
  year =        {2015},
  pages =       {85-93},
  booktitle =   {CVWW 2015: Proceedings of the 20th Computer Vision Winter Workshop},
  editor =      {Paul Wohlhart, Vincent Lepetit},
  publisher =   {Graz University of Technology},
  address =     {Graz, Austria},
  isbn =        {978-3-85125-388-7},
  book_pages =  {137},
  month =       {2},
  day =         {9-11},
  venue =       {Schloss Seggau, Seggau, Austria},
  annote =      {In USAR (Urban Search and Rescue) missions, robots
                  are often required to operate in an unknown
                  environment and with imprecise data coming from
                  their sensors. However, it is highly desired that
                  the robots only act in a safe manner and do not
                  perform actions that could probably make damage to
                  them. To train some tasks with the robot, we utilize
                  reinforcement learning (RL). This machine learning
                  method however requires the robot to perform actions
                  leading to unknown states, which may be
                  dangerous. We develop a framework for training a
                  safety function which constrains possible actions to
                  a subset of really safe actions. Our approach
                  utilizes two basic concepts. First, a "core" of the
                  safety function is given by a cautious simulator and
                  possibly also by manually given examples. Second, a
                  classifier training phase is performed (using
                  Neyman-Pearson SVMs), which extends the safety
                  function to the states where the simulator fails to
                  recognize safe states.},
  keywords =    {safe exploration, reinforcement learning,
                  Neyman-Pearson, SVM, machine learning, robotics,
                  rescue robotics},
  project =     {GACR 14-13876S, SGS13/142/OHK3/2T/13, FP7-ICT-609763 TRADR},
  doi =         {10.3217/978-3-85125-388-7},
  psurl       = {[PDF, 810 KB]},
  www         = {http://cvww2015.icg.tugraz.at},
  acceptance_ratio = { 19/32 },
}