Reusing risk-aware stochastic abstract policies in robotic navigation learning

Valdinei Freire Da Silva, Marcelo Li Koga, Fábio Gagliardi Cozman, Anna Helena Reali Costa

Resultado de la investigación: Capítulo del libro/informe/acta de congresoContribución a la conferencia

1 Cita (Scopus)

Resumen

In this paper we improve learning performance of a risk-aware robot facing navigation tasks by employing transfer learning; that is, we use information from a previously solved task to accelerate learning in a new task. To do so, we transfer risk-aware memoryless stochastic abstract policies into a new task. We show how to incorporate risk-awareness into robotic navigation tasks, in particular when tasks are modeled as stochastic shortest path problems. We then show how to use a modified policy iteration algorithm, called AbsProb-PI, to obtain risk-neutral and risk-prone memoryless stochastic abstract policies. Finally, we propose a method that combines abstract policies, and show how to use the combined policy in a new navigation task. Experiments validate our proposals and show that one can find effective abstract policies that can improve robot behavior in navigation problems. © 2014 Springer-Verlag Berlin Heidelberg.
Idioma originalInglés estadounidense
Título de la publicación alojadaReusing risk-aware stochastic abstract policies in robotic navigation learning
Páginas256-267
Número de páginas12
ISBN (versión digital)9783662444672
DOI
EstadoPublicada - 1 ene 2014
Publicado de forma externa
EventoLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) -
Duración: 1 ene 2018 → …

Serie de la publicación

NombreLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volumen8371 LNAI
ISSN (versión impresa)0302-9743

Conferencia

ConferenciaLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Período1/01/18 → …

Huella Profundice en los temas de investigación de 'Reusing risk-aware stochastic abstract policies in robotic navigation learning'. En conjunto forman una huella única.

  • Citar esto

    Da Silva, V. F., Koga, M. L., Cozman, F. G., & Costa, A. H. R. (2014). Reusing risk-aware stochastic abstract policies in robotic navigation learning. En Reusing risk-aware stochastic abstract policies in robotic navigation learning (pp. 256-267). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8371 LNAI). https://doi.org/10.1007/978-3-662-44468-9_23