UP - logo
University of Primorska University Library (UPUK)
  • Qualitative control of cart and pole on uneven terrain [Elektronski vir]
    Šoberl, Domen ; Bratko, Ivan, računalništvo
    Qualitative modelling can be applied to the control of dynamic systems by following these steps: (1) learning a qualitative model of the controlled dynamic system from the system’s behaviors in time, ... (2) using the learned model for deriving a qualitative plan for the control task, and (3) executing the qualitative plan on the actual dynamic system. This approach has been demonstrated in the usual cart-pole control domain to be much more efficient in terms of learning sample size than usual variants of reinforcement learning. In this paper, we generalize the cart-pole problem to uneven terrain, such driving over a crater. We study whether the learned flat-surface qualitative controller can be advantageously transferred to the tasks of negotiating uneven terrain. Experiments show that the flat-surface qualitative controller is remarkably robust on the new, more difficult tasks.
    Source: QR 2022 [Elektronski vir] : proceedings of the 35th International Workshop on Qualitative Reasoning : co-located at International Conference on Artificial Intelligence (IJCAI’22), Vienna, July 23rd, 2022 (Str. 6-12)
    Type of material - conference contribution
    Publish date - 2022
    Language - english
    COBISS.SI-ID - 120894467

    Link(s):

    https://itemsweb.esade.edu/wi/research/JUICE/QR18-Proceedings.pdf

Shelf entry