A Reinforcement Learning-Based Model for Mapping and Goal-Directed Navigation Using Multiscale Place Fields

Bekarys Dukenbaev; Andrew Gerstenslager; Alexander Johnson; Ali A. Minai

A Reinforcement Learning-Based Model for Mapping and Goal-Directed Navigation Using Multiscale Place Fields

Neural and Evolutionary Computing 2026-01-08 v1 Artificial Intelligence Robotics

Authors: Bekarys Dukenbaev , Andrew Gerstenslager , Alexander Johnson , Ali A. Minai

Abstract

Autonomous navigation in complex and partially observable environments remains a central challenge in robotics. Several bio-inspired models of mapping and navigation based on place cells in the mammalian hippocampus have been proposed. This paper introduces a new robust model that employs parallel layers of place fields at multiple spatial scales, a replay-based reward mechanism, and dynamic scale fusion. Simulations show that the model improves path efficiency and accelerates learning compared to single-scale baselines, highlighting the value of multiscale spatial representations for adaptive robot navigation.

Keywords

robot navigation reinforcement learning imitation learning

Cite

@article{arxiv.2601.03520,
  title  = {A Reinforcement Learning-Based Model for Mapping and Goal-Directed Navigation Using Multiscale Place Fields},
  author = {Bekarys Dukenbaev and Andrew Gerstenslager and Alexander Johnson and Ali A. Minai},
  journal= {arXiv preprint arXiv:2601.03520},
  year   = {2026}
}

Comments

11 pages, 8 figures. Submitted to IEEE Transactions on Cognitive and Developmental Systems

A Reinforcement Learning-Based Model for Mapping and Goal-Directed Navigation Using Multiscale Place Fields

Abstract

Keywords

Cite

Comments

Related papers