A Neuromorphic Robot as a Digital Twin for Rodent Spatial Representation in Dynamic Environment Navigation

Document Type

Article

Publication Date

1-1-2026

Abstract

Spatial perception, localization, and adaptive navigation are core capabilities required for immersive Extended Reality (XR) systems. Achieving low-latency, energy-efficient, and adaptive spatial intelligence on resource-constrained XR platforms remains a major challenge. Inspired by spatial cognition in the nervous system, we propose a neuromorphic digital twin framework that serves as an XR cognitive backend for dynamic spatial environments and neural rehabilitation scenarios. Specifically, we present a neuromorphic robotic system that functions as an embodied digital twin platform for spatial cognition. The system implements multiple spatial cell models, including place cells, grid cells, head direction cells, and border cells, within a neuromorphic robot. Orientation-selective visual perception ensembles are integrated with these spatial representations. Through synaptic plasticity, visual inputs become linked to specific spatial locations, leading to the emergence of place-cell-like activity and adaptive spatial representations suitable for dynamic XR scenarios. To provide biologically grounded calibration, we use the rodent Dynamic Environment Navigation (DEN) task as a reference paradigm. Rodent trajectories and hippocampal electrophysiological recordings guide system calibration and enable direct comparison between neuromorphic spatial models and biological neural activity. By bridging experimental rodent neuroscience and embodied neuromorphic systems, the proposed framework establishes a biologically grounded digital twin cognitive backend for XR. This early-stage platform demonstrates how neuromorphic robotics and spatial-cell modeling can provide a foundation for future XR systems.

Publication Title

IEEE Journal on Emerging and Selected Topics in Circuits and Systems

Link to Full Text

Share

COinS