Celestial navigation relies critically on extracting directional cues from skylight, solar position, and polarization patterns. In insects, the central complex integrates these signals to create a robust, time-compensated heading reference for orientation and homing. This work introduces a computational model and prototype sensor inspired by insect neuroanatomy, specifically the fan-like polarization filter arrays of the compound eye and the ring attractor network of the central complex. The sensor captures light intensity and polarisation information, while the model processes inputs into stable activity patterns representing heading directions. Temporal compensation is achieved via phase-synchronised activity oscillations, reflecting the endogenous clocks and light-sensitive protein cycles underlying circadian stability in natural systems. By combining these spatial and temporal features, the geocentric compass output mirrors insect performance in dynamic celestial conditions. This approach demonstrates biologically grounded principles for decentralised, low-energy navigation, providing insight for both biological understanding and engineering autonomous systems that exploit natural celestial signals for robust orientation.
Significance: Celestial navigation research benefits from linking biological computation and sensing to practical design. This project bridges insect neuroethology with sensor engineering, illustrating how natural time-compensation and polarisation processing can be mimicked in artificial devices. The results advance understanding of animal navigation under real-world conditions and open up new possibilities for satellite-free, energy-efficient systems for robotics, environmental monitoring, and autonomous vehicles.