Fly Whiskers Meet LiDAR: Robust Navigation under Sensor Failure
Abstract
We extend FLYNN, a leaky integrator recurrent network wired on the Drosophila melanogaster connectome, with a 32-ray 2D LiDAR interface that requires no learned fusion layer beyond a scalar input gain. A 270 degree frontal scan at 8 m range is injected as inverted proximity into the 37 BM Vib head-bristle neurons. Across 24 paired comparisons in a four- state vision matrix with LiDAR on or off, FLYNN-LiDAR has 17 significant wins, 5 nonsignificant comparisons, and 2 significant losses against three LiDAR-capable baselines. Both losses are to MobileNet with the left eye off. With full vision, FLYNN- LiDAR reaches 96.6% success. Under total vision loss, the scan raises success from 45.2% to 64.6%; the baselines reach 44.6% to 56.6% with the scan on. Under out of distribution textures, it is the only LiDAR-capable model whose success improves with the scan (44.2% to 58.4%); both CNNs score 0%, and SmallWorld declines from 26.8% to 7.4%. Under periodic vision dropout, FLYNN-LiDAR achieves 93.6% and 89.6% success at 25% and 50% duty with fewer collisions. Profiling the evaluated forward paths gives 3.93M multiply accumulate operations per step; EfficientNet and MobileNet require 4.60x and 2.94x as many. Final- state analysis shows near-linear sensory superposition (cosine 0.9955).
Related
- Seeing Through Uncertainty: Free-Energy-Inspired Real-Time Adaptation for Robust Visual Navigation
- GEDI canopy cover assessment using airborne LiDAR across eco-climate regions of the United States
- Simplifying complexity: the reliability of lidar-derived forest structural complexity metrics for conservation management
- Monitoring Carbon Hotspots in Patchy Landscapes from Sparse Lidar and Dense Satellite Time Series
- Greener but not taller: LiDAR shows South Asia's forest greening is concentrated in structurally depleted forest
BibTeX
@misc{mahzarnia2026whiskers,
title = {{Fly Whiskers Meet LiDAR: Robust Navigation under Sensor Failure}},
author = {Ali Mahzarnia and Homayoun Jamshidi},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.03897},
eprint = {2611.03897},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.03897},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Fly Whiskers Meet LiDAR: Robust Navigation under Sensor Failure
AU - Ali Mahzarnia
AU - Homayoun Jamshidi
PY - 2026
DA - 2026/11/02
PB - arXiv
DO - 10.48550/arxiv.2611.03897
UR - https://arxiv.org/abs/2611.03897
AB - We extend FLYNN, a leaky integrator recurrent network wired on the Drosophila melanogaster connectome, with a 32-ray 2D LiDAR interface that requires no learned fusion layer beyond a scalar input gain. A 270 degree frontal scan at 8 m range is injected as inverted proximity into the 37 BM Vib head-bristle neurons. Across 24 paired comparisons in a four- state vision matrix with LiDAR on or off, FLYNN-LiDAR has 17 significant wins, 5 nonsignificant comparisons, and 2 significant losses against three LiDAR-capable baselines. Both losses are to MobileNet with the left eye off. With full vision, FLYNN- LiDAR reaches 96.6% success. Under total vision loss, the scan raises success from 45.2% to 64.6%; the baselines reach 44.6% to 56.6% with the scan on. Under out of distribution textures, it is the only LiDAR-capable model whose success improves with the scan (44.2% to 58.4%); both CNNs score 0%, and SmallWorld declines from 26.8% to 7.4%. Under periodic vision dropout, FLYNN-LiDAR achieves 93.6% and 89.6% success at 25% and 50% duty with fewer collisions. Profiling the evaluated forward paths gives 3.93M multiply accumulate operations per step; EfficientNet and MobileNet require 4.60x and 2.94x as many. Final- state analysis shows near-linear sensory superposition (cosine 0.9955).
ER -