Why Event Cameras Are Still Not Widely Used in Industrial Robotics: A Barrier Analysis for Warehousing Logistics

Abstract

Event cameras report per-pixel intensity changes asynchronously with microsecond latency, a high dynamic range and low power consumption, which makes them a natural fit for the high-speed autonomous mobile robots emerging in warehousing logistics. Despite this fit, they remain largely absent from documented industrial mobile-robot deployments. This paper examines the structural reasons for that gap through a barrier analysis along four dimensions: data availability, tooling and integration maturity, representation heterogeneity, and application fit. Drawing on a structured review of the event-based vision literature and on practical experience, we find that warehouse-specific datasets remain scarce and narrow in speed range even though event data are abundant overall, that drivers and software stacks remain fragmented across vendor and community projects even though calibration is largely solved, that no dominant event representation has emerged, and that the class of scenarios in which event cameras offer a decisive advantage is narrower than commonly assumed. We conclude that event cameras are genuinely useful for a well-defined subset of warehousing applications, above all high-speed navigation and perception under challenging lighting, but that broader adoption depends on progress in data, interfaces and representation together with a selection criterion based on the failure mode of the incumbent sensing stack.