Real-world issues in warehouse navigation

Abstract
The MDARS security robotics program has successfully demonstrated the simultaneous control of multiple robots autonomously navigating within an industrial warehouse environment. This real-world warehouse system installation required adapting a navigational paradigm designed for highly structured environments such as office corridors (with smooth walls and regularly spaced doorways) to a semi-structured warehouse environment (with few walls and within which odd-shaped objects unpredictably move about from day to day). A number of challenges, some expected and others unexpected, were encountered during this transfer of the system to the test/demonstration site. This paper examines these problems (and others previously encountered) in the historical context of the ongoing development of the navigation and other technologies needed to support the operations of a security robotic system, and the evolution of these technologies from the research lab to an operational warehouse environment. A key lesson is that a system's robustness can only be ensured by exercising its capabilities in a number of diverse operating environments, in order to (1) uncover latent system hardware deficiencies and software implementation errors not manifested in the initial system hardware or initial development environment; and (2) identify sensor modes or processing algorithms tuned too tightly to the specific characteristics of the initial development environment.

This publication has 0 references indexed in Scilit: