Poster abstract: Detecting faulty street lamps with illumination maps

Abstract

Badly lit roads usually lead to vehicle accidents and encourage crime in the area. Therefore, it is important to detect faulty street lamps rapidly and report them to related authorities to keep roads safe. Currently, communities still mostly depend on electrical inspectors to check street lamps regularly, which may result in long time delays for repair. Recent studies focus on add networking capability into street lamp poles to enable real-time reports on the healthy status of lamps. However, such a smart system increases costs to add sensors and network modules in every street lamp; therefore, it is nearly impossible to realize this kind of system in a short term. We propose a new method to detect faulty lamps. We designed equipment that could be installed on fixed bus routes and collect the lighting intensity along the routes. We created illumination maps in meter-level resolution. The differences between illumination maps created at different times can help identify the changes of lighting intensity in specific locations. We executed a proof-of-concept experiment that shows our method is feasible. This method can be extended to a citywide scale at low cost. As a result, this would detect faulty street lamps along main roads and prevent accidents and crime by shortening the duration of badly lit streets.