Accepted paper at ACM SIGSPATIAL 2022 Workshop IWCTS 2022

Authors

Lukas Rottkamp, Niklas Strauss, Matthias Schubert

The 15th International Workshop on Computational Transportation Science (IWCTS 2022) co-located with the 30th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM SIGSPATIAL2022), 
01–04 November 2022, Seattle, Washington, USA

Abstract

When placing sensors in an environment, it may not be possible to directly cover all entities of interest with sensors due to cost or other restraints. This leads to a sensor placement problem in which only a subset of all sensible sensor locations is equipped with sensors. If data concerning the system to be measured is already available or easily procured, sensor locations can be selected in a data-driven approach. Without data, alternative methods have to be applied. In this paper, we present and compare various data-driven and data-agnostic methods for selecting parking sensor locations in a city environment. Experiments using real-world data show that methods only requiring parking bays' locations compare reasonable well to data-driven approaches requiring environment data which may be expensive to acquire.