Optimising Antenna Positioning for Maximum Coverage: The Case Study of Cattle Tracking in Austrian Alps Using Long Range (LoRa) Based Monitoring System

Abstract

This paper presents an approach to solve the antenna coverage location problem (ACLP) in the context of tracking cattle in the Austrian Alps. In cooperation with ViehFinder a process is determined for optimizing the placement of the antennas using a DEM with a resolution of 1m. The paper identifies several constraints, such as mobile reception, proximity to a street and a maximum slope, that influence the number of candidate positions. Further, the demand nodes are restricted to the alp areas. Of the possible objectives for the ACLP, such as (1) complete coverage, (2) maximum coverage with a budget and (3) backup coverage, this paper focuses on maximizing the coverage in respect to a given budget. The optimization model discussed here uses visibility analyses to determine the viewsheds - i.e. the covered area - of the antennas and Integer Linear Programming (ILP) to solve the locational problem of placing the antennas. It will be applied to two study areas. The first study area is located near Graz in Styria containing about 277.3 million candidate positions and demand points. The second is close to the western border of Carinthia and consists of 47.5 million demand nodes and possible candidate locations. Due to the size of the problem computational problems might result, when using ILP. Therefore the use of genetic algorithms and heuristics is considered as well. Especially NSGA-II that has been suggested as one of the viable solutions for problems of similar nature and complexity. The sheer size of the problem is unprecedented, thus future work will explore the boundaries of the proposed approach.