MoL-2023-23: How Many Oracle Calls Does It Take to Locate a Lightbulb?

MoL-2023-23: Danco, Dominique Marie (2023) How Many Oracle Calls Does It Take to Locate a Lightbulb? [Report]

Text MoL-2023-23.text.pdf - Published Version Download (518kB)

Abstract

Wireless nodes in smart buildings have many benefits, but one of their drawbacks is the added complication of mapping the nodes on a floor plan after installation. Currently, it is common practice to manually record the location of each placed node. This is labor intensive and prone to error. In this project, we attempt to create a system that will reduce this manual labor by locating the nodes using the Received Signal Strength Indicator (RSSI) data that each node receives from its neighbors. By comparing the possible locations of each node with the strength of these signals, we can craft a constraint solving problem to determine the most likely mappings of the nodes. RSSI data is notoriously inaccurate and inconsistent, so we have built a system that combines constraint solving and human input, and has a high tolerance for error. Our system has been shown to suggest the correct solution after asking a small number of questions to the user, to clarify between symmetric placement options or ambiguous node placement due to inconsistent signal strength. While the system will need improvements to scale out to scenarios with hundreds or thousands of nodes, it is a successful proof of concept that demonstrates that a balance can be struck between automatic and manual strategies, given a constraint problem with unreliable data.

Item Type:	Report
Report Nr:	MoL-2023-23
Series Name:	Master of Logic Thesis (MoL) Series
Year:	2023
Subjects:	Computation Logic
Depositing User:	Dr Marco Vervoort
Date Deposited:	26 Sep 2023 13:13
Last Modified:	26 Sep 2023 13:13
URI:	https://eprints.illc.uva.nl/id/eprint/2273

Actions (login required)

View Item