Enabling Opportunistic Low-cost Smart Cities By Using Tactical Edge Node Placement

TL;DR

This paper proposes a cost-effective approach for smart city IoT connectivity by leveraging public transit nodes as edge points, using algorithms to optimize coverage and latency, significantly reducing costs and delays.

Contribution

It introduces novel algorithms for opportunistic edge node placement using transit networks, improving coverage and latency over traditional centrality-based methods.

Findings

Over 77% reduction in coverage budget compared to traditional methods.

Over 20 minutes reduction in latency.

Algorithms outperform centrality metrics in real transit datasets.

Abstract

Smart city projects aim to enhance the management of city infrastructure by enabling government entities to monitor, control and maintain infrastructure efficiently through the deployment of Internet-of-things (IoT) devices. However, the financial burden associated with smart city projects is a detriment to prospective smart cities. A noteworthy factor that impacts the cost and sustainability of smart city projects is providing cellular Internet connectivity to IoT devices. In response to this problem, this paper explores the use of public transportation network nodes and mules, such as bus-stops as buses, to facilitate connectivity via device-to-device communication in order to reduce cellular connectivity costs within a smart city. The data mules convey non-urgent data from IoT devices to edge computing hardware, where data can be processed or sent to the cloud. Consequently, this paper focuses on edge node placement in smart cities that opportunistically leverage public transit networks for reducing reliance on and thus costs of cellular connectivity. We introduce an algorithm that selects a set of edge nodes that provides maximal sensor coverage and explore another that selects a set of edge nodes that provide minimal delivery delay within a budget. The algorithms are evaluated for two public transit network data-sets:Chapel Hill, North Carolina and Louisville, Kentucky. Results show that our algorithms consistently outperform edge node placement strategies that rely on traditional centrality metrics(betweenness and in-degree centrality) by over 77% reduction in coverage budget and over 20 minutes reduction in latency.