A centralized mapping architecture for Automated Streetlight Fault Detection via Wireless Sensor Mesh Networks

1 Introduction

Street lighting is a critical component of modern urban infrastructure, contributing significantly to road safety, public security, and the overall quality of life in smart cities. As urban populations continue to grow, municipalities are deploying increasingly large streetlight networks to ensure adequate illumination of public spaces. However, the maintenance and monitoring of these systems remain challenging due to their distributed nature and the high costs associated with routine inspections and fault diagnosis. Conventional streetlight maintenance strategies primarily rely on manual inspection and public reporting, resulting in delayed fault detection, increased operational costs, and reduced system reliability.

Recent advancements in the Internet of Things (IoT), Wireless Sensor Networks (WSNs), and smart city technologies have enabled the development of intelligent street lighting systems capable of remote monitoring and control. Various studies have proposed sensor-based solutions for monitoring energy consumption, lamp status, and environmental conditions, while wireless communication technologies such as ZigBee, LoRaWAN, and mesh networking have been employed to facilitate data transmission between distributed lighting nodes and centralized control systems. Although these approaches improve monitoring capabilities, most existing systems require dedicated sensing and communication hardware for each individual lamp post, significantly increasing deployment costs, maintenance complexity, and environmental impact.

Furthermore, current streetlight monitoring systems often focus solely on lamp status detection, neglecting critical electrical faults such as leakage currents, abnormal power consumption, and potential electrocution hazards. Such faults can pose serious safety risks to pedestrians and maintenance personnel while also contributing to energy wastage and infrastructure degradation. Consequently, there is a growing need for a cost-effective and scalable monitoring architecture capable of detecting multiple fault conditions in real time while minimizing hardware requirements.

To address these challenges, this study proposes a Centralized Mapping Architecture for Automated Streetlight Fault Detection via Wireless Sensor Mesh Networks. The proposed architecture employs a wireless sensor mesh network to continuously monitor the electrical characteristics of streetlight clusters and transmit real-time operational data to a centralized monitoring station. Unlike conventional approaches that require dedicated sensing units on every lamp post, the proposed system utilizes a single sensor node to monitor a group of ten streetlights through aggregated power consumption analysis. By analyzing the collective electrical behavior of each streetlight cluster, the system can identify abnormal operating conditions, including faulty bulbs, leakage currents, and potential electrocution risks, thereby significantly reducing hardware deployment costs.

To enhance deployment sustainability, the proposed sensor nodes are powered using standalone solar energy systems, eliminating the need for additional grid power infrastructure and reducing the overall carbon footprint of the monitoring network. The mesh networking architecture further improves system scalability and communication reliability by enabling sensor nodes to relay information through neighboring nodes, thereby extending network coverage in large urban environments. Additionally, a centralized Geographic Information System (GIS)-based mapping interface is integrated into the architecture to provide real-time visualization of streetlight faults. This allows maintenance personnel to accurately identify the geographical location of a fault and rapidly dispatch corrective actions, thereby reducing maintenance response times and improving operational efficiency.

The main contributions of this work are summarized as follows:

• Development of a centralized monitoring architecture for real-time streetlight fault detection using wireless sensor mesh networks.
• Introduction of a cluster-based monitoring strategy in which a single sensor node supervises the electrical behavior of ten streetlights, substantially reducing hardware and deployment costs.
• Real-time detection of multiple fault conditions, including faulty lamps, leakage currents, abnormal power consumption, and potential electrocution hazards.
• Integration of a GIS-based mapping platform for accurate fault localization and maintenance management.
• Implementation of solar-powered sensor nodes to support sustainable and energy-efficient operation while minimizing environmental impact.
• Enhancement of network scalability and communication robustness through mesh-network-based data transmission.

The proposed architecture aims to provide municipalities and smart city operators with a practical, scalable, and cost-effective solution for intelligent streetlight infrastructure management. By combining centralized fault monitoring, wireless mesh communication, renewable energy integration, and geographic fault visualization, the system offers a comprehensive framework for improving the reliability, safety, and sustainability of future urban lighting networks.

2 Literature review and survey

2.1 Wireless sensor networks for infrastructure monitoring

Wireless Sensor Networks (WSNs) have emerged as a key enabling technology for large-scale infrastructure monitoring due to their low deployment cost, scalability, and ability to provide real-time data acquisition. Akyildiz and Vuran. (2010), Karl and Willig (2005), Sohraby et al. (2007) established the theoretical foundations of WSN architectures, communication protocols, and deployment strategies for distributed monitoring applications. These studies demonstrated that sensor networks can effectively collect and transmit environmental and operational data from geographically dispersed infrastructures.

Fault detection within WSNs has been extensively studied to improve network reliability and operational continuity. Wang et al. (2008) investigated fault detection mechanisms in sensor networks, while Mahapatro and Khilar (2013) presented a comprehensive survey of fault diagnosis techniques for WSN environments. Ammari et al. (2020) further reviewed fault management frameworks and highlighted the importance of autonomous fault detection in large-scale sensor deployments. Xu et al. (2014) proposed the LEDFD algorithm to reduce energy consumption while maintaining effective fault detection capabilities. Similarly, Asim et al. (2010) introduced a self-managing fault management mechanism for improving network resilience under node failures. Security challenges and reliability concerns associated with WSN deployments were comprehensively discussed by Padmavathi and Shanmugapriya (2009).

These studies demonstrate the effectiveness of WSNs for distributed monitoring applications; however, most existing approaches focus primarily on sensor node failures rather than monitoring electrical infrastructure faults such as leakage currents, electrocution hazards, and streetlight malfunctions.

2.2 Smart street lighting systems

The rapid growth of smart city initiatives has accelerated research on intelligent street lighting systems. Early work by Jing et al. (2007) proposed a wireless sensor network-based streetlight monitoring and control framework capable of remote operation and management. Martirano (2011) developed an intelligent lighting control system aimed at reducing energy consumption in urban environments.

Several researchers have investigated energy-efficient streetlight control strategies. Kaleem et al. (2014) introduced a ZigBee-based smart LED streetlight system, while Toubal et al. (2016) proposed an energy-efficient streetlight control architecture utilizing wireless sensor technologies. Shahzad et al. (2016) developed a traffic-adaptive street lighting system capable of dynamically adjusting illumination levels based on vehicular activity. Abdullah et al. (2019), Kim et al. (2020) further demonstrated the effectiveness of sensor-driven adaptive lighting systems in reducing energy consumption.

Comprehensive reviews by Pizzuti et al. (2013), Mahoor et al. (2020) summarized the evolution of smart streetlight technologies and emphasized the need for scalable monitoring infrastructures capable of supporting large urban deployments. Although these systems successfully improve energy efficiency and lighting control, they generally require dedicated sensing hardware for each lamp post, increasing implementation cost and maintenance complexity.

2.3 IoT and smart city integration

The integration of IoT technologies has significantly expanded the capabilities of urban monitoring systems. Ejaz et al. (2017) highlighted the role of IoT-enabled energy management systems in smart cities, while Talari et al. (2017) reviewed the broader application of IoT technologies in urban infrastructure management. Stallings (2021) discussed modern IoT architectures and networking principles that facilitate large-scale real-time monitoring systems. Giudice (2018) further emphasized the importance of interconnected sensing infrastructures in enabling sustainable smart city development.

Castro et al. (2013) proposed smart lighting solutions integrated within urban IoT ecosystems, demonstrating how lighting infrastructures can serve as intelligent sensing platforms. Jain et al. (2019) extended this concept by integrating deep learning techniques into smart city management systems, enabling predictive analytics and automated decision-making.

Despite these advancements, many IoT-based streetlight monitoring systems remain focused on energy optimization and illumination control, while comprehensive electrical fault monitoring and geographic fault localization remain relatively unexplored.

2.4 Fault detection and diagnostic systems for street lighting

Fault diagnosis has become an essential requirement for maintaining reliable urban lighting infrastructure. Lee and Huang (2015) developed a non-invasive fault detection system capable of identifying defective streetlights through electrical parameter monitoring. Kumar et al. (2016) proposed a mobile sensor-based monitoring platform for urban lighting infrastructure inspection. Nanduri et al. (2020) introduced an IoT-based damaged streetlight detection system capable of identifying lamp failures remotely.

Advanced fault diagnosis methodologies have also been investigated. Velasquez and Passino (2015) employed fuzzy logic techniques for fault-tolerant lighting control, while Seabra et al. (2016) developed an IoT-based intelligent diagnostic framework for fault detection applications. Dahan et al. (2020) identified several challenges associated with designing reliable smart lighting systems, including communication reliability, maintenance requirements, and deployment costs.

Although these systems improve fault identification capabilities, most focus exclusively on bulb failures or lighting control. Few studies address broader electrical safety concerns such as leakage currents and electrocution hazards while simultaneously providing centralized visualization and localization capabilities.

2.5 Wireless mesh networking for smart lighting applications

Wireless mesh networking has gained considerable attention due to its scalability, fault tolerance, and self-healing communication characteristics. Pantoni and Brandao (2013) proposed a gradient-based routing scheme specifically designed for streetlight sensor networks. Song et al. (2020) investigated mesh networking architectures for smart streetlight communication and demonstrated improved coverage and reliability compared to traditional star-based topologies.

Mesh-based communication infrastructures offer several advantages for urban monitoring systems, including extended communication range, redundancy, and improved network resilience. These characteristics make mesh networks particularly suitable for large-scale streetlight deployments where direct communication with a centralized station may not always be feasible.

2.6 Research gap and motivation

The literature review reveals significant advancements in smart street lighting, wireless sensor networks, IoT-enabled monitoring, and fault diagnosis systems. However, several limitations remain. First, most existing systems require dedicated sensing hardware for each individual lamp post, increasing deployment costs and environmental impact. Second, current solutions primarily focus on lamp status monitoring and energy optimization while providing limited support for detecting leakage currents, abnormal power consumption, and potential electrocution hazards. Third, few studies integrate real-time geographic fault mapping to facilitate rapid maintenance response and infrastructure management. Finally, the use of solar-powered monitoring nodes for sustainable operation remains insufficiently explored.

To address these limitations, the present work proposes a Centralized Mapping Architecture for Automated Streetlight Fault Detection via Wireless Sensor Mesh Networks. The proposed system employs a cluster-based monitoring strategy in which a single sensor node supervises the electrical behavior of multiple streetlights, thereby reducing hardware requirements and deployment costs. Furthermore, the architecture integrates wireless mesh communication, solar-powered sensor nodes, real-time fault diagnosis, and GIS-based fault localization to provide a scalable, sustainable, and cost-effective solution for smart city streetlight management as shown in Figure 1.

3 Methodology

The proposed architecture relies on a decentralized Wireless Sensor Network (WSN) deployed in a multi-hop mesh topology. The system leverages microcontrollers (e.g., ESP32) at each node to continuously monitor electrical parameters. To ensure diagnostic accuracy, the network first establishes a statistical baseline using a subset of fully functional nodes, which serves as the comparative threshold for real-time fault detection across the entire grid.

3.1 System architecture and baseline calibration

Let the network be represented by a directed graph G = ( V , E ), where V = { v 0 , v 1 , v 2 , … , v N } is the set of nodes and E represents the set of available communication links. Node v 0 serves as the central control station (sink), while V * = V \ { v 0 } represents the street light nodes.

The system initializes by monitoring a predefined baseline set of 10 fully operational street lights, denoted as B ⊂ V *, where | B | = 10. For each node v b ∈ B, the sensor suite records an operational state vector at discrete time t:where V ( t ) is the node voltage, I i n ( t ) is the incoming current, and I out ( t ) is the outgoing current. Over a calibration period T, the central station computes the nominal expected state vector μ base and the covariance matrix Σ base:

X b t = V t I i n t I out t (1)

μ base = 1 10 T ∑ b ∈ B ∑ t = 1 T X b t (2)

Σ base = 1 10 T − 1 ∑ b ∈ B ∑ t = 1 T X b t − μ base X b t − μ base T (3)

3.2 Mathematical formulation of fault detection

During standard operation, any node v i ∈ V * (Equations 1–3) continuously samples its state vector X i ( t ). To detect deviations from the baseline, we utilize the Mahalanobis distance, which accounts for the variance and covariance of the baseline electrical parameters:

D M X i t = X i t − μ base T Σ base − 1 X i t − μ base (4)

A primary fault flag is triggered if D M ( X i ( t ) ) > γ, where γ is a statistically derived threshold. Upon triggering, the ESP32 microcontroller classifies the fault using deterministic residual equations based on Kirchhoff’s laws (Equation 4). Let Δ I ( t ) = | I i n ( t ) − I out ( t ) | denote the residual differential current.

• Faulty Bulb (Open Circuit): Characterized by a near-zero current draw while supply voltage remains nominal.

F bulb t = 1 , if I i n t ≤ ϵ I ∧ | V t − V nom | ≤ ϵ V 0 , otherwise (5)

• Leakage Current: Characterized by a steady differential current exceeding safe limits, without triggering an overcurrent trip.

F leak t = 1 , if I leak t h ≤ Δ I t < I short t h 0 , otherwise (6)

• Earthing/Ground Fault: Characterized by a severe, transient differential current coupled with an instantaneous voltage drop.

F earth t = 1 , if Δ I t ≥ I short t h ∧ d V t d t ≤ − κ 0 , otherwise (7)

3.3 Resilient multi-hop data transmission protocol

When a fault F type ( t ) = 1 (Equations 5–7) is detected at a node v i (e.g., Node A), the diagnostic payload P A is generated. To preserve energy and overcome distance limitations, data is transmitted to the central station v 0 via a multi-hop routing protocol (e.g., A → B → C → v 0).

Each intermediate node acts as both a sensor and a relay. Let R i be the routing table at node v i. The total transmitted payload M i from node v i to its next-hop neighbor v j is the union of its own data and the aggregated data from its downstream children C i:

M i = P i ∪ ⋃ k ∈ C i M k (8)

To ensure fault tolerance, the network utilizes a dynamic cost function to establish routing paths. The edge weight w ( e i j ) for a link from v i to v j is inversely proportional to the Link Quality Indicator (LQI) and the residual energy E res of node v j:where α and β are weighting coefficients. If an intermediate node (e.g., Node B) fails, the periodic beaconing process will detect an timeout t ack > T timeout. Node A will dynamically update its routing table to skip Node B, selecting the next optimal neighbor (e.g., Node D) that minimizes the path cost to v 0:where N ( v i ) is the active neighbor set of v i, and C ( v j , v 0 ) is the cumulative cost from v j to the central station.

w e i j = α 1 LQI i j + β E max E res,j (9)

NextHop v i = arg min v j ∈ N v i w e i j + C v j , v 0 (10)

3.4 Reliability and system availability analysis

The reliability of the mesh network is modeled using a probabilistic approach (Equations 8–10). Let p i j be the probability of successful transmission across edge e i j. The reliability of a single path Π k comprising m hops is:

R path Π k = ∏ u , v ∈ Π k p u v (11)

Because the mesh topology allows for K independent or disjoint paths from a sensor node to the central station, the overall transmission reliability R sys ( v i ) from node v i becomes highly fault-tolerant. The probability that the data successfully reaches v 0 despite node failures is bounded by:

R sys v i = 1 − ∏ k = 1 K 1 − R path Π k (12)

This formulation mathematically guarantees that as the node density increases, R sys → 1, ensuring resilient data delivery even under cascading hardware failures (Equations 11 , 12).

3.5 Covariance estimation, threshold selection, and robustness

Equation 4 depends critically on a well-conditioned, statistically sound estimate of μ base and Σ base, on a principled choice of the detection threshold γ, and on robustness to sensor and ADC noise. This section details these three aspects, along with the computational cost of evaluating Equation 4 on the ESP32 platform.

3.5.1 Feature vector and sample size

To improve discriminability against transient noise, the raw instantaneous state vector in Equation 1 is extended to a d-dimensional feature vector x b ( ℓ ) ∈ R d (d = 7, as summarized in Table 1) computed over each observation window of N W = 500 samples (spanning W = 5 mains cycles at f s = 5 kHz):where V rms and I rms are RMS voltage/current, T H D V and T H D I are total harmonic distortion up to harmonic order H = 10, cos ⁡ ϕ is the power factor, and Δ I rms is the RMS residual current. Windowing and harmonic averaging suppress single-sample ADC noise and transient switching spikes prior to distance computation. With K = 10 baseline luminaires and M 0 = 1,440 windows collected per luminaire during commissioning, the baseline sample size is n base = K × M 0 = 14,400 ≫ d, which is sufficient to estimate a 7 × 7 covariance matrix reliably (a common rule of thumb requires n base ≳ 10 d–20 d; here n base / d ≈ 2057).

x b ℓ = V rms I i n , r m s I out,rms T H D V T H D I cos ⁡ ϕ Δ I rms T (13)

TABLE 1. CSFDS design parameters and nominal values.

3.5.2 Covariance estimation and regularization

The baseline covariance matrix is estimated as in Equation 3. To guard against ill-conditioning (e.g., near-collinear features such as V rms and cos ⁡ ϕ under stable grid conditions) and to prevent an unstable or non-invertible Σ base, we apply Ledoit–Wolf shrinkage regularization:where λ ∈ [ 0,1 ] is the shrinkage intensity (estimated analytically from the baseline data via the Ledoit–Wolf formula) and ν is the average eigenvalue used as an isotropic target. This guarantees Σ ̂ base is well-conditioned (bounded condition number) and invertible even if individual features are highly correlated during calibration. Σ ̂ base − 1 is computed once, offline, at the central station (v 0) during the calibration phase and the resulting 7 × 7 inverse matrix (49 floating-point values, 196 bytes) is then pushed to each ESP32 node matrix inversion itself is never performed on the sensor node (Equation 14).

Σ ̂ base = 1 − λ Σ base + λ ν I d , ν = t r Σ base d (14)

3.5.3 Threshold selection

Under the approximate multivariate-Gaussian assumption for the baseline electrical parameters, the squared Mahalanobis distance D M 2 ( X i ( t ) ) is asymptotically chi-squared distributed with d degrees of freedom, D M 2 ∼ χ d 2. The threshold γ is therefore chosen analytically to achieve a target false-alarm probability α (Table 1: α = 1 0 − 3):where χ d , 1 − α 2 is the ( 1 − α )-quantile of the chi-squared distribution with d degrees of freedom. For d = 7 and α = 1 0 − 3, χ 7 , 0.999 2 ≈ 24.32, giving γ = 24.32 ≈ 4.93, consistent with the value reported in Table 1. This closed-form threshold avoids ad hoc tuning and allows γ to be recomputed automatically if d or α are changed for a different deployment (Equation 15).

γ = χ d , 1 − α 2 (15)

3.5.4 Robustness to noisy measurements

Three mechanisms are used to make the detector robust to sensor and environmental noise: (i) window-based averaging, where each feature in Equation 13 is computed over N W = 500 samples rather than a single ADC reading, suppressing zero-mean quantization and thermal noise by a factor of approximately N W; (ii) persistence filtering, where a fault flag is only confirmed after D M ( t ) > γ is observed for W fail consecutive windows (analogous to the beacon-loss criterion in Table 1), rejecting single-window outliers caused by transient grid disturbances; and (iii) shrinkage-regularized covariance estimation (Equation 14), which prevents the inverse covariance from amplifying noise along low-variance feature directions, a known failure mode of the unregularized Mahalanobis distance when Σ base is near-singular.

3.5.5 Computational complexity on ESP32

Because Σ ̂ base − 1 is precomputed offline and distributed to nodes as a constant d × d matrix, the online cost of Equation 4 reduces to one matrix–vector product and one dot product per window:which requires O ( d 2 ) multiply-accumulate operations, 49 multiplications and 42 additions for d = 7, plus a single square root. On the ESP32 (dual-core Xtensa LX6, 240 MHz, hardware floating-point unit), this computation executes in well under 10 μ s, and the constant memory footprint (196 B for Σ ̂ base − 1 plus 28 B for μ base) is negligible relative to the ESP32’s available SRAM. Consequently, the anomaly-detection step contributes negligibly to the node’s active-mode energy budget (Figure 2) and does not affect the δ * = 0.233 duty cycle or the 7.5-day backup duration reported in Table 1.

D M 2 X i t = δ T Σ ̂ base − 1 δ , δ = X i t − μ base (16)

3.6 Power consumption model

Given the deployment of low-power ESP32 microcontrollers, energy optimization is paramount. The total energy consumed by node v i over a time period T is the sum of sensing, processing, transmitting, receiving, and idling energies:where N t x and N r x are the number of transmitted and received packets, respectively. By aggressively utilizing deep-sleep modes between sampling intervals and only transmitting aggregated M i payloads upon fault detection or scheduled heartbeats, E t x and E r x are minimized, drastically extending the lifespan of the WSN infrastructure (Equation 17). Algorithm 1 outlines the step-by-step execution of the proposed automated fault detection and dynamic routing protocol.

E total i = E sens + E proc + N t x E t x + N r x E r x + E idle (17)

Algorithm 1 Automated Fault Detection and Resilient Multi-Hop Routing.

As detailed in Algorithm 1, the operational lifecycle of the network is divided into a calibration phase and a continuous monitoring phase. Initially, the system aggregates data from 10 established, fully functional streetlights to formulate a highly accurate statistical baseline, represented by the mean vector μ base and covariance matrix Σ base. During runtime, each ESP32 node continuously samples its voltage and current metrics, applying the Mahalanobis distance to identify anomalies without requiring heavy computational overhead. If the distance metric D M exceeds the calibrated threshold γ, the node utilizes residual current equations to instantly classify the anomaly as a bulb failure, leakage, or earthing fault. Upon classification, the node shifts from a power-saving deep-sleep state to an active transmission state. The algorithm guarantees data delivery to the central station ( v 0 ) by calculating dynamic edge weights based on link quality (LQI) and the residual energy of neighboring nodes. Crucially, the routing loop incorporates a timeout mechanism (Lines 22-25); if an intermediate relay node fails to acknowledge receipt, the algorithm dynamically prunes the failed node from the routing table and recalculates an alternate optimal path, ensuring robust fault tolerance and preventing data loss within the smart city infrastructure.

3.7 Summary of key system parameters

The CSFDS design parameters outlined in Table 1 define a robust, low-power wireless sensor network. The architecture balances high-frequency environmental sampling ( f s = 5 kHz ) with a low-duty-cycle TDMA networking protocol ( δ * = 0.233 ). This efficiency minimizes the average node power to 184.6 mW, ensuring a resilient 7.5-day backup duration under typical urban channel conditions.

4 Results and discussion

The proposed Centralized Mapping Architecture for Automated Streetlight Fault Detection via Wireless Sensor Mesh Networks was evaluated through a comprehensive simulation environment designed to emulate a real-world urban street lighting deployment. The results demonstrate that the proposed system is capable of accurately detecting multiple fault conditions, including faulty bulbs, leakage currents, and ground faults, with an overall detection accuracy exceeding 97%. The Mahalanobis distance-based anomaly detection mechanism successfully identified abnormal electrical behavior while maintaining a low false alarm rate, thereby improving the reliability of the monitoring framework. The mesh communication architecture exhibited excellent network performance, achieving a packet delivery ratio of approximately 99% while maintaining low communication latency. Furthermore, the self-healing routing mechanism effectively rerouted diagnostic data through alternate communication paths when intermediate nodes failed, ensuring uninterrupted data transmission to the centralized monitoring station. Reliability analysis revealed that the overall network availability increased with node density, confirming the robustness of the proposed multi-hop mesh topology for large-scale smart city deployments as shown in Figures 3–6.

Energy consumption analysis demonstrated the effectiveness of the solar-powered sensor nodes and low-duty-cycle operation strategy. By utilizing deep-sleep modes and event-driven communication, the average node power consumption remained low, enabling extended autonomous operation without additional grid power requirements. In addition, the cluster-based monitoring approach significantly reduced hardware deployment costs by allowing a single sensor node to supervise a group of ten streetlights. Compared with conventional one-sensor-per-pole architectures, the proposed system reduced sensor hardware requirements by approximately 90%, leading to substantial reductions in implementation cost and environmental impact. The GIS-based fault localization module successfully identified and visualized fault locations in real time, enabling maintenance personnel to rapidly locate defective streetlight clusters. This capability is expected to reduce maintenance response times and improve overall operational efficiency. Collectively, the simulation results validate the feasibility of the proposed architecture as a scalable, cost-effective, energy-efficient, and reliable solution for next-generation smart streetlight monitoring and fault management systems as shown in Figures 2, 7–9.

The correlation heatmap Figure 10 reveals that voltage stability and current stability exhibit strong positive relationships with fault detection accuracy and network reliability. Furthermore, mesh connectivity significantly influences packet delivery ratio and fault localization performance, confirming the effectiveness of the proposed multi-hop communication architecture. The results indicate that stable sensor measurements contribute directly to improved fault diagnosis accuracy, while solar-powered operation and battery health positively impact overall energy efficiency and long-term system sustainability.

Table 2 highlights the advantages of the proposed architecture compared with existing streetlight monitoring and fault management systems. Most previously reported approaches primarily focus on energy-efficient lighting control, basic fault monitoring, or communication network design. However, the majority of these systems require dedicated sensing hardware for individual streetlight poles, resulting in increased deployment cost and maintenance complexity. Furthermore, leakage current detection, electrocution risk monitoring, GIS-based fault localization, and renewable energy integration are generally absent from existing solutions.

TABLE 2. Detailed comparison of the proposed centralized mapping architecture with existing state-of-the-art streetlight monitoring and fault detection systems.

In contrast, the proposed architecture combines wireless sensor mesh networking, centralized GIS-based fault mapping, real-time leakage current monitoring, and solar-powered operation within a unified framework. Moreover, the cluster-based monitoring strategy enables a single sensor node to supervise up to ten streetlights simultaneously, reducing sensor hardware requirements by approximately 90% compared with conventional one-sensor-per-pole architectures. These features collectively improve scalability, sustainability, fault detection capability, and deployment feasibility for future smart city streetlight infrastructures.

The hardware experimental setup as shown in Figure 11 comprises three primary units designed to validate the proposed monitoring system. Figure 11a shows the initial gateway development utilizing an Arduino microcontroller integrated with a Bolt IoT wireless module for data logging. The core sensor node prototype is illustrated in Figure 11b, featuring an ESP32 microcontroller interfaced with a switching relay module and an LED array to simulate streetlight loads. Lastly, Figure 11c depicts the power regulation circuit and signal-conditioning modules, including a switched-mode power supply (SMPS) and a digital status display for localized debugging.

5 Conclusion

This paper proposed a centralized wireless sensor mesh network architecture for real-time streetlight fault detection and monitoring. The system combines fault diagnosis, GIS-based fault localization, solar-powered operation, and mesh networking within a unified framework. By enabling a single sensor node to monitor multiple streetlights, the proposed approach substantially reduces hardware requirements and deployment costs. Simulation results confirmed high fault detection accuracy, reliable communication performance, and improved operational efficiency. The proposed architecture demonstrates strong potential for scalable and sustainable smart city streetlight management.

6 Future scope

Future work will focus on implementing a full-scale hardware prototype and validating the proposed architecture under real-world operating conditions. Advanced machine learning algorithms may be incorporated to enable predictive maintenance and fault forecasting. Additionally, future studies will explore integration with digital twin platforms, edge computing technologies, and large-scale smart city infrastructures to further enhance monitoring accuracy, system intelligence, and operational efficiency.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Sajal Biring, Ming Chi University of Technology, Taiwan

Reviewed by: Surajit Chattopadhyay, Ghani Khan Choudhury Institute of Engineering and Technology (GKCIT), India

Chittajit Sarkar, Islampur College, India