Edge-AI robotic swarms for predictive maintenance in utility-scale solar farms: a systematic review and meta-analysis

Abstract

The rapid expansion of utility-scale solar photovoltaic infrastructure poses critical operational challenges for maintaining reliability across large, dispersed networks. Conventional manual inspection and reactive maintenance approaches are increasingly ineffective, leading to reduced energy yield, increased downtime, and higher operational costs. This study presents a systematic review and meta-analysis evaluating the performance of autonomous robotic swarm technologies for predictive maintenance in solar farms, following PRISMA 2020 and Cochrane methodologies. A total of 20 eligible studies were analyzed using random-effects modeling, forest plot interpretation, subgroup evaluation, publication bias assessment, and sensitivity analysis. Results indicate a significant pooled effect size (Hedges g = 7.80) with substantial improvements in detection accuracy, efficiency, and cost reduction, supported by low publication bias and strong robustness. The discussion emphasizes the theoretical and practical implications of decentralized multi-agent coordination, while highlighting the limitations of heterogeneous methodologies and limited real-world validation. The study concludes that swarm robotics offers transformative potential for scalable, intelligent maintenance ecosystems and recommends future research that includes large-scale field deployments and standardized evaluation frameworks.