Edge Computing Methods and Applications in Wireless Seismic Sensor Networks: A Review

The centralized architecture of traditional wireless seismic sensor network (WSSN) is constrained by several bottlenecks, including high latency, high energy consumption, and the risk of single-point failures. To address these challenges, we provide a comprehensive review of the distributed intelligent processing paradigm based on edge computing (EC), which achieves real-time performance and energy efficiency optimization of the seismic exploration system by reconstructing the architecture and innovating algorithms. We focus on three layers of optimization architecture, in which the localized data processing layer employs wavelet transform (WT) and compressive sensing (CS) technology, the EC layer combines with deep learning (DL), deep reinforcement learning (DRL) and federated learning (FL), and intelligent optimization of the edge imaging layer. Subsequently, three critical challenges in edge intelligence are systematically analyzed, including latency boundary limiting complex model deployment, network jitter reducing the convergence stability of distributed learning, and multi-dimensional data heterogeneity making generalization errors fluctuate. Finally, the development prospect is elaborated to promote the continuous evolution of seismic exploration to the paradigm of independent cognition, and to provide theoretical support and technical reference for the intelligent transformation of microtremor survey.