DART: A High-Rate Data Acquisition Protocol With Relative Topology Reconstruction for Sensor Networks

Abstract

Dense wired sensor networks are required in applications where many sensing points must be sampled at high rates with deterministic timing, and high reliability. Conventional fieldbuses provide robust connectivity, but request / response polling limits acquisition rates. Higher-throughput alternatives can address these limitations, but typically require additional hardware, increasing cost and integration complexity. This paper presents DART, a high frequency data acquisition protocol for dense wired sensing systems. DART replaces shared bus polling with cascaded point-to-point UART links coordinated by trigger signals. Each node forwards an aggregate packet upstream and collects data from up to two downstream branches, forming a deterministic data acquisition process that supports both linear and binary tree topologies using at most two UART interfaces per node. In addition, DART embeds compact topology information in the acquisition stream, allowing the data logger to reconstruct the relative parent child layout of the network. The protocol is evaluated using an ESP32 based testbed and a simulation based scalability analysis across different network sizes, payloads, and topologies. The proposed solution offers a cost-effective, scalable architecture for dense sensing, bridging the gap between low speed fieldbuses and complex high bandwidth networks.