Integrating Smart Rocks, UAV-Based Sensing, and Large-Scale Field Experiments for Bedload Transport Analysis in River Systems

Bedload transport is vital for river dynamics but difficult to measure directly. To address this, we conducted a large-scale field experiment at Landaoxi in the HueiShui Forest Station, featuring a 25-meter-long, 2-meter-wide channel with significant inflow (~1.5 cms) to replicate realistic river conditions. This large-scale setup enabled us to capture complex sediment transport behaviors that are often unobservable in smaller laboratory environments. Six UAVs were used to record the entire experiment, applying Particle Tracking Velocimetry (PTV) to capture surface flow velocity. Additionally, we reconstructed the flow surface using Structure-from-Motion (SfM) techniques.

This study emphasizes the deployment of smart rock technology for direct tracking of bedload motion. Each smart rock, built on Arduino-based controllers, incorporates a 9-axis accelerometer, timer, and SD card reader. The shells were fabricated through 3D printing and molded with epoxy-sand mixtures to match the density of natural particles. Signal processing of accelerometer data facilitated reconstruction of particle paths, velocities, and dynamic responses. Path reconstruction involves integrating the accelerometer data twice to estimate velocity and position. To mitigate drift errors inherent in integration, we applied noise filtering, coordinate alignment using gyroscopic data, and final position correction based on known retrieval locations. Advanced techniques, such as Kalman filtering and cross-validation with UAV-based flow data, were employed to enhance accuracy and ensure reliable path visualization.

The smart rock measurements were compared with (1) seismic signal monitoring and (2) optical fiber sensing. This multi-method approach within a large-scale experimental framework provides new insights into sediment transport processes in river systems.