SandHound: A Quadruped Robotic Platform for Integrated Coastal Topographic and Geotechnical Mapping
Authors:
- Arthur TrembanisArthur Trembanis
- Jacob StasiewiczJSJacob Stasiewicz
- Jack PuleoJPJack Puleo
- David YoungDYDavid Young
- Brian McFallBMBrian McFall
- Gaurav SavantGSGaurav Savant
- Nicholas SporeNSNicholas Spore
Preprint
- 2 February 2026
- preprint
- Published by Elsevier in SSRN Electronic Journal
Abstract
Coastal zones are critical regions that support vital economic and ecological systems, yet they are increasingly threatened by erosion, sea-level rise, and extreme weather. Effective management of these environments can be supported by high-resolution, co-registered topographic and sedimentological datasets. Conventional methods to obtain subaerial topographic datasets such as ground-based real-time kinematic GPS surveys, Unmanned Aerial System (UAS) based Structure-from-Motion (SfM) photogrammetry, and fixed position light detection and ranging (lidar) are limited by safety, mobility, and the inability to capture both elevation and sediment properties simultaneously in dynamic coastal settings. We introduce SandHound, a quadruped unmanned ground vehicle (UGV-Q), modified and operated to conduct a complete and multi-faceted survey in the subaerial beach zone. SandHound’s integrated payload includes lidar for topographic mapping, and stereoscopic cameras for onboard grainsize prediction via SandSnap. Field experiments at the U.S. Army Corps of Engineers Field Research Facility (FRF) in Duck, North Carolina, were conducted to validate the system’s core functions. The evaluation focused on three components: (1) accuracy of SandHound’s topographic data compared to SfM and fixed position lidar, (2) effectiveness of the SandSnap system validated against physical sediment samples (Camsizer), and (3) assessmentof system mobility across a study area characterized by varied near-surface compaction (measured using external Dynamic Cone Penetrometer-DCP-tests). SandHound achieved comparable or improved elevation accuracy (RMSE ∼0.15 m) when compared to the fixed position lidar as a ground truth, out-performing UAS-based SfM (RMSE ∼ 0.24 m). SandHound achieved spatial coverage of (∼ 89%) of the study area, and reliable operation across the range of measured sediment conditions. Site-specific tuning of SandSnap yielded a median absolute percent error of ∼ 13.4% in median grain size (D50) when compared with Camsizer analyses, while DCP values from 0.47–5.02 MPa confirmed reliable quadruped operation across loose to moderately compact sediment. These results demonstrate that UGV-Q platforms can provide a safe, mobile, and efficient means of integrated topographic and sedimentological measurements, offering a robust new tool for coastal monitoring, erosion assessment, and post-storm reconnaissance.Keywords
This publication has 11 references indexed in Scilit:
- SandSnap filtering techniques2025
- Measurement of Changes in Beach Sand Soil Stiffness due to Fluctuating TidesJournal of Waterway, Port, Coastal, and Ocean Engineering, 2025
- SandSnap: Measuring and mapping beach grain size using crowd-sourced smartphone imagesCoastal Engineering, 2024
- SediNet: a configurable deep learning model for mixed qualitative and quantitative optical granulometryEarth Surface Processes and Landforms, 2019
- Coastal Lidar and Radar Imaging System (CLARIS) lidar data report : 2011 - 20172019
- Investigation of spatial and short-term temporal nearshore sandy sediment strength using a portable free fall penetrometerCoastal Engineering, 2019
- An Easy-to-Use Airborne LiDAR Data Filtering Method Based on Cloth SimulationRemote Sensing, 2016
- Modelling storm impacts on beaches, dunes and barrier islandsCoastal Engineering, 2009
- Developing Terrestrial-LIDAR-Based Digital Elevation Models for Monitoring Beach Nourishment PerformanceJournal of Coastal Research, 2008
- Dynamics of marine sands1997