Autonomous Quadruped Achieves 89% Coastal Coverage Delivering Rapid In-Situ Geotechnical Data
SandHound is a four-legged coastal surveying robot equipped with LiDAR, GPS, cameras, and mechanical probes. It traverses complex beach terrains to accurately measure topography, sand grain size, and sediment compaction for coastal erosion and flood modeling.
Background
Coastal surveying and geotechnical mapping are essential for understanding shoreline evolution, sediment transport, and coastal resilience. Characterizing complex terrains, including dunes, berms, and swash zones, provides critical physical data such as surface elevation, grain size distributions, and soil compaction. These parameters are vital for calibrating numerical erosion and flood models that protect coastal communities. Because coastal environments are continuously reshaped by waves, tides, and storms, researchers require reliable, repeatable methods to capture both topographic and subsurface conditions across expansive areas.
However, existing coastal monitoring approaches struggle to deliver complete datasets across these rugged landscapes. Manual human surveys are slow, labor-intensive, and expose personnel to significant safety hazards in active swash zones and steep dunes. Remote sensing alternatives, such as aerial photogrammetry from uncrewed aerial vehicles, are frequently constrained by strong coastal winds and adverse weather. Meanwhile, fixed dune-mounted LiDAR stations suffer from restricted spatial coverage caused by terrain occlusions and dense vegetation. Crucially, aerial and stationary remote sensing platforms cannot physically probe the ground, leaving a critical gap in collecting in-situ geotechnical measurements like sediment looseness and compaction resistance.
Technology
SandHound is a quadruped robotic platform built on the Ghost Robotics Vision 60 chassis for coastal topographic and geotechnical mapping. Engineered to navigate dunes, berms, and swash zones, it integrates a Velodyne VLP-16 LiDAR and Trimble R12 RTK-GPS to capture elevation data, using cloth simulation filtering to remove vegetation from digital elevation models. Additionally, it features a ZED 2 stereoscopic camera running SandSnap for grain size analysis and a Dynamic Cone Penetrometer to measure sediment compaction. Planned upgrades include thermal imaging, higher-resolution cameras, and an autonomous miniature cone penetration system.
SandHound is differentiated from traditional coastal surveying methods, such as manual surveys, wind-sensitive UAVs, and fixed LiDAR, by combining rugged mobility with simultaneous in-situ sensing. Unlike remote platforms that lack subsurface mechanical measurements or stationary systems limited by spatial blind spots, SandHound traverses loose sands to achieve superior spatial coverage (88.9% versus 84.13% for fixed LiDAR) while minimizing human safety risks. By unifying precise topography, grain size data, and soil compaction resistance within one mobile platform, it delivers critical physical parameters required for erosion and flood-risk models like XBeach.
Benefits
- Navigates complex and loose coastal terrains such as dunes, berms, and swash zones where traditional survey methods struggle
- Combines high-resolution topographic mapping, real-time grain size analysis, and in-situ soil compaction testing in a single platform
- Achieves greater spatial coverage than fixed dune-mounted LiDAR systems while maintaining high elevation accuracy through vegetation filtering
- Enhances surveyor safety and avoids wind sensitivity limitations associated with manual surveys and aerial drones
- Rapidly gathers detailed physical sediment and elevation data required to improve coastal erosion and flood-risk models
- Supports modular sensor upgrades and autonomous geotechnical probing to minimize post-processing and human intervention
Applications
- Coastal erosion monitoring
- Flood risk assessment mapping
- Beach nourishment surveying
- Coastal geotechnical site assessment
- Dune stability testing
Additional info
SandHound is an autonomous quadruped robotic platform engineered for coastal topographic and geotechnical mapping. Built on the Vision 60 chassis, it combines a Velodyne VLP-16 LiDAR and Trimble RTK-GPS to produce high-resolution digital elevation models using cloth simulation filtering. It simultaneously integrates a ZED 2 stereoscopic camera for real-time grain size analysis and an onboard Dynamic Cone Penetrometer to measure sediment compaction and bearing resistance across dynamic coastal terrains.