Ground Vehicle Systems Center (GVSC) is currently tasked to develop the Next-Generation NATO Reference Mobility Model (NG-NRMM). NG-NRMM is a simulation capability to compute and display the mobility on deformable soil as a map of GO/NOGO capability. An essential input or the NG-NRMM model is the terrain mechanical properties. Traditionally, the terrain mechanical properties are determined by direct measurements of soil bearing capacity using cone penetration tests. However, for military operations in hostile territories, the direct measurement could place the soldier in harm's way. So, developing alternate approaches to characterize the terrain's mechanical properties is critical for the U.S. Army. Remote sensing is an alternate approach to the direct measurement for determining terrain mechanical properties that need to be explored. Previous research by Dr. Oommen's team has demonstrated that thermal remote sensing can characterize the soil strength, and hyperspectral remote sensing can be used for soil type characterization. The team also explored the power of combining data from thermal and hyperspectral sensing utilizing machine learning algorithms to improve the terrain strength characterization. The proposed work as part of this STTR will advance the previous efforts to identify the components required to develop remote sensing capabilities for offline and online mobility planning, develop the relationship between hyperspectral, multispectral, thermal, and optical sensing, and terrain mechanical properties such as cohesion, friction, cone index, and vehicle cone index, conduct initial integration of the components to demonstrate feasibility online and offline terrain strength estimation using remote sensing for ground vehicle mobility planning, and complete a proof-of-concept field demonstration that supports a short (<0.5km) vehicle traversal with nearly real-time terrain information. The work will explore strategies for both online and offline mobility planning. The combined team (technical and industry partners) have extensive expertise in remote sensing terrain properties, Unmanned Aerial Vehicle (UAV) based large data collection, multi-sensor approaches, and mobility planning. The principal investigator is part of a provisional patent application on the use of hyperspectral sensing for soil gradation determination.
