A23-013 - Real-Time Analytics for Engineer Reconnaissance - KIHOMAC

Visual Simultaneous Location and Mapping (VSLAM) is a robotics and artificial intelligence supporting technology with many potential benefits to the Engineer Reconnaissance function. However, the technology has several key constraints due to camera stabilization, sensor quality, optical quality, and processing speeds of the attached computer. Environment factors, such as noisy, movement of objects (swaying branches in wind) and poor illumination are significant problems currently being encountered by the research community at large. These environments put significant constraints on the technological success in unstructured, complex, and large-scale scans of the typical Army Area of Operation (AO). Additionally, the balance between real-time and accuracy of the various methods is a significant trade-off yet to be understood in the Engineer Reconnaissance aspect.KIHOMACÆs Crossing Support System, known as CROSSFIRE, is posed to adapt, advance, and incorporate VSLAM technology into the ENFIRE program. CROSSFIRE is currently integrated in the ENFIRE software application and offers an enhanced bridge Military Load Classifications (MLC) calculation process beyond the capabilities of traditional methodology. The calculation tools include vector, Computer Vision (CV) and lidar functionality. While currently tailored to the MLC function, these technologies are engineering task agnostic. CROSSFIRE fully supports the current engineer doctrine of assured mobility and provides critical input to the commandersÆ front-end mission analysis (MA) and intelligence preparation of the battlefield (IPB) requirements to support the Military or Rapid Decision-Making Process (MDMP/RDMP).A decomposition of the research topic, Real-Time Analytics for Engineer Reconnaissance, into Capability Need Statements (CNS) from a software solution that answers intelligence requirements for the Army Engineers and decision-makers and informs command decisions through:Transformation of ground-based full-motion video (FMV), and other sensors within the ENFIRE kit, of a mobility corridor into a digital 3D scan (sometimes called a 3D point cloud)Manipulation, visualization, and automation of obstacles analysis within the digital 3D scan, i.e. Semantic VSLAMKIHOMAC proposes a research and development effort to apply VSLAM technology within a mobility assessment system, fully leveraging critical code, libraries and tools provided by CROSSFIRE and already incorporated within ENFIRE.Ā The research and development efforts would accomplish the feasibility analysis, transfer, and convergence of these technologies into the currently fielded ENFIRE kit.