SBIR-STTR Award

Imagery analysts must frequently identify targets in rapidly changing environments where mistakes can have tragic consequences. LADAR is an emerging sensor technology that provides a rich data source for human analysts in support of combat identification. However, the primary focus of prior research has been on developing LADAR as a technology without adequately considering how an analyst will exploit LADAR data. Additionally, providing 3D LADAR data to an analyst when they are already trying to manage ever-increasing amounts of data in multiple forms can promote data overload. Further, issues of trust in automation must be examined. We will leverage innovative research methodologies and cognitive task analyses to identify the cognitive demands and challenges of combat identification analysts and determine how 3D LADAR data can address these gaps. We will develop visualization concepts based on ecological interface design principles to support 3D data exploitation. This effort will utilize a decision-centered design approach to develop a truly collaborative system that will allow a human analyst and assisted target recognition systems to work together to exploit 3D LADAR data. These concepts will generalize to other domains where the exploitation of 3D data is critical including medical imagery, border surveillance, and airport screening.

Benefit:
Within the military, this research will have direct applicability to military intelligence and any analysts who are responsible for exploiting 3D LADAR data for combat identification. The proposed system will provide innovative visualization concepts that will allow analysts to interpret and act on 3D data. This ability has direct relevance to any government or private sector group that uses 3D data, including DHS (border surveillance and airport screening) and the medical community (diagnosis using 3D imaging).

Keywords:
Ladar, Combat Identification, Cognitive Task Analysis, Imagery Analysis, 3d Data Visualization, 3d Data Exploitation, Assisted Target Recognition

Current LADAR analysis support tools and processes are inherently flawed due to the data-driven approaches that have guided the development of interfaces, automations, and other supports. Analyst cognitive demands, requirements, and preferences have had little influence in the design of these tools. As a result, these so called “decision aids” can often inhibit the decision making process. In Phase II of this effort, we propose to continue our Phase I Cognitive Systems Engineering approach to develop an analyst-centric LADAR analysis support tool, leveraging the nascent Opticks LADAR add-on framework as a starting point. We will extend the cognitive task analysis and simulation interview protocols developed and executed in Phase 1 to fully describe and detail the decision points, cues and strategies, performance barriers, and current/future benefits associated with LADAR data analysis, leveraging a range of operational LADAR PED cell analysts and likely future NASIC users. Our end product will promote efficient and effective analysis by: leveraging automation only where the analyst will need and accept it (e.g., improved line of sight analysis, automated colorization range enhancements, product support development); developing cognitively-based user interfaces; and implementing the Visual Checklist tool, which will describe the critical decision points in the analysis process.

Benefit:
Analyst-centric LADAR PED decision support tools will greatly enhance the ability of analysts to provide more effective and customized products to their customers. Military benefactors of such tools (and the research we will conduct to derive them) include operational PED cells in Afghanistan and elsewhere as well as NASIC analysts and the Air Force LADAR R&D community. Multiple entities within the Department of Homeland Security would also benefit greatly from cognitively-based LADAR analysis support, whether they are managing borders or disasters.

Keywords:
Ladar, Cognitive Systems Engineering, Analyst-Centric, Geoint Analysis, Lidar, 3d Data Visualization, Combat Identification, Simulation Interview ---------- The demand for 3D LADAR imaging is growing both in the DoD and commercial sectors.However, exploitation processes and tools have lagged advances in LADAR technologies.In a prior Phase II SBIR effort, 361 Interactive leveraged a Cognitive Systems Engineering approach to develop a LADAR analysis tool that was implemented as a Quick Terrain Modeler plug-in.The tool consisted of automated feature extraction and machine learning-based object classification algorithms along with a novel user interface for interacting with the algorithms and their output.The current effort will build upon that prototype product to expand and enhance both the algorithm and UI capabilities.Our overall objective for this Phase II follow-on effort is to develop a mission-driven, analyst-centric LADAR analysis capability that fosters human-machine teaming to improve the accuracy, efficiency and effectiveness of the LADAR exploitation workflow.We will employ a user-centric approach involving frequent and substantive interaction with end users, and incorporation of state-of-the-art machine learning algorithms to segment and classify LADAR data sets.The final product will seamlessly integrate into the analysts workflow, provide much better accuracy, and significantly expand the range of classes and subclasses that can be detected.