Separating Cognition Performance from Execution Environment (SCoPE)
Abstract
The ability to measure and, more importantly, predict required capabilities and effectiveness by units to execute assigned missions and tasks is paramount to prepare our military forces for mission success. Military research is still trying to understand the relationships between preparation for combat and unit performance effectiveness [1]; this challenge applies to all levels (strategic, operational, and tactical) and phases (planning, execution) of warfighting mission readiness. Among the many aspects to be considered in such research is the definition of measures of performance, the required input data and the feasibility to obtain such information in time, so Naval commanders can make informed decisions while defining a particular military strategy. Human performance models have been able to integrate user characteristics with environmental characteristics to more accurately represent vulnerabilities brought to a system’s performance by the human interaction [2]. A key feature in such models is precisely the ability to represent, independently, both spaces: user and environmental conditions. Military unit performance models should consider a similar approach. In the ‘fog of war’, timely individual soldier performance measures might be difficult to obtain so prediction of mission readiness will be defined mainly from conditions in the area of operation and the commander’s intuition about the expected unit performance. Defining performance models that take into account this potential lack of information, will provide our warfighters with better representation and assessment of the readiness to engage an enemy. Moreover, these models can be adapted for predictive analytics on the readiness of the enemy, providing our troops military advantage. Our overall objective in Separating Cognition Performance from Execution Environment (SCoPE2) is to enhance Naval commander s decision aids (e.g., mission readiness, predictive analytics) by integrating models that estimate the ability of units to accomplish tasks to expected standard under specified conditions, isolating human performance factors. SCoPE2 will develop a framework for measuring unit performance during exercises that allows the separation of human performance differences from physical, military and civil conditions. Developed models will weight features based on relative importance. The proposed approach in SCoPE2 will leverage our research and development efforts on the Model Integrity and Discovery Suite (MInDS) program (ONR Contract N00173-07-C-4018). MInDS was developed as an analytically rigorous meta-fusion process to prioritize the features and events a decision maker should consider while evaluating a situation. As is described below, we will mature our research on this area and define a framework to capture relevant features to describe the individual soldier, units, tasks and external conditions during military training exercises. These templates will allow the implementation of a flexible model to evaluate unit performance on multiple missions and corresponding mission tasks. A novel machine learning technology, conceptual spaces, will be used as the basis for our proposed framework. Conceptual spaces [3] introduces a geometrical way to represent human thought. The theory is relevant not only to cognition, but appears to be robust enough to handle other types of problems. The use of conceptual spaces will provide a framework for representing the various mission types and the requirements for a successful mission as well as the ability to assess a plan’s distance from feasibility as well as from optimality.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 08, 2016
- Source ID
- N000141512376
Entities
People
- Moises Sudit
Organizations
- Calspan-University of Buffalo Research Center
- Office of Naval Research
- United States Navy