Accession Number:

AD1093491

Title:

Reconciling Simultaneous Evolution of Ground Vehicle Capabilities and Operator Preferences

Descriptive Note:

Technical Report,01 Jan 2019,01 Nov 2019

Corporate Author:

US ARMY CCDC GVSC WARREN United States

Personal Author(s):

• Slon,Christopher
• Pandey,Vijitashwa
• Gorsich,David
• Jayakumar,Paramsothy

Report Date:

2019-11-08

Pagination or Media Count:

10.0

Abstract:

An objective evaluation of ground vehicle performance is a challenging task. This is further exacerbated by the increasing level of autonomy, dynamically changing the roles and capabilities of these vehicles. In the context of decision making involving these vehicles, as the capabilities of the vehicles improve, there is a concurrent change in the preferences of the decision makers operating the vehicles that must be accounted for. Decision based methods are a natural choice when multiple conflicting attributes are present, however, most of the literature focuses on static preferences. In this paper, we provide a sequential Bayesian framework to accommodate time varying preferences. The utility function is considered a stochastic function with the shape parameters themselves being random variables. In the proposed approach, initially the shape parameters model either uncertain preferences or variation in the preferences because of the presence of multiple decision makers. We consider this utility distribution as the prior and update it to a posterior with feedback that can be acquired from actual system use. The framework improves the utility function and thereby the decisions made for the next generation systems, allowing continuous improvement. We present our approach on a ground vehicle selection problem.

Descriptors:

• ground vehicles
• bayesian networks
• random variables
• test and evaluation
• probabilistic models

Subject Categories:

• Surface Transportation and Equipment
• Statistics and Probability

Distribution Statement:

APPROVED FOR PUBLIC RELEASE