Integrating adjustable autonomy in an intelligent control framework

Abstract

Currently, multiple humans are needed to operate a single uninhabited aerial vehicle (UAV). In the near future, combat techniques will involve single operators controlling multiple uninhabited ground and air vehicles. This situation creates both technological hurdles as well as interaction design challenges that must be addressed to support future fighters. In particular, the system will need to negotiate with the operator about proper task delegation, keeping the operator appropriately apprised of autonomous actions. This in turn implies that the system must know what the user is doing, what needs to be done in the present situation, and the comparative strengths for of the human and the system in each task. Towards building such systems, we are working on an Intelligent Control Framework (ICF) that provides a layer of intelligence to support future warfighters in complex task environments. The present paper presents the Adjustable Autonomy Module (AAM) in ICF. The AAM encapsulates some capabilities for user plan recognition, situation reasoning, and authority delegation control. The AAM has the knowledge necessary to support operator-system dialogue about autonomy changes, and it also provides the system with the ability to act on this knowledge. Combined with careful interaction design, planning and plan-execution capabilities, the AAM enables future design and development of effective human-robot teams.

Using GOMS and Bayesian plan recognition to develop recognition models of operator behavior

“…Without a model that specifies, for example, that an operator (a) should attend to threats, (b) normally attends to threats using standard procedures, and (c) should attend to the threat within a specific amount of time, there would be no way for the system to identify a point when system autonomy might be adjusted. Performing this form of task recognition is difficult [6]. Significant categories of relevant information observable by the system include operator mental processes and noncomputer-based communication processes (e.g.…”

Situation Reasoning in an Integrating Adjustable Autonomy Framework

“…The information is composed of a variety of measures including cognitive workload, the user's overall task load, the speed with which tasks need to be accomplished(versus when they need to be completed), all of the METT-TC factors, and other measures. The situation reasoning component is responsible for characterizing the system's current situation based on a combination of task modeling and world modeling [7].…”

“…And situation reasoning contains processing of data such as some sensor processing/fusion or a replan assessment component that monitors for and assesses the impact of contingencies [6]. A frame representative of situation reasoning in adjustable autonomous framework is the Intelligent Control Framework investigated by the company of Soar Technology [7]. The major challenge in situation reasoning is reasoning under uncertainty.…”

Studies on situation reasoning approach of autonomous underwater vehicle under uncertain environment

“…The framework of SR is based upon existing research in framework of SA. The main frames used for SA are data fusion architecture [6], intelligent circulation [7], Joint Directors of Laboratories data fusion model [8], OODA (Observation, Orientation, Decesion, Action) loop [9], SA global assessment technique [1], Dasarathy model [10], Omnibus model [11], The extended OODA model [12], knowledge representation model [13], intelligent control framework [14,15] and model‐based diagnostics [16]. Several knowledge representation approaches can be found in the literature implementing information transfer protocols for robotics [17].…”