Jerome Azarewicz, Glenn Fala, Ralph Fink, and Christof Heithecker. Plan recognition for airborne tactical decision making. In Proceedings of the Fifth National Conference on Artificial Intelligence, pages 805--811, Philadelphia, 1986.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....must coordinate their activity instead by observing the actions of other agents and recognizing their plans. Both systems use this plan recognition to allow agents to coordinate without explicit communication. Plan recognition has also been used by the military for tactical decision making [AFFH86, AFH89] Enemy plans are recognized from observations of individual enemy ship and airplane activity. 2.3 Natural Language One of the largest areas of plan recognition research has been in the field of natural language processing. Human communication is filled with intended recognition, which ....

Jerome Azarewicz, Glenn Fala, Ralph Fink, and Christof Heithecker. Plan recognition for airborne tactical decision making. In Proceedings of the Fifth National Conference on Artificial Intelligence, pages 805--811, Philadelphia, 1986.

.... language, or inexpressivity of the common communication language; creating a need for an agent tracking capability for effective collaboration[14] This agent tracking capability is closely related to plan recognition, which involves recognizing agents plans based on observations of their actions[15, 2, 23]. One key difference is that plan recognition efforts typically focus on tracking a narrower (planbased) class of agent behaviors, as seen in static, single agent domains. In particular, they assume that agents rigidly follow plans step by step. Agent tracking, in contrast, can involve tracking a ....

J. Azarewicz, G. Fala, R. Fink, and C. Heithecker. Plan recognition for airborne tactical decision making. In Proceedings of the National Conferenceon Artificial Intelligence, pages 805--811. Menlo Park, Calif.: AAAI press, 1986.

....Thus, the pilot agents need to continually track their opponents actions, such as turns, and infer unobserved actions, high level goals and behaviors, such as the fpole, beam or missile firing behaviors. This agent tracking capability is related to plan recognition [ Kautz and Allen, 1986; Azarewicz et al. 1986 ] The key difference is that plan recognition efforts typically focus on tracking a narrower (plan based) class of agent behaviors, as seen in static, single agent domains. In particular, they assume that agents rigidly follow plans step by step. In contrast, agent tracking involves the novel ....

J. Azarewicz, G. Fala, R. Fink, and C. Heithecker. Plan recognition for airborne tactical decision making. In Proceedings of the National Conference on Artificial Intelligence, pages 805--811. Menlo Park, Calif.: AAAI press, 1986.

....must coordinate their activity instead by observing the actions of other agents and recognizing their plans. Both systems use this plan recognition to allow agents to coordinate without explicit communication. Plan recognition has also been used by the military for tactical decision making [AFFH86, AFH89] Enemy plans are recognized from observations of individual enemy ship and airplane activity. 2.3 Natural Language One of the largest areas of plan recognition research has been in the eld of natural language processing. Human communication is lled with intended recognition, which ....

Jerome Azarewicz, Glenn Fala, Ralph Fink, and Christof Heithecker. Plan recognition for airborne tactical decision making. In Proceedings of the Fifth National Conference on Arti cial Intelligence, pages 805-811, Philadelphia, 1986.

....or generating actions designed to influence the behavior itself. Researchers in AI have studied this problem of plan recognition for several kinds of tasks, including discourse analysis [18] user modeling [23] traffic monitoring [37] collaborative planning [20] and adversarial planning [1]. Existing plan recognition approaches illuminate several key issues in the knowledge representation and inference subtasks of plan recognition. Modeling the uncertainty inherent in most planning domains provides one of the most difficult challenges to a recognizer. Many approaches use ....

....subsequence lengths which add to the total length of four. If np derives a string of length one and vp a string of length three, then the only possible levels of abstraction for each are three and one, respectively, since all others will have zero fi values. Therefore, we insert the production s np[1,3] vp[3,1] into the domain of P 141 , where the numbers in brackets correspond to the subsequence length and level of abstraction, respectively. The conditional probability of this value, given that N 141 = S, is the product of the probability of the production, fi(np; 1; 3) and fi(vp; 3; 1) ....

[Article contains additional citation context not shown here]

Jerome Azarewicz, Glenn Fala, Ralph Fink, and Christof Heithecker. Plan recognition for airborne tactical decision making. In Proceedings of the Fifth National Conference on Artificial Intelligence, pages 805--811, 1986.

.... to a speaker s utterances) must recognize the plan that the speaker intends to convey; this is referred to as intended recognition[CSSvB91] The latter situation is representative of adversarial settings, such as warfare, where an agent might actively attempt to thwart recognition of his plan[Pol87, AFFH86]. However, little published research has examined such interactions. Although early work on plan recognition offered much promise and the contribution of plan recognition to robust adaptive systems has been widely recognized, a number of serious problems have hampered the use of plan recognition ....

....and novel plans and for revising system beliefs to account for misconceptions or novel plans that originally went undetected. In addition to noise in the input, robustness is affected if the agent is deliberately attempting to thwart the plan inference process. Although Azarewicz et. al[AFFH86, AFH89] investigate plan recognition in an adversarial domain, they expand their system s knowledge base to encode plans that they expect an adversary might pursue and do not propose principled mechanisms for hypothesizing how an agent might attempt to conceal his actual plan with misleading actions. 6 ....

Jerome Azarewicz, Glenn Fala, Ralph Fink, and Christof Heithecker. Plan Recognition for Airborne Tactical Decision Making. In Proceedings of the Fifth National Conference on Artificial Intelligence, pages 805--811, Philadelphia, Pennsylvania, 1986.

.... language, or inexpressivity of the common communication language; creating a need for an agent tracking capability for effective collaboration[14] This agent tracking capability is closely related to plan recognition, which involves recognizing agents plans based on observations of their actions[15, 2, 23]. One key difference is that plan recognition efforts typically focus on tracking a narrower (planbased) class of agent behaviors, as seen in static, single agent domains. In particular, they assume that agents rigidly follow plans step by step. Agent tracking, in contrast, can involve tracking a ....

J. Azarewicz, G. Fala, R. Fink, and C. Heithecker. Plan recognition for airborne tactical decision making. In Proceedings of the National Conferenceon Artificial Intelligence, pages 805--811. Menlo Park, Calif.: AAAI press, 1986.

....Thus, the pilot agents need to continually track their opponents actions, such as turns, and infer unobserved actions, high level goals and behaviors, such as the fpole, beam or missile firing behaviors. This agent tracking capability is related to plan recognition [ Kautz and Allen, 1986; Azarewicz et al. 1986 ] The key difference is that plan recognition efforts typically focus on tracking a narrower (plan based) class of agent behaviors, as seen in static, single agent domains. In particular, they assume that agents rigidly follow plans step by step. In contrast, agent tracking involves the novel ....

J. Azarewicz, G. Fala, R. Fink, and C. Heithecker. Plan recognition for airborne tactical decision making. In Proceedings of the National Conference on Artificial Intelligence, pages 805--811. Menlo Park, Calif.: AAAI press, 1986.

....goals. Agent tracking is an important requirement for intelligent interaction. It involves monitoring other agents observable actions as well as inferring their unobserved actions or high level goals, plans and behaviors. Agent tracking is closely related to plan recognition (Kautz Allen 1986; Azarewicz et al. 1986), which involves recognizing agents plans based on observations of their actions. The key difference is that 1 I thank Paul Rosenbloom and Ben Smith for detailed feedback on this effort. Thanks also to Lewis Johnson, Piotr Gmytrasiewicz and the anonymous reviewers for helpful comments. This ....

Azarewicz, J.; Fala, G.; Fink, R.; and Heithecker, C. 1986. Plan recognition for airborne tactical decision making. In Proceedings of the National Conference on Artificial Intelligence, 805--811. Menlo Park, Calif.: AAAI press.

....be restricted due to cost, risk, lack of a common protocol etc. 13] The key to this capability is tracking an agent s flexible mix of goal driven and reactive behaviors, as seen in dynamic, interactive, multi agent domains. This contrasts with previous work in the related area of plan recognition[15, 2], which mostly focuses on recognizing plans in static, single agent domains. This paper takes a step beyond tracking individual agents the current state of the art in agent tracking and plan recognition by focusing on team tracking. We (humans) see team activity all around, e.g. teamwork ....

J. Azarewicz, G. Fala, R. Fink, and C. Heithecker. Plan recognition for airborne tactical decision making. In Proceedings of the National Conference on Artificial Intelligence, pages 805--811. Menlo Park, Calif.: AAAI press, 1986.

....situation awareness, may be adequate for comparing the relative effectiveness of different tactics against a given enemy tactic, but does not provide an adequate representation of tactical decision making by the pilots and does not take account of pre mission briefing of pilots. Azarewicz et: al: [ 3 ] have developed a system which is based on recognizing the an opponent s plans from a given set of possible plans, based on observing manoeuvres and actions. In this paper we have extended this approach and developed a formalism for recognising beliefs, desires and intentions of other agents in ....

J. Azarewicz, G. Fala, R. Fink, and C. Heitheker. Plan recognition for airborne tactical decision making. In Proceedings of the Fifth National Conference on Artificial Intelligence (AAAI-86), pages 805--811, 1986.

.... Introduction Plan recognition and agent modeling capabilities are valuable for intelligent tutoring (Corbett et al. 1990; Johnson, 1986) as well as other areas such as natural language processing (Charniak Goldman, 1991) expert consultation (Calistri, 1990) and tactical decision making (Azarewicz et al. 1986). However, such capabilities are difficult to implement and employ effectively, for the following reasons. Plan recognition techniques can be rigid they assume the agent is following a known plan step by step, and have difficulty interpreting deviations from the plan. The modeling process can be ....

Azarewicz, J., Fala, G. and Fink, R. and Heighecker, C. (1986). Plan recognition for airborne tactical decision making. Proceedings of the Fifth National Conference on Artificial Intelligence , pp. 805-811, 1986.

.... particular, in previous investigations in the related areas of plan situation recognition (Kautz Allen 1986; Song Cohen 1991; Dousson, Gaborit, Ghallab 1993; Van Beek Cohen 1991; Carberry 1990a) including one investigation focused on plan recognition in airborne tactical decision making(Azarewicz et al. 1986) these issues have not been addressed. With regard to the first two issues, plan recognition models have not been applied in such dynamic, interactive multi agent situations, and hence do not address strong interactions among agents or the resulting flexibility and reactivity in agent ....

Azarewicz, J., G. Fala, R. Fink and C. Heithecker 1986. Plan recognition for airborne tactical decision making. In Proceedings of the National Conference on Artificial Intelligence, 805--811. Menlo Park, Calif.: AAAI press.

....required for intelligent interaction (Tambe Rosenbloom 1995; Ward 1991; Rao 1994) It involves monitoring other agents observable actions and inferring their unobserved actions or high level goals, plans and behaviors. This capability is closely related to plan recognition (Kautz Allen 1986; Azarewicz et al. 1986), which involves recognizing agents plans based on observations of their actions. One key difference is that plan recognition efforts generally assume that agents are executing plans that rigidly prescribe the actions to be performed. Agent tracking, in contrast, involves recognizing a broader ....

Azarewicz, J.; Fala, G.; Fink, R.; and Heithecker, C. 1986. Plan recognition for airborne tactical decision making. In Proceedings of the National Conference on Artificial Intelligence, 805--811. Menlo Park, Calif.: AAAI press.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC