Event Tracking in Complex Multi-agent Environments

نویسندگان

  • Milind Tambe
  • Paul S. Rosenbloom
چکیده

exercises, the automated pilots must act in a Abstract realistic manner, i.e., like trained human pilots. Otherwise, both the training and tactics The Soar-IFOR project is aimed at developing development in these environments will not be intelligent automated pilots for simulated tactical realistic. air-combat. One key requirement for an automated To act in a realistic manner, an automated pilot pilot in this environment is event tracking: the must, among other things, be responsive to events ability to monitor or track events instigated by in its environment  it must modify and adapt its opponents, so as to respond to them appropriately. own maneuvers in response to relevant events. These events include the opponents’ low level These events may correspond to simple actions of actions, which the automated pilot may directly other pilots, such as changes in heading or altitude, observe, as well as opponents’ high level plans and which the automated pilot may directly observe on actions, which the automated pilot can not observe its radar. Alternatively, these events may involve (but only infer). This paper analyzes the challenges the execution of complex, high-level actions or that an automated pilots must face when tracking plans of other pilots, which the automated pilot can events in this environment. This analysis reveals not directly observe. For instance, one crucial some novel constraints on event tracking that arise event is an opponent’s firing a missile at an from the dynamic multi-agent interactions in this automated pilot’s aircraft, threatening its very environment. In previous work on event tracking, survival. Yet, the automated pilot cannot directly which is primarily based on single-agent see the missile until it is too late to evade it. environments, these constraints have not been Fortunately, the automated pilot can monitor the addressed. This paper proposes one solution for opponent’s sequence of maneuvers, and infer the event tracking that appears better suited for possibility of a missile firing based on them, as addressing these constraints. The solution is shown in Figure 1. The automated pilot is in the demonstrated via a simple re-implementation of an dark-shaded aircraft, and its opponent is in the existing automated pilot agent for air-combat light-shaded one. 1 simulation .

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Event Tracking for an Intelligent Automated Agent

In a dynamic, multi-agent environment, an the intelligent agents act as automated pilots automated intelligent agent is often faced with the for the simulated aircraft. These automated possibility that other agents may instigate events that pilots will take part in exercises with human actually hinder or help the achievement of its own goals. fighter pilots, where they will aid in tactics To ac...

متن کامل

Adaptive Neural Network Method for Consensus Tracking of High-Order Mimo Nonlinear Multi-Agent Systems

This paper is concerned with the consensus tracking problem of high order MIMO nonlinear multi-agent systems. The agents must follow a leader node in presence of unknown dynamics and uncertain external disturbances. The communication network topology of agents is assumed to be a fixed undirected graph. A distributed adaptive control method is proposed to solve the consensus problem utilizing re...

متن کامل

Event Tracking in a Dynamic Multiagent Environment

In a dynamic, multi-agent environment, an automated intelligent agent is often faced with the possibility that other agents may instigate events that hinder or help the achievement of its own goals. To act intelligently in such an environment, an automated agent needs an event tracking capability to continually monitor the occurrence of such events and the temporal relationships among them. Thi...

متن کامل

Roadmap-Based Pursuit-Evasion in 3D Structures

We present an approach to the pursuit-evasion problem which is applicable to complex, multi-level environments. Studying each aspect of this problem in 3D structured environments is a distinct extension over many previous approaches. We also utilize our roadmap-based approach to multi-agent behavior when tracking agents of interest. Results are presented in three multi-level environments to hig...

متن کامل

Adaptive neural control of nonlinear fractional order multi- agent systems in the presence of error constraintion

In this paper, the problem of fractional order multi-agent tracking control problem is considered. External disturbances, uncertainties, error constraints, transient response suitability and desirable response tracking problems are the challenges in this study. Because of these problems and challenges, an adaptive control and neural estimator approaches are used in this study. In the first part...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1994