Beyond-Visual-Range Air Combat Tactics Auto-Generation by Reinforcement Learning

With the development of information technology, the degree of intelligence in air combat is increasing, and the demand for automated intelligent decision-making systems is becoming more intense. Based on the characteristics of over-the-horizon air combat, this paper constructs a super-horizon air combat training environment, which includes aircraft model modeling, air combat scene design, enemy aircraft strategy design, and reward and punishment signal design. In order to improve the efficiency of the reinforcement learning algorithm for the exploration of strategy space, this paper proposes a heuristic Q-Network method that integrates expert experience, and uses expert experience as a heuristic signal to guide the search process. At the same time, heuristic exploration and random exploration are combined. Aiming at the over-the-horizon air combat maneuver decision problem, the heuristic Q-Network method is adopted to train the neural network model in the over-the-horizon air combat training environment. Through continuous interaction with the environment, self-learning of the air combat maneuver strategy is realized. The efficiency of the heuristic Q-Network method and effectiveness of the air combat maneuver strategy are verified by simulation experiments.

Download Full-text

Development of an Air Combat Performance Measure

Proceedings of the Human Factors Society Annual Meeting ◽

10.1177/154193128803201804 ◽

1988 ◽

Vol 32 (18) ◽

pp. 1207-1211

Author(s):

Gary S. Thomas ◽

David C. Miller

Keyword(s):

Performance Feedback ◽

Diagnostic Performance ◽

Performance Measure ◽

Regression Analyses ◽

Rater Agreement ◽

Air Combat ◽

Measure Of Performance ◽

Development And Validation ◽

Neutral Posture ◽

Visual Range

The purpose of this research was to formulate a unitary measure of performance for simulated one-versus-one, within visual range, air-to-air combat. The measure will serve as a criterion for the development and validation of specific measures of ACM skill that can be used to provide diagnostic performance feedback to pilots. Two experiments were conducted in which fighter pilots served as judges and rank-ordered, from most to least desirable, hypothetical ACM engagement outcomes. Outcome variables included (1) whether or not the hypothetical pilot achieved a “kill,” (2) whether or not he survived the mission, (3) the percent of time the pilot was in an offensive, defensive, or neutral posture, (4) length of engagement, and (5) posture at the beginning and end of the engagement (offensive, defensive, or neutral). In order to determine inter-rater agreement among judges in Experiment I, their rankings were correlated. Correlations ranged from .93 to .99. Pilots' rankings of engagement outcomes were subjected to linear regression analyses to derive equations that could be used as a unitary measure of ACM success. The regression equation in Experiment I accounted for 95% of the variance in rankings, and the composite regression model calculated in Experiment II accounted for more than 70% of the variance.

Download Full-text

Validation of the Explanatory Concept for Decision Support in Air-to-Air Combat

Proceedings of the Human Factors and Ergonomics Society Annual Meeting ◽

10.1177/107118139704100107 ◽

1997 ◽

Vol 41 (1) ◽

pp. 28-31

Author(s):

Jeffrey L. Crick ◽

Stephen J. Selcon ◽

Maddalena Piras ◽

Craig Shanks ◽

Chris Drewery ◽

...

Keyword(s):

Decision Support ◽

Data Fusion ◽

Graphical Representation ◽

Ecological Validity ◽

Graphical Displays ◽

Air Combat ◽

Control Version ◽

Explanatory Concept ◽

Military Pilots ◽

Visual Range

A decision-support aid developed for use by pilots in air-to-air combat was evaluated in a simulated beyond-visual-range combat scenario in which military pilots competed against one another head-to-head. Combat performance was assessed on a range of operationally-valid measures with three different versions of a head-down display showing integrated information derived from data fusion. One version presented graphical, dynamic representations of both ownship's and the enemy's missile performance envelopes (launch success zones); another showed only the launch success zones of the enemy aircraft; and a third, control version showed neither form of graphical representation. Superior attacking performance was demonstrated with the display showing both ownship and enemy launch success zones, while more successful evasive performance was associated with the display showing only enemy launch success zones. Greater levels of situation understanding were associated with the display showing both ownship and enemy launch success zones. The results lend ecological validity to the use of explanatory graphical displays in providing decision support for pilots in air-to-air combat.

Download Full-text