Abstract:The Autonomous Air Combat technique has been a lasting research topic for decades. However, no complete solutions seem to have appeared because of the highly dynamic and complex nature of the Autonomous Air Combat problem. In devising the Autonomous Air Combat solutions, we follow similar methodologies in the robotics community, and divide the overall scheme into two folds: the perception of other (enemy/friendly) aircraft, and the guidance/control for own aircraft. While the perception in the first fold serves as a foundation, this paper is mainly focused on the second one. Based on our survey, a review of own aircraft guidance/control in the (primarily one-to-one) Autonomous Air Combat solutions is presented. We divide different Autonomous Air Combat solutions into three groups, i.e. mathematics-based, knowledge-encoded, and learning-driven. In each group, we present the representative methods first; problem definition, solution, and a brief overview of the historical development are illustrated. We also comment on both weakness and strengths for each group/method. We point out certain technical paths/challenges that need to be addressed in the future Autonomous Air Combat development, i.e. to abstract and emulate the human pilot experiences, and to develop the online learning capabilites. Inspired by the state-of-art techniques in other similar fields (robotics, autonomous driving), we also propose potential solutions, i.e. traditional approaches enhanced by the novel data-driven technique. Via this paper, we hope to deliver an in-depth analysis of past experiences and potential challenges/solutions for the Autonomous Air Combat technique. We also advocate referring to the approaches/techniques that are utilized in other similar fields in devising the Autonomous Air Combat solutions.

Research on enemy aircraft bank angle estimation for autonomous air combat

Fast Detection and Recognition Method of UAV in Sky Background

Visual Perception-Based Target Aircraft Movement Prediction for Autonomous Air Combat

Knowledge-Driven Accurate Opponent Trajectory Prediction for Gun-Dominated Autonomous Air Combat

Guidance and Control for Own Aircraft in the Autonomous Air Combat: A Historical Review and Future Prospects

Research on recognizing aiming-point of airplane's infrared sequence images at endgame

Orientation Estimation for Airplane Targets in SAR Images Based on Keypoint Detection

Towards Micro Air Vehicle Flight Autonomy Research on the Method of Horizon Extraction

The Research of Air Combat Intention Identification Method Based on BiLSTM + Attention

A Neural Network Based Inverse System Method for High Angle-of-attack Control

A GRU-Based Method for Predicting Intention of Aerial Targets

An Efficient Target Detection and Recognition Method in Aerial Remote-sensing Images Based on Multiangle Regions-of-Interest

State estimation of micro autonomous helicopter based on computer vision

RAIH-Det: An End-to-End Rotated Aircraft and Aircraft Head Detector Based on ConvNeXt and Cyclical Focal Loss in Optical Remote Sensing Images

Estimation of drone intention using trajectory frequency defined in radar's measurement phase planes

Maneuvering Strategy and Own Aircraft Movement Prediction in Trial Input Method - Low Angle of Attack.

Vision-Based 3D Aerial Target Detection and Tracking for Maneuver Decision in Close-Range Air Combat

Vision-based Horizon Extraction for Micro Air Vehicle Flight Control

Autonomous UAV maneuvering decisions by refining opponent strategies

Axis Estimation of Spaceborne Targets via Inverse Synthetic Aperture Radar Image Sequence Based on Regression Network

Precise Interception Flight Targets by Image-based Visual Servoing of Multicopter