In the down-looking operating mode of space-based multichannel radar, the pronounced broadening of mainlobe clutter severely degrades the target detection performance. To address this issue, a reduced-dimension clutter rejection approach based on beam-subpulse smoothed coherent integration is proposed. The proposed method first introduces a spatial reduced-dimension matrix to transform two-dimensional (2D) space-time data into the beam-pulse domain. Next, multiple sets of beam-subpulse 2D data are obtained using subpulse smoothing technique. Subsequently, the clutter-plus-noise covariance matrix is estimated from one or multiple sets of beam-subpulse data and shared across different subaperture data groups to construct the corresponding clutter suppression weight vectors. Finally, clutter is suppressed in each set of beam subpulse data using its respective weight vector, and the resulting residuals from all sets are coherently summed to form the final output. The reliability and effectiveness of the proposed approach are verified by both simulation and experimental data.
This research focuses on the time-varying formation control of multi-UAV systems, specifically addressing their resilience against sensor faults and denial-of-service (DoS) attacks. A security-aware control architecture is proposed for UAV formations by combining a distributed resilient controller with an RTS-CKF. Building on this framework, a distributed resilient control law is developed to ensure formation maintenance by using the neighbor information available during DoS attacks. The RTS-CKF estimates UAV states and sensor faults through a decoupled procedure and supplies corrected state information for feedback. Finally, comprehensive simulations of nonlinear UAV formations under concurrent sensor faults and DoS attacks demonstrate the robustness and effectiveness of the proposed scheme.
Owing to fast platform motion, the mainlobe clutter spectrum of airborne multichannel radar is severely broadened, causing slow-moving targets to be buried in the strong clutter background. To tackle this problem, we propose a reduced-dimension space-time adaptive processing (STAP) approach based on space-time subaperture synthesis. In this approach, the raw spatio-temporal two-dimensional (2D) echo data is first processed using a spatial sliding window to generate multiple sets of spatial subaperture data. These subaperture data are then synthesized spatially to form subspace-pulse 2D data, achieving spatial dimensionality reduction. Subsequently, time-domain sliding window processing and vectorization are applied to the synthesized subspace-pulse data to obtain multiple sets of time-domain subaperture data. One or more of these sets are selected to construct the clutter-plus-noise covariance matrix (CNCM). Following this, temporal synthesis is performed on all time-domain subapertures to obtain the final space-time subaperture data, realizing temporal dimensionality reduction. Finally, adaptive clutter suppression is performed on the synthesized spatio-temporal subaperture data to enhance target detection performance in strong clutter environments. The efficacy and practicality of the present approach are validated through simulation and experimental data.
In highly dynamic and confrontational environments, the motion of aerial targets exhibits distinct spatiotemporal coupling characteristics and intricate maneuver patterns. The strategic behaviors of aerial targets are influenced by their motion states and dynamic interactions between opposing sides, which makes traditional trajectory prediction approaches insufficient for supporting subsequent maneuver strategy inference. To address this issue, this paper proposes an aerial target maneuver strategy inference framework integrating the Spatiotemporal Attention Transformer (STA-Transformer) and matrix game theory. First, a trajectory behavior recognition model based on STA-Transformer is established. Maneuver-aware positional encoding and a dual-pathway spatiotemporal feature separation mechanism are introduced to separately model temporal evolution patterns and correlations among multi-dimensional state variables, thereby enhancing the representation capability of spatiotemporal maneuver patterns under complex motion conditions. Second, based on predicted trajectory states, a dynamic threat payoff function is constructed by incorporating target motion characteristics, radar operating conditions, and tactical interactions. A matrix game model between aerial targets and ground radars is further formulated. Linear programming is adopted to solve the mixed-strategy Nash equilibrium, enabling the transformation from trajectory behavior perception to maneuver strategy inference. Simulation results demonstrate that under the test scenarios and experimental settings established in this work, the proposed method achieves accuracy of 95.3% and 96.4% for trajectory pattern recognition and maneuver strategy inference, respectively, outperforming comparative approaches. The findings verify the effectiveness of the proposed framework in aerial target behavior modeling and strategy inference, and provide a feasible scheme for integrating data-driven behavior modeling and game-based strategy inference in complex air defense scenarios.
This work presents a method for leveraging intercepted Radio Frequency (RF) signals to inform a constrained Reinforcement Learning (RL) policy for robust control of a satellite operating in contested environments. Noncooperative satellite engagements with nation-state actors prompts the need for enhanced maneuverability and agility on-orbit. However, robust, autonomous and rapid adversary avoidance capabilities for the space environment is seldom studied. Further, the capability constrained nature of many space vehicles does not afford robust space situational awareness capabilities that can be used for well informed maneuvering. We present a “Cat & Mouse” system for training optimal adversary avoidance algorithms using RL. We propose utilizing intercepted radio frequency communication and dynamic spacecraft state as multi-modal input that could inform paths for a mouse to out maneuver the cat satellite. Given the current ubiquitous use of RF communications, our proposed system can be applicable to a diverse array of satellites. In addition to providing a comprehensive framework for training and implementing a constrained RL policy capable of providing control for robust adversary avoidance, we compare our method to several optimization based methods for adversarial avoidance. These methods were tested on real-world data obtained from the Space Surveillance Network (SSN) to analyze the benefits and limitations of different avoidance methods.
