Airborne radar signal environment
Airborne radar faces severe ground and sea clutter. Because clutter is often strong and widely distributed, and because the platform moves, the radar clutter spectrum is broadened. Therefore, clutter suppression is a key technical challenge for airborne radar. Traditional clutter suppression methods that operate only in the time or frequency domain are often insufficient, making detection of small targets difficult. New theories and techniques are required to address airborne radar clutter suppression.
Signal distribution of airborne radar
Ground clutter for airborne radar exhibits a two-dimensional space-time coupled spectrum. This implies that clutter suppression fundamentally becomes a two-dimensional space-time filtering problem, and the two-dimensional processing must be implemented adaptively and in real time. That is, clutter suppression requires space-time adaptive processing (STAP). Adaptive processing can effectively match complex external environments and partially compensate system errors, significantly improving system performance. However, adaptive processing typically requires flexible formation of multiple beams and online real-time computation of adaptive weights, which involves substantial computational load. The rapid development of digital beamforming (DBF) and very-large-scale integration (VLSI) circuits has enabled precise control of space-time adaptive weights and faster processing, creating favorable conditions for practical application of two-dimensional space-time adaptive processing.
Comparison: spatial, temporal, and space-time 2D filtering
Development history
Space-time adaptive processing is an evolution of spatial adaptive techniques. In the late 1960s, Van Atta and others proposed spatial adaptive ideas to suppress spatial clutter and interference. In 1973, Brennan first introduced the concept of two-dimensional space-time adaptive processing, extending array adaptive processing from element signals to a two-dimensional data field consisting of pulses and array element samples.
Because STAP has an extremely large computational burden, real-time processing was nearly impossible for many years, and research stalled. In the early 1980s, Dr. Klemm proposed auxiliary channel reduction (ACR) to reduce dimensionality, making engineering implementation feasible and triggering renewed research activity. Academician Bao Zheng was a pioneer in STAP research in China, followed by researchers such as Yongliang Wang and Guisheng Liao.
Current STAP algorithms can be classified into four main categories.
One of the most cited reports on STAP is J. Ward's "Space-Time Adaptive Processing for Airborne Radar," which has been cited extensively.
Basic principles of STAP
An antenna array processes two-dimensional spatial and temporal data over a coherent processing interval (CPI) to detect and locate target signals in noise and clutter environments. Spatial data come from the array elements, and temporal data come from a sequence of pulses within the coherent processing time. N antenna elements, M pulses, and L range bins form a three-dimensional data cube. This cube contains all information about targets, clutter, interference, and noise. Data received from each range gate form a snapshot, and snapshots across range gates accumulate into the data cube. The data cube is used to determine whether a target exists in each range gate, while adaptive two-dimensional beamforming maximizes the signal-to-noise ratio for each range gate.
Overview of the Mountain Top program and experimental systems
Major STAP experimental systems include the Mountain Top program, the MCARM program, MIT Lincoln Laboratory's KASSPER project, which integrates various prior knowledge into traditional STAP algorithms, and parallel STAP implementations at the University of Southern California on large-scale parallel machines such as IBM SP2, Cray T3D, and Intel Paragon. In addition, Chinese institutions such as Xidian University, Tsinghua University, the National University of Defense Technology, and the Chinese Academy of Sciences have also conducted significant experimental work on STAP.
In the 1990s, the U.S. Air Force and the Advanced Research Projects Agency (ARPA) launched the Mountain Top program to investigate advanced signal processing and related technologies for next-generation airborne early warning radar. The program installed experimental radars on several mountain tops at test ranges in New Mexico and Hawaii, and a dataset was collected in those ranges around 1993.
