The Engineering Calculus Behind SB-AMTI:

 Eight Tradeoffs That Will Define Whether It Works

May 30, 2026 Companion Analysis — Sensors / Signal Processing

Bottom Line Up Front

The $4.16 billion SpaceX SB-AMTI award commits the U.S. to a high-band radar constellation for airborne moving target indication from LEO — a technically achievable mission, but one that confronts a cascade of interlocking design tradeoffs. The hardest are not political or industrial: they are fundamental physics. Platform velocity approximately 30 times that of an AWACS aircraft explodes the mainlobe clutter Doppler spread, driving STAP complexity and minimum detectable velocity in ways that simply do not exist for airborne radar. The Space Force's dual-band architecture — a high-band production constellation paired with a $140 million low-band wide-area search tier — is a direct engineering response to these constraints. Whether the resulting system can achieve fire-control-quality custody of a maneuvering cruise missile by 2028 is the central unresolved question.

From the range equation's brutal arithmetic to the clutter non-stationarity that defeated every previous space-based radar program, the design choices confronting the SB-AMTI constellation reveal why moving the airborne moving target indicator mission to orbit is genuinely hard — and why the 2028 IOC claim demands scrutiny.

Technical Analysis Desk May 30, 2026 Sensors & Electronic Warfare

▶  Companion to: "Space-Based Eyes for the Kill Chain: SB-AMTI and HBTSS as the Essential Sensors for Hypersonic Defense" — same issue.

WASHINGTON — The announcement of SpaceX's $4.16 billion Space-Based Airborne Moving Target Indicator contract has dominated defense headlines for the straightforward reasons: scale, timing, and the identity of the recipient. What has received less attention is the more consequential question of whether the physics problem SB-AMTI is being asked to solve can actually be solved within the parameters the Space Force has set — a deployed constellation by 2028, using a radar payload on a small satellite bus derived from commercial Starshield architecture. Answering that question requires engaging a set of design tradeoffs that have, in various combinations, defeated every previous attempt to field a militarily useful space-based radar moving target indicator over the past three decades.

Tradeoff 1: The Range Equation — The Fundamental Tax

Everything in space-based radar begins with the radar range equation, and the news is not encouraging. Signal-to-noise ratio scales as the inverse fourth power of range. An airborne radar like the E-7 Wedgetail's MESA system operates at roughly 9–12 km altitude above the target area; a LEO satellite at 500 km altitude is approximately 40–55 times farther away. That factor, taken to the fourth power, means a space-based radar requires roughly three million times more radiated power-aperture product to achieve the same SNR as an airborne radar — before any other complicating factors are introduced. In practice, Aviation Week's own analysis of the space-based GMTI problem put the required antenna aperture at approximately 113 m² in LEO at 500 km to match the SNR of a 4.43 m² antenna on an airborne platform at 45,000 ft, assuming equal transmit power.

Radar Range Equation — Key Dependencies

SNR scales as: SNR ∝ (P × Ae²) / (R⁴ × σ_clutter)
where P = transmit power, Ae = effective aperture, R = slant range, σ_clutter = clutter cross-section.

For LEO at 500 km vs. airborne at 9 km: R ratio ≈ 55× → R⁴ ratio ≈ 9.2 × 10⁶×

Recovery via aperture: doubling Ae gains 6 dB. Recovering 70 dB via aperture alone requires Ae_space ≈ 3,000 × Ae_airborne — plainly impossible on a small satellite bus. The practical path is to combine modest aperture gains, higher transmit power, higher PRF where range ambiguity permits, and onboard STAP processing — with all of the associated satellite mass, thermal, and cost penalties each entails.

The mitigation options are well understood and all expensive. A larger phased-array antenna reduces the aperture gap but adds mass and drives up satellite cost beyond the unit price point that a proliferated constellation requires. Higher transmit power accelerates thermal management challenges and reduces satellite lifetime. Higher pulse repetition frequency improves coherent integration time but introduces range ambiguity that must be resolved by waveform design. Lower orbit improves the range equation but increases atmospheric drag, shortens orbital lifetime, and reduces the satellite's dwell time over any given target area. The FY2027 budget documents confirm that the SpaceX SB-AMTI award covers a high-band radar system — almost certainly X-band or Ku-band — where shorter wavelengths allow smaller antenna apertures for equivalent angular resolution, at the cost of greater atmospheric attenuation and higher sensitivity to precipitation and sea-surface clutter at low grazing angles.

Tradeoff 2: Platform Velocity and the STAP Problem

Here the SB-AMTI engineering problem diverges most dramatically from the familiar airborne AMTI experience. An E-7 Wedgetail or E-3 Sentry flies at roughly 250 m/s. A LEO satellite at 500 km altitude moves at approximately 7,500 m/s — thirty times faster. The consequence for Moving Target Indication is severe and not widely appreciated outside the radar community.

In airborne AMTI, the platform's motion smears the clutter Doppler spectrum. A target must exceed the clutter spread — the Minimum Detectable Velocity (MDV) — to be extracted from the clutter background. MDV scales directly with platform velocity: at 7,500 m/s, the mainlobe clutter Doppler spread from a LEO satellite covers an enormous range of radial velocities, and a target flying tangentially to the satellite's ground track — precisely the geometry relevant to a cruise missile approaching from a perpendicular bearing — will have a radial velocity component that may fall squarely within the clutter notch.

Space-Time Adaptive Processing (STAP) is the signal processing architecture designed to address this. By jointly processing across both spatial (antenna array element) and temporal (pulse-to-pulse) dimensions, STAP can adaptively null the clutter in the joint angle-Doppler space, recovering targets that simple Doppler filtering cannot reach. The challenge from a spaceborne platform is that the classical STAP assumption of independent and identically distributed (IID) training data breaks down: Earth's curvature means the clutter Doppler center frequency varies as a function of range, producing what the literature terms "range-dependent non-stationarity." The IID training samples from adjacent range cells that a ground-based or airborne STAP processor would use to estimate the clutter covariance matrix are no longer statistically homogeneous — they come from different points on the Earth's surface with different look angles, different clutter reflectivities, and different platform-relative geometries.

This is not an exotic edge case. It is the defining challenge of spaceborne radar signal processing, and it is one reason that a 2002 IEEE paper on space-based radar moving target detection concluded that Displaced Phase Center Antenna (DPCA) techniques — conceptually simpler than STAP but requiring precise inter-element phase matching — "may well provide the better cost/performance trade-off for SBR" despite the STAP-heavy literature in the airborne domain. Modern answers to the non-stationarity problem include sparse recovery STAP, which exploits the known sparsity of clutter in the angle-Doppler domain to estimate the covariance matrix with fewer training samples, and sparse Bayesian learning formulations that eliminate the need to select regularization parameters. These techniques are computationally intensive and require significant onboard processing capability — which cycles back directly to satellite mass, power, and cost.

"I think there's a lot to be learned from GMTI, how do we apply it to AMTI, and then really just fleshing out all the different phenomenologies, and what's the best in the AOA combination. Some targets will be bigger, some smaller, faster movers. What's the best phenomenology of how you would detect those from space, which is very different than doing it on a very close airborne or ground-based platform?" — Lt. Gen. DeAnna Burt, Deputy Chief of Space Operations, August 2025

Tradeoff 3: The Dual-Band Architecture — Search vs. Track

The FY2027 budget documents and SpaceNews reporting on the Pentagon's budget request have confirmed that SB-AMTI is not, in fact, a single-sensor architecture. The program funds a high-band radar component — the production constellation toward which SpaceX's $4.16 billion is directed — and separately, a $140 million development effort for complementary lower-frequency sensing for wide-area search. This dual-band split is a direct engineering response to the physics of the AMTI mission from LEO.

High-band radar (X or Ku band) offers several advantages for precision tracking: finer Doppler resolution for a given integration time, smaller antenna aperture requirement for equivalent angular resolution, and better compatibility with small satellite bus constraints. But it is limited as a search sensor. Wide-area coverage at high band requires either a very large antenna, a very high PRF (with the attendant range ambiguity), or a very long dwell time — all of which are in tension with the need to search a large volume of airspace continuously.

Low-band radar (L or S band) inverts these tradeoffs: coarser angular resolution but broader beam coverage, longer wavelengths that are less sensitive to precipitation clutter, and a Doppler resolution regime that tolerates greater target velocity uncertainty without ambiguity. As a search and cue sensor rather than a precision tracker, low-band provides the "find" function; high-band provides the "fix" and "track" functions. The $140 million investment for low-band development acknowledges that the high-band constellation alone cannot perform uncued wide-area search of contested airspace. It will need a cue — either from the low-band tier or, for missile targets, from the HBTSS/SDA Tracking Layer infrared sensors.

This is not a new insight. GlobalSecurity's summary of the earlier Space Based Radar program analysis noted explicitly that "SBR AMTI would not have a capability to search airspace and must instead either be cued by another system or maintain track from the point of launch." The dual-band architecture is the Space Force's acknowledgment that this remains true in 2026.

Tradeoff 4: Constellation Size, Revisit Rate, and the Persistence Requirement

The central operational requirement for AMTI — as distinct from SAR imaging or GMTI of slow ground vehicles — is not occasional coverage but continuous custody. A cruise missile tracked for 90 seconds and then lost during a satellite handoff is an operationally useless data product; it provides geolocation at a moment in time but not the persistent track-before-shoot that a fire control system requires. Achieving continuous custody from LEO requires either overlapping satellite fields of regard or sufficiently rapid revisit that track prediction can bridge the gap.

The numbers are sobering. For a constellation at 500 km altitude, each satellite has a field-of-regard footprint of roughly 2,500–3,000 km diameter for a wide-angle radar. At LEO orbital velocity, the satellite passes over a fixed theater in approximately 8–12 minutes, then is gone for the remainder of its ~90-minute orbital period. A single satellite provides roughly 12% duty cycle over a given theater. Nine satellites in an optimized Walker constellation provide perhaps 65% coverage probability — meaningful for some intelligence functions but grossly inadequate for weapons-quality tracking. Twenty-one satellites approach continuous theater coverage. The leaked figure of approximately 600 satellites for a global SB-AMTI constellation, reported in advance of the contract announcement, is the logical product of this arithmetic applied globally rather than to a single theater.

The 2028 initial operational capability milestone almost certainly represents a regional rather than global capability — a subset of the eventual constellation that provides coverage of priority theaters (likely the Indo-Pacific and North Atlantic) with the same limitations on persistence that any small constellation imposes. The FY2027 budget language's reference to "regional operational requirements" progressing "toward global coverage" confirms this phased interpretation.

Constellation Scaling — Persistence vs. Cost

• ~9 satellites at 53° inclination: ~65% probability of coverage over a given theater; gaps of 30–40 minutes per orbit period. Adequate for periodic intelligence collection; inadequate for continuous weapons custody.
• ~21 satellites: Near-continuous multi-theater access. Approaches the minimum for operationally useful AMTI tracking of a maneuvering threat over a sustained engagement timeline.
• ~600 satellites (reported full constellation): Global persistent coverage with handoff overlap. Required for the homeland defense mission against all-azimuth threat axes. Implies ~100–200 satellite launches per year for a 3–5 year replacement cycle.
• Handoff custody problem: Each inter-satellite track handoff introduces a gap during which the target's state must be propagated from last known track. A cruise missile at 250 m/s introduces 2.5 km positional uncertainty per 10-second handoff gap — potentially outside the next satellite's acquisition gate at high-band angular resolution.

Tradeoff 5: Orbit Inclination and the Polar Threat Corridor

The choice of orbital inclination is one of the most consequential and least discussed design decisions in the SB-AMTI architecture. For homeland defense against Russian and Chinese hypersonic threats, the dominant attack corridors are polar and sub-polar: over-the-pole trajectories for Russian ICBMs and Avangard HGVs, depressed-trajectory approaches from the Arctic for Kinzhal-class systems launched from long-range bombers, and maritime-launched cruise missiles from submarines operating in the Norwegian Sea or Arctic Ocean. These geometries all require coverage above 60° latitude — precisely where mid-inclination Walker constellations at 45–55° have minimal dwell time.

Senator Lisa Murkowski's persistent advocacy for AWACS replacement in the AWACS debate throughout 2025 and 2026 was not parochial: Alaska's geographic position directly astride these threat corridors makes arctic airspace coverage a first-order operational requirement, not a regional preference. A sun-synchronous orbit (approximately 98° inclination) provides polar coverage but creates a retrograde geometry that complicates Doppler processing and produces uneven revisit at lower latitudes. High-inclination prograde orbits (70–80°) balance arctic coverage with reasonable mid-latitude dwell time but require more satellites than a lower-inclination system to achieve equivalent coverage of equatorial and tropical theaters. The six-hundred-satellite figure likely incorporates this high-inclination architecture, which explains why it is so much larger than constellations sufficient for theater-only coverage.

Tradeoff 6: Radar Waveform — PRF, Ambiguity, and LPI

For airborne AMTI radar, waveform design is well understood. The pulse repetition frequency must be chosen to avoid range ambiguity (high PRF blinds ranges beyond PRF interval) while achieving sufficient Doppler unambiguity to separate target returns from clutter. The classic tradeoff between high PRF (good velocity coverage, range-ambiguous) and low PRF (unambiguous range, poor MDV) is addressed by staggered PRF and coded waveforms. From a LEO platform, this tradeoff is more severe: ranges to the ground extend from zero (nadir) to the radar horizon at 2,500+ km, spanning a range gate depth that would require a PRF below approximately 100 Hz for unambiguous coverage — producing a coherent integration burst far too short for adequate Doppler resolution against slow targets.

The practical answer is waveform diversity with a priori knowledge of the engagement geometry. If the system is cued to a specific target location by the low-band search tier or by HBTSS, the high-band tracker can be initialized to a narrow range gate around the expected target range, relieving the need for unambiguous wide-area coverage. This is another manifestation of the cued versus uncued architecture distinction: the uncued case is essentially unsolvable with a small satellite bus; the cued case is addressable with modern waveform and processing techniques.

The low probability of intercept dimension adds complexity. The SB-AMTI radar must transmit to detect targets. Any transmitted signal is detectable by an adversary's electronic intelligence system, revealing the satellite's orbital parameters, waveform characteristics, PRF stagger, and scan patterns — exactly the information needed to design a repeater jammer, predict tracking gates, or time an ASAT engagement. Spread-spectrum LPI waveforms address this partially but reduce peak power and thus SNR per pulse, tightening the already constrained link budget. The classification of specific sensor parameters in the SpaceX contract announcement suggests the program is treating waveform and frequency choice as sensitive information — a reasonable posture given that an adversary who reverse-engineers the scan pattern from emissions intelligence can significantly degrade the system's effectiveness against a prepared threat.

Tradeoff 7: The Distributed Formation Flying Alternative

The dominant architecture assumption underlying the SpaceX award — many moderately capable individual satellites — is not the only engineering answer to the space-based radar problem. Array Labs, a Seattle-area startup that received an Office of Naval Research contract in October 2025 to study space-based AMTI feasibility, is pursuing a fundamentally different approach: formation-flying radar satellites that function as a distributed aperture, with along-track and cross-track baselines between satellites substituting for a physically large antenna on a single platform.

The distributed aperture approach addresses the range equation problem directly. By coherently combining received signals from multiple formation-flying satellites separated by hundreds of meters to kilometers along their orbital track, the system synthesizes an effective aperture far larger than any single satellite can carry. A longer along-track baseline (ATB) improves the MDV by sharpening the Doppler null around the clutter ridge — the fundamental mechanism by which DPCA and interferometric AMTI reduce MDV below the limit imposed by a single-aperture sensor. The IEEE Transactions on Geoscience and Remote Sensing analysis by Chen et al. (2022) established the signal models for distributed space-based radar AMTI and confirmed that the longer ATB of a formation system "is a good candidate due to the longer along-track baseline and spatial power synthesis" for improving MDV performance against weak targets.

The engineering costs are different but equally real. Formation flying requires precise relative orbit determination and station-keeping — Array Labs reports single-pass 3D imaging capability from two test satellites launched in late 2024, which implies the inter-satellite metrology is achievable at useful accuracy levels. But the inter-satellite baseline stability required for coherent combining is orders of magnitude tighter than that required for independent passive imaging, and the grating lobe problem in AMTI — where the sparse configuration of a distributed formation produces spatial ambiguities that "cause the non-continuous detection phenomenon of an air moving target" — requires careful ATB optimization. Array Labs' approach is still in the study phase for AMTI specifically; its commercial product line is focused on 3D surface mapping, where AMTI's demanding coherence requirements don't apply.

Whether the formation flying approach or the proliferated single-satellite approach ultimately proves more cost-effective for weapons-quality AMTI custody is an open question. The Space Force's multi-vendor strategy — nine companies in the OTA pool, with SpaceX receiving the first major award — is partly an explicit hedge against not knowing the answer. Additional vendor awards over the coming year may reveal whether any other pool members are pursuing formation-based architectures.

Tradeoff 8: Kill Chain Latency and the JADC2 Integration Tax

Even a technically perfect sensor is operationally useless if its output arrives at the shooter after the engagement window has closed. Against a cruise missile traveling at Mach 0.8 (approximately 270 m/s), a five-second track-to-shooter latency represents 1.35 km of positional uncertainty — manageable for a terminal seeker but problematic for a long-range interceptor that must receive its launch cue before the target is within engagement range. Against a hypersonic glide vehicle at Mach 8–15, the same five-second latency produces 12–20 km of uncertainty, which likely falls outside the engagement envelope of a Glide Phase Interceptor if the uncertainty is not characterized and bounded.

The end-to-end latency budget for a space-based track-to-shooter chain includes: satellite onboard STAP processing time, uplink to the intersatellite link mesh, transit through the SDA transport layer (or the SpaceX Space Data Network Backbone), downlink to a ground terminal or airborne relay, ingestion by the battle management system, translation into a J-series BOM fire control message compatible with CEC architecture, delivery to the shooter, and the shooter's own engagement timeline. Each link adds latency; each translation between data formats adds latency and introduces potential for track association errors. The Space Force's requirement that SpaceX integrate SB-AMTI sensors with the Space Data Network Backbone — the companion $2.29 billion Starshield communications contract awarded three days earlier — is the programmatic acknowledgment that sensor and pipe must be co-designed, not bolted together after the fact.

The onboard processing demand is particularly acute. Classical STAP processing of a full-DOF covariance matrix for a multi-channel phased array radar is computationally intensive: the computational load scales as the cube of the number of space-time degrees of freedom, which for a useful spaceborne system means teraflop-class processing on a radiation-hardened platform with tight power and thermal constraints. Reduced-dimension and reduced-rank STAP implementations trade some performance for tractable computational load, and AI-enabled processing — explicitly referenced in the Space Systems Command's program description — can be interpreted as a label for neural network-based clutter covariance estimation and target detection that bypasses some of the classical STAP computational bottleneck. Whether the AI processing approach can achieve the required probability of detection and false alarm rate on a representative target population against realistic clutter statistics remains an open systems engineering question, not yet answered in open literature.

"Radar based AMTI from space is feasible. The technology has matured and it's commoditized." — Derek Tournear, outgoing SDA Director, Breaking Defense, 2025

Synthesis: What the 2028 Milestone Actually Means

Reading the tradeoffs together, a coherent picture of the SB-AMTI architecture emerges that is significantly more nuanced than the headline contract numbers suggest. The high-band constellation SpaceX is building — almost certainly using a phased-array radar payload on a Starshield bus — is a cued tracker, not a wide-area search sensor. It requires a cue: from the low-band wide-area search tier still in development, from HBTSS or the SDA Tracking Layer for missile targets, from ground-based early warning radar, or potentially from national technical means through the NRO collaboration that has been quietly underway for several years. The 2028 IOC constellation will provide initial regional coverage over priority theaters, with the inter-satellite handoff custody problem partially mitigated by onboard track prediction and AI-assisted state estimation, and with STAP processing implemented in reduced-dimension form that accepts some MDV penalty in exchange for onboard computational tractability.

The dual-band architecture represents a genuinely thoughtful systems engineering response to the physics. By separating the search function (low-band, wider beams, coarser resolution, lower cost per satellite, fewer required in constellation for coverage) from the precision tracking function (high-band, tighter beams, better Doppler resolution, but limited to a pre-designated volume), the Space Force avoids demanding that a single satellite design simultaneously optimize for mutually contradictory requirements. This is architecturally sound — but it also means the full operational capability of the system depends on both tiers being operational, integrated, and mutually cued, which is a more complex integration problem than a single-band system.

Tournear's assessment that radar-based AMTI from space is "feasible" and "commoditized" is defensible at the level of individual technology readiness. Phased-array radar on a small satellite bus is not exotic; coherent Doppler processing is understood; STAP algorithms exist in the literature. What is not commoditized is the integration of all of these elements into a system that achieves weapons-quality custody of a maneuvering cruise missile in realistic clutter environments, with sub-second end-to-end latency, at the constellation scale required for persistent global coverage — and delivers it on a 2028 schedule from a standing start in 2026. The engineering challenge is real. The execution risk is real. And the adversary — who is reading the same open literature from Xidian University's National Key Laboratory of Radar Signal Processing that the Space Force is — will not be standing still.

Summary: Key SB-AMTI Design Tradeoffs

Design Dimension	Higher Performance Choice	Cost / Risk Accepted	Space Force Architecture Response
Orbit altitude	Lower LEO → better range equation	Shorter pass time; drag; shorter satellite life	~500 km LEO expected; proliferated constellation compensates for short dwell
Antenna aperture	Larger → better SNR & MDV	Mass, cost, thermal; incompatible with small bus	High-band (X/Ku) reduces aperture needed per beam; formation flying studied as alternative
Radar frequency band	High-band: finer resolution, smaller aperture per beam	Poorer wide-area coverage; atmospheric attenuation	Dual-band: high-band ($7B) for precision track + low-band ($140M) for wide-area search
Platform velocity / STAP	Reduced-dim STAP: computationally tractable	MDV penalty vs. full-DOF; non-stationarity unsolved by simple methods	AI-enabled processing referenced in program docs; sparse STAP variants likely
Constellation size	More satellites → persistence, global handoff	Cost, launch rate, production scale; ~600 for global continuous coverage	Phased approach: regional IOC 2028; global expansion post-2028 contingent on $7B FY27 funding
Orbit inclination	High inclination → Arctic / polar threat corridor coverage	Larger constellation required for equivalent lower-latitude coverage	Inclination not publicly disclosed; Arctic coverage politically required (Alaska AWACS gap)
Cued vs. uncued search	Uncued: autonomous wide-area airspace surveillance	Physically requires low-band search tier; high-band alone cannot search uncued	Cued tracking architecture; cue from low-band, HBTSS, ground radar, or NRO assets
Track-to-shooter latency	Lower latency: onboard processing, direct ISL relay	Onboard compute mass/power; JADC2 format translation adds time	SDNB co-design with AMTI; Starshield ISL for low-latency relay; AI processing for onboard STAP
LPI / electronic attack resistance	Spread spectrum, low average power: harder to exploit	Reduced peak SNR; tightens already constrained link budget	Waveform parameters classified; Space Force has not disclosed frequency or PRF approach
Single-satellite vs. formation flying	Formation flying: larger effective aperture; better MDV	Inter-satellite metrology demands; ATB grating lobe problem; station-keeping cost	Array Labs / ONR studying distributed approach; SpaceX award uses single-satellite architecture

Verified Sources & Citations

1. SpaceNews, "Pentagon Budget Affirms Space Force Role Tracking Moving Targets from Orbit," April 28, 2026. https://spacenews.com/pentagon-budget-affirms-space-force-role-tracking-moving-targets-from-orbit/
2. The Defense News (thedefensenews.com), "US Space Force Advances Space-Based AMTI System with $140 Million Low-Band Radar Investment," April 28, 2026. https://www.thedefensenews.com/news-details/US-Space-Force-Advances-Space-Based-AMTI-System-with-140-Million-Low-Band-Radar-Investment/
3. Military Times, "SpaceX Awarded $4 Billion Space Force Contract to Track Airborne Threats," May 29, 2026. https://www.militarytimes.com/industry/techwatch/2026/05/29/spacex-awarded-4-billion-space-force-contract-to-track-airborne-threats/
4. Air and Space Forces Magazine, "Space Force Gives SpaceX $4.2B for Satellites to Track Airborne Targets," May 29, 2026. https://www.airandspaceforces.com/space-force-spacex-4b-airborne-target-tracking-satellites/
5. The War Zone (TWZ), "Tracking Moving Aircraft Via Radar Satellites Instead of Surveillance Jets Still Far From Reality," Aug. 4, 2025. https://www.twz.com/space/track-moving-aircraft-via-radar-satellites-instead-of-surveillance-jets-still-far-from-reality
6. The War Zone (TWZ), "Pentagon's Plans To Track Aircraft From Orbit Accelerated With New $4B SpaceX Deal," May 29, 2026. https://www.twz.com/space/pentagons-plans-to-track-aircraft-from-orbit-accelerated-with-new-4b-spacex-deal
7. Breaking Defense, "Space Force Testing Space-Based Sensors to Track Airborne Targets," May 15, 2025. https://breakingdefense.com/2025/05/space-force-testing-space-based-sensors-to-track-airborne-targets/
8. Aviation Week Network, "Space-Based Ground Moving Target Tracker Revealed," June 2021. https://aviationweek.com/defense/sensors-electronic-warfare/space-based-ground-moving-target-tracker-revealed
9. Aviation Week Network, "Array Labs To Study Space-Based AMTI For Office Of Naval Research," Oct. 9, 2025. https://aviationweek.com/space/satellites/array-labs-study-space-based-amti-office-naval-research
10. Aviation Week Network, "Debrief: Array Labs' Old-New School Approach To Space-Based AMTI," Jan. 14, 2026. https://aviationweek.com/defense/sensors-electronic-warfare/debrief-array-labs-old-new-school-approach-space-based-amti
11. Array Labs, "Array Labs Awarded U.S. Navy Contract to Study Space-Based AMTI Concepts," Oct. 6, 2025. https://www.arraylabs.io/updates/2
12. Array Labs, "Array Labs Unveils Satellite Design and First Production Cluster," April 2, 2025. https://www.arraylabs.io/updates/7
13. ExecutiveGov, "Space Force Launches SB-AMTI Multi-Vendor Strategy," April 28, 2026. https://www.executivegov.com/articles/space-force-space-based-amti-multi-vendor-strategy
14. Chen, J. et al., "Multi-channel Signal Modeling and AMTI Performance Analysis for Distributed Space-based Radar Systems," IEEE Transactions on Geoscience and Remote Sensing, 2022. DOI: 10.1109/TGRS.2022.3202567. https://udspace.udel.edu/items/f5c25ee9-2f94-470c-aed9-24e286e39c09
15. Deng, Z. et al., "Approach for AMTI Formation Design in a Distributed Space-based Radar System," IEEE IGARSS 2024. DOI: 10.1109/IGARSS53475.2024.10641395. https://ieeexplore.ieee.org/document/10641395/
16. Guang, X. et al., "Clutter Suppression with Doppler Frequency Shifted LMS Filtering in LEO Satellite-Based Passive Radar," Remote Sensing 17(17):3096, Sept. 2025. DOI: 10.3390/rs17173096. https://www.mdpi.com/2072-4292/17/17/3096
17. Liu, C. et al., "A Fast Space-Time Adaptive Processing Algorithm Based on Sparse Bayesian Learning for Airborne Radar," Sensors 22(7):2664, March 2022. DOI: 10.3390/s22072664. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9003390/
18. Ren, B. and Wang, T., "Space-Time Adaptive Processing Based on Modified Sparse Learning via Iterative Minimization for Conformal Array Radar," Sensors 22(18):6917, Sept. 2022. DOI: 10.3390/s22186917. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9500770/
19. IEEE Xplore, "Spaceborne Radar Design Equations and Concepts," (including AMTI mode design constraints). DOI: 10.1109/IGARSS.1996.574402. https://ieeexplore.ieee.org/document/574402/
20. Klemm, R., Principles of Space-Time Adaptive Processing. Relevant sections on spaceborne radar range-dependent clutter non-stationarity, DPCA vs. STAP for SBR, and MDV scaling with platform velocity. IEE/IET, 3rd Ed., 2006. (Reference standard in the discipline.)
21. GlobalSecurity.org, "Space Based Radar (SBR) — Configuration and AMTI Analysis," derived from DoD/RAND program documentation. https://www.globalsecurity.org/space/systems/sbr-config.htm
22. GlobalSecurity.org, "Space Based Radar and Missile Defense — AMTI Cueing Requirement." https://www.globalsecurity.org/space/systems/sbr-md.htm
23. DTIC, "Space Based Radar — System Architecture Design and Tradeoffs" (MIT thesis, archived). ADA329019. https://apps.dtic.mil/sti/tr/pdf/ADA329019.pdf
24. Sun, K. et al., "Airborne Radar STAP using Sparse Recovery of Clutter Spectrum," arXiv:1008.4185. https://arxiv.org/pdf/1008.4185
25. Yee Wei Law et al., "Detecting and Tracking Hypersonic Glide Vehicles: A Cybersecurity-Informed Analysis," Proc. ICCWS 2025. https://papers.academic-conferences.org/index.php/iccws/article/download/950/955