Disaggregated Satellite Systems: Architecture of the Processing Core, Sensor Modules, and Operational Rollout

Key Takeaways

• Disaggregated satellites split computing, control, and sensing across swappable independent modules
• A central processing hub manages power, pointing, and data routing for all attached sensor payloads
• Modular designs let hyperspectral, EO, and SAR sensors be swapped or upgraded independently

Why Monolithic Spacecraft Design Has Reached Its Limits

The first commercially available synthetic aperture radar small satellite, ICEYE-X1, launched in January 2018 with a 3.25-metre deployable antenna folded into a microsatellite chassis. That engineering feat required every subsystem on board to share the same power budget, the same thermal envelope, the same attitude control system, and the same structural frame. If the radar antenna needed more power, the onboard computer got less. If the optical communications terminal required a clear field of view, the radar beam geometry had to accommodate it. Everything was a trade-off, and every trade-off was permanent from the moment the rocket left the launch pad.

That trade-off model has defined satellite engineering since Sputnik. A satellite is designed around a specific mission, integrated over years, launched once, and operated with whatever capabilities survived the process. Changing the mission means building a new satellite. Upgrading one sensor means redesigning the entire bus. Adding a second payload type means recalculating structural load paths, thermal dissipation profiles, and power management sequences all over again. The design is monolithic by nature, and monolithic designs make the spacecraft both expensive to build and brittle in operation.

The monolithic satellite architecture starts to break down when mission requirements become more complex. Intelligence, surveillance, and reconnaissance missions increasingly demand that a single observation asset deliver electro-optical imagery in sunlit conditions, synthetic aperture radar data through cloud cover, hyperspectral spectral fingerprints for material identification, and signals intelligence simultaneously or in rapid succession. Packing all of that capability into a single spacecraft at sufficient performance levels drives mass, volume, and cost to levels that make launch options scarce and total mission costs prohibitive. The result is that the most capable spacecraft are also the most expensive to replace, which makes them high-value targets in contested environments and slow to adapt when sensor technology improves.

Disaggregation offers a different logic. Instead of consolidating all mission functions into one structure, a disaggregated system distributes them across multiple modules that communicate through defined interfaces. Each module performs a specific function. One carries the processing and control infrastructure that all other modules depend on. Others carry mission payloads such as cameras, radar antennas, or spectrometers. The modules can operate in physical contact, in close proximity using wireless links, or as loosely coupled formation-flying nodes in the same orbital plane. What holds the system together is not a shared frame but a shared protocol.

What the Processing and Control Module Actually Does

The processing and control module is the organizational core of a disaggregated satellite system. It is sometimes called the service module, the bus hub, or the mission processing backbone, but whatever the terminology, its function is consistent: it aggregates telemetry from all attached sensor modules, routes commands from ground operators, manages the shared power and thermal budget, maintains attitude knowledge for the entire cluster, and handles downlink to ground infrastructure. Every sensor module plugs into this core, either physically through standardized electrical connectors or wirelessly through inter-module transceivers.

The computing capacity inside this module is substantially larger than what would be found in a dedicated sensor satellite. In a disaggregated system, the processing module provides onboard computing resources that individual sensor modules draw on rather than duplicating separately. A hyperspectral imager, for instance, generates several gigabits of raw spectral data per second during an observation pass. Processing that data on the sensor module itself would require a powerful, power-hungry computer embedded in a payload that already has demanding thermal and mechanical requirements. Moving that computational load to the central processing module, which is purpose-built for high-throughput data handling, lets the sensor module be smaller, lighter, and cheaper.

The architecture of the processing module typically includes several layers. At the lowest level sits the flight computer, which runs the operating system and manages housekeeping functions like attitude determination and control, power conditioning, and thermal management. Above that layer runs the data management system, which handles ingestion of payload data streams, applies compression algorithms, and stages data for downlink. The highest software layer contains mission management logic: task scheduling, observation planning, sensor tasking commands, and autonomous decision-making routines that allow the satellite to act without waiting for ground contact.

Modern processing modules increasingly incorporate field-programmable gate arrays alongside conventional microprocessors. FPGAs allow the same hardware to perform different signal-processing functions depending on which sensor module is active. An FPGA configured for radar signal processing handles the matched filtering, range compression, and azimuth compression that raw SAR data requires. Reconfigured via software upload, the same FPGA can instead run spectral unmixing algorithms on hyperspectral data cubes. This reconfigurability is one reason why a centralized processing module outperforms a collection of individually-embedded processors. It can be retasked from the ground as mission requirements change, without physical intervention.

Power management is another function the processing module centralizes. Sensor modules have very different power consumption profiles. A synthetic aperture radar payload can draw several hundred watts during active transmission, while a hyperspectral imager running in passive mode draws a fraction of that. The processing module’s power management system allocates available solar-generated power dynamically across the cluster, prioritizing active payloads and ensuring the power budget stays within battery capacity. Without centralized coordination, multiple modules drawing peak power simultaneously could collapse the power bus and put the entire cluster in safe mode.

Attitude control is similarly centralized. Pointing a SAR antenna precisely enough to generate usable imagery requires knowledge of the spacecraft’s orientation accurate to fractions of a degree. Pointing a hyperspectral imager for push-broom imaging requires stable nadir-pointed flight with controlled pitch rotation. The processing module hosts the star trackers, gyroscopes, and reaction wheel controllers that maintain the cluster’s pointing. Sensor modules receive pointing commands and attitude state vectors from the processing module rather than managing their own attitude systems. That centralization reduces mass and complexity in each sensor module and ensures that all payloads operate with consistent attitude knowledge.

How Modules Communicate: Interfaces, Protocols, and the Plug-and-Play Problem

Getting separate modules to behave as a coherent system requires rigorous standardization of interfaces. In a physically connected disaggregated system, modules connect through electrical harnesses using standardized connector footprints that carry power, data, and timing signals. The mechanical interface specifies attachment points, separation distances, and structural load limits. The electrical interface specifies voltage rails, current limits, and connector pin assignments. The data interface specifies the protocol, baud rate, and packet format that the sensor module and processing module use to exchange commands and telemetry.

DARPA’s System F6 program, originally called the Future, Fast, Flexible, Fractionated, Free-Flying Spacecraft United by Information Exchange, explored the wireless end of this problem starting in 2007. The program’s core insight was that the most flexible form of disaggregation treats each module as a uniquely addressable node on an IP-based wireless network. Rather than routing data through physical connectors, modules communicate through inter-module transceivers that form a self-organizing mesh network in orbit. Commands addressed to a specific sensor module travel across the mesh and arrive at the module regardless of where it is positioned relative to the hub. Data flows back the same way. The processing module sees all modules as network peripherals, not as physically integrated subsystems.

The F6 program developed two key artifacts to make this approach scalable. The F6 Developer’s Kit provided open-source interface standards and protocol specifications that any organization could use to build a module compatible with the cluster without a direct contractual relationship with the prime contractor. The F6 Technology Package was a hardware instantiation of the wireless connectivity and encryption stack that could be embedded into an existing spacecraft bus to make it cluster-capable. Together, these tools were meant to create a marketplace of interoperable modules. Although DARPA cancelled the program’s flight demonstration in May 2013 before hardware reached orbit, the technical specifications it developed informed a generation of subsequent modular satellite programs.

Physical connection standards have advanced considerably since the F6 era. The Aerospace Corporation’s Slingshot 1 mission, a 12U CubeSat carrying 19 payloads, demonstrated a plug-and-play interface called Handle that allows sensor payloads to communicate with the host spacecraft bus regardless of the bus manufacturer’s underlying design. Handle standardizes the electrical interface layer so that a payload developed independently can be integrated into the satellite at any point in the development timeline, including late changes that would normally require extensive rework. The mission, which hosted payloads for attitude control, GPS tracking, propulsion, and laser communications simultaneously, showed that a modular interface standard can accommodate diverse payload types on a single host without customization.

In practice, most deployed disaggregated satellite systems today use physical connections for the most demanding data interfaces and reserve wireless links for slower telemetry, inter-satellite coordination, and formation flying management. SAR signal processing, for instance, produces raw data rates that can exceed 1 gigabit per second during active collection. Routing that data wirelessly between a SAR payload module and a processing module would require a high-bandwidth link that consumes more power and introduces more latency than a direct high-speed serial bus. Hyperspectral data, at lower raw rates, is more compatible with wireless transfer. Future systems are expected to move toward optical inter-module links using laser communications technology, which can support multi-gigabit transfer rates over short distances in orbit.

Electro-Optical Modules: Imaging in the Visible and Infrared

An electro-optical (EO) sensor module contains an optical telescope, a focal plane array detector, and the electronics needed to digitize the image and transfer it to the processing module. The telescope collects light reflected from the Earth’s surface in wavelengths ranging from ultraviolet through shortwave infrared. The focal plane array converts that light into a digital signal, one pixel at a time across the detector’s area. The image produced represents what the sensor can see, bounded by spatial resolution, swath width, and spectral band coverage.

EO modules have been evolving for decades. The Landsat program, which began with the launch of Landsat 1 in 1972, established the basic framework of multispectral imaging from orbit. Modern commercial EO modules have compressed those capabilities into increasingly small packages. Maxar Technologies operates WorldView-3, which delivers panchromatic imagery at 31-centimetre resolution and multispectral imagery at 1.24 metres from geosynchronous transfer orbit altitudes. Planet Labs operates a constellation of Dove satellites that sacrifices resolution for revisit frequency, imaging the entire Earth’s land surface daily at 3 to 5-metre resolution.

In a disaggregated architecture, an EO module’s design is constrained primarily by the aperture of its optical system and the sensitivity of its focal plane array. A larger aperture collects more light and supports finer spatial resolution. A more sensitive detector array reduces the integration time needed per frame, which allows the satellite to image faster and collect more frames per orbit pass. The processing module’s role is to receive the raw frame data, apply radiometric calibration coefficients, correct for geometric distortions introduced by the satellite’s motion, and stage the corrected imagery for downlink or onboard analysis.

EO modules designed for disaggregated systems tend to be optimized around specific observation tasks. A wide-area imaging module might use a shorter focal length telescope to maximize swath width at the expense of resolution, covering large areas per orbit pass. A high-resolution spot imaging module uses a longer focal length and a smaller detector to achieve sub-metre resolution across a narrow swath. In a modular system, operators can configure the cluster with both types of modules simultaneously, using the wide-area module for initial area coverage and cueing the high-resolution module to points of interest identified in the first pass. That sequential cueing is only possible when the processing module coordinates the tasking of both sensor modules against a common observation timeline.

Thermal management is a significant engineering challenge for EO modules. Focal plane array detectors produce their best signal-to-noise performance when cooled to low temperatures. A cooled near-infrared or shortwave infrared detector might operate at 150 to 200 Kelvin, requiring either a dedicated cryogenic cooler or a passive radiator pointed away from the Earth and sun. In a monolithic satellite, the cooler’s heat rejection competes with the thermal environment of other subsystems. In a disaggregated system, the EO module’s thermal design can be optimized independently, with the processing module aware of the module’s thermal state through telemetry monitoring and able to adjust observation scheduling to prevent thermal overruns.

Synthetic Aperture Radar Modules and All-Weather Persistence

A synthetic aperture radar sensor module transmits microwave pulses toward the Earth’s surface and receives the reflected signals. The reflected signal carries phase and amplitude information that encodes the surface’s geometry and dielectric properties. The SAR processing algorithm, which runs largely in the processing module given its computational intensity, uses the satellite’s orbital motion to synthesize a very long virtual antenna aperture, producing images with spatial resolution far finer than the physical antenna size would suggest.

The practical advantage of SAR over EO is independence from sunlight and weather. Microwave signals penetrate clouds, rain, smoke, and darkness. ICEYE, the Finnish company that operates the world’s largest commercial SAR constellation with reported 2025 revenues exceeding 250 million euros, built its business around exactly this advantage. Defense and intelligence customers who need to monitor specific locations regardless of weather conditions cannot rely solely on optical systems. A SAR module in a disaggregated satellite cluster provides that all-condition visibility as a complement to, rather than a replacement for, optical modules.

SAR modules come in several operating modes, each suited to different applications. Spotlight mode focuses the radar beam on a small area for an extended dwell time, achieving very high spatial resolutions sometimes below 0.3 metres. Stripmap mode maintains a fixed antenna pointing angle as the satellite passes over an area, producing a continuous swath at moderate resolution. ScanSAR and wide-swath modes sacrifice resolution for coverage, imaging very large areas per orbit pass. In a disaggregated architecture, the processing module can configure the SAR module into different modes on successive passes over the same area, providing both a wide-area situational awareness product from one pass and a high-resolution spot image of a specific target from the next.

The power demands of a SAR module present the most significant challenge for integration into a disaggregated system. An X-band SAR payload on a microsatellite may peak at 200 to 500 watts of RF output power during transmission, with total system power consumption considerably higher. The processing module must ensure that the power management system reserves sufficient charge in the battery bank before a SAR collection event begins, restricts other power-consuming activities during collection, and allows thermal dissipation after the collection ends before commanding another pass. Capella Space, which achieves sub-0.25-metre SAR resolution from small satellites in low Earth orbit, manages similar power constraints across its constellation operations.

SAR module antennas pose mechanical design challenges in a disaggregated context. A phased array antenna panel that provides electronically steerable beam control offers flexibility but adds mass and complexity. A deployable reflector antenna that unfolds after launch can provide a larger aperture than a body-mounted panel but introduces articulation points that can become failure modes. The choice between these options depends on the mission profile and the mechanical interface standards defined by the processing module’s attachment structure. A modular system that standardizes on fixed-dimension attachment ports will tend to favor compact folded-array or deployable designs that fit within the defined envelope.

Interferometric SAR modes, in which two or more SAR modules collect data over the same area at the same time from slightly different positions, are particularly compelling in a disaggregated architecture. Interferometric SAR (InSAR) uses the phase difference between two collected images to measure surface deformation at centimetre-scale accuracy, enabling applications like earthquake damage mapping, volcanic deformation monitoring, and infrastructure subsidence detection. In a monolithic system, InSAR requires two physically separate satellites, each with its own complete suite of avionics, communications, and power systems, flying in a carefully controlled formation. In a disaggregated cluster, a second SAR module attached to or flying near the same processing hub could share the hub’s communications, computing, and ground infrastructure while operating as the second antenna element for InSAR collection. The incremental cost of adding InSAR capability to an existing cluster would be the mass and cost of the second SAR module, not an entirely new spacecraft.

Bistatic SAR modes extend this concept further. A bistatic configuration separates the transmitting antenna from the receiving antenna, with one element actively transmitting pulses and one or more other elements passively receiving reflections from the illuminated scene. Bistatic geometry reduces the illumination footprint’s specular reflection component and can reveal target signatures that monostatic SAR, in which transmitter and receiver are co-located, would miss. Commercial bistatic SAR constellation concepts, including those explored by European and American providers, require two or more cooperating spacecraft. A disaggregated multi-module architecture where one module transmits and another receives makes bistatic collection operationally manageable by centralizing the timing, command, and data correlation functions in the shared processing hub. Airbus Defence and Space has explored bistatic SAR concepts using its TerraSAR-X and TanDEM-X satellites, which fly in a close formation to produce digital elevation models through interferometric processing, providing a model for how cooperating radar modules could work within a disaggregated architecture.

Hyperspectral Imaging Modules: Reading the Spectral Fingerprint

Hyperspectral imaging extends EO capability into a much richer spectral space. Where a conventional multispectral camera captures four to ten broad spectral bands, a hyperspectral imaging module captures hundreds of narrow contiguous bands spanning the visible through shortwave infrared spectrum. Each band is typically less than 10 nanometres wide. The resulting data product is a three-dimensional hyperspectral cube with two spatial dimensions and one spectral dimension, where every pixel in the image contains a detailed spectral signature of the surface material at that location.

The spectral signatures encoded in hyperspectral data are physically unique to different materials. Healthy vegetation reflects strongly in the near-infrared and absorbs in the red portion of the visible spectrum, producing a characteristic signature that distinguishes it from stressed vegetation, bare soil, or open water. Different mineral types reflect and absorb light at different wavelengths in the shortwave infrared, allowing a hyperspectral sensor to identify rock composition from orbit. Industrial chemicals and petroleum products have spectral signatures in the mid-infrared that can reveal pipeline leaks or surface spills invisible to conventional cameras.

Pixxel, the US-India space technology company, launched the first three of its commercial Firefly hyperspectral satellites on January 14, 2025 as rideshare payloads aboard SpaceX’s Transporter-12 mission. Three additional Firefly satellites followed on August 26, 2025, completing the first operational phase of Pixxel’s constellation at six satellites. Each Firefly captures imagery across more than 135 spectral bands at 5-metre spatial resolution with a 40-kilometre swath width, making it the highest-resolution commercial hyperspectral offering as of April 2026. Pixxel holds contracts with both NASA and the National Reconnaissance Office (NRO), reflecting the sensor type’s appeal to both scientific and intelligence users.

In a disaggregated satellite system, a hyperspectral module poses specific data management challenges. A 300-band hyperspectral imager operating at 5-metre resolution over a 40-kilometre swath can generate data at rates exceeding 5 gigabits per second during continuous collection. The central processing module must ingest that data stream, buffer it in solid-state storage, and apply compression before downlink. Lossless compression of hyperspectral data is limited by the data’s information density. Lossy compression can reduce file sizes substantially but risks discarding subtle spectral features that analysts need. Onboard processing algorithms that perform targeted analysis, retaining only the spectral anomalies of interest, can reduce downlink volume by an order of magnitude compared to raw data transmission.

The Copernicus Hyperspectral Imaging Mission for the Environment (CHIME), the European Space Agency’s planned hyperspectral Earth observation mission, was designed with a projected data downlink requirement of 3.7 gigabits per second, more than 20 times the data rate of the Sentinel-2 multispectral mission. Managing that throughput requires a processing module with high-speed storage interfaces, powerful compression processing, and flexible downlink scheduling. In a disaggregated system where the hyperspectral module operates as a separate node from the processing hub, the data interface between the two becomes the bottleneck that determines how much spectral data can actually be exploited.

Imec, the Belgian semiconductor research center, demonstrated a new generation of hyperspectral sensors at the Small Satellite Conference in August 2024 that employ line-based spectral filters deposited directly on the image sensor chip. This approach creates push-frame hyperspectral sensors that relax pointing accuracy requirements, particularly relevant for small satellites where attitude control is less precise than on larger platforms. In a disaggregated architecture, a sensor module using this type of integrated filter design would be lighter and less mechanically complex than one using traditional dispersive spectrometer optics, while still delivering the spectral coverage needed for material identification applications.

AIS, Signals Intelligence, and Additional Sensor Payloads

Disaggregated satellite systems are not limited to imaging sensors. The same modular architecture that accommodates a hyperspectral imager or a SAR antenna can host signals collection payloads, Automatic Identification System (AIS) receivers, radio frequency monitoring equipment, and other non-imaging sensors. The processing module’s standardized interface and power management system treats each payload type the same way regardless of whether it produces a two-dimensional image or a stream of RF signal data.

AIS receivers detect the radio transmissions that ships are required to broadcast under international maritime regulations. Each transmission carries the vessel’s identity, position, course, and speed. A satellite-borne AIS receiver can hear transmissions from ships across a wide maritime area and aggregate thousands of position reports per orbit pass. In a disaggregated system, an AIS receiver module operates in parallel with an EO or SAR module, allowing operators to correlate the optical or radar images of a specific ocean area with AIS-reported vessel positions in that same area. Vessels that appear in the imagery but do not match any AIS transmissions represent dark ships, a category of maritime intelligence interest.

Signals intelligence (SIGINT) payloads are more complex, collecting and characterizing radio frequency emissions from ground-based transmitters. A SIGINT module in a disaggregated cluster can geolocate emitters using time-difference-of-arrival techniques across multiple modules in different orbital positions, or can characterize the technical parameters of a detected signal to identify the type of system producing it. The processing module’s role in SIGINT operations is particularly demanding. Raw signal intercept data is highly voluminous and computationally expensive to process. Matching detected signals against libraries of known emitter signatures requires a database and pattern-matching capability that benefits from the processing module’s centralized computing resources.

Spire Global operates a constellation of small satellites that carry multiple payload types including GPS radio occultation receivers for weather sensing, AIS receivers, and ADS-B receivers for aircraft tracking. The company’s approach of combining multiple sensor types on a common satellite bus represents an early commercial implementation of the disaggregated concept at the constellation level, even if individual satellites are not themselves disaggregated. The logical extension of Spire’s model to a fully disaggregated architecture would have each sensor type hosted as a swappable module on a common processing hub, allowing the ratio of weather sensors to maritime sensors to AIS receivers to be adjusted as market demand shifts.

Environmental monitoring payloads represent another growing category of additional sensor modules. Methane and carbon dioxide sensors that detect greenhouse gas concentrations using spectroscopic measurement in the infrared can operate as sensor modules on a disaggregated satellite alongside conventional EO or hyperspectral modules. Orbital Sidekick, a US company operating a constellation of hyperspectral satellites built on Astro Digital’s Corvus bus, uses its GHOSt satellites to pinpoint methane leaks across oil and gas pipeline networks by capturing nearly 500 spectral bands of data. That combination of a capable sensor module and a standardized satellite bus is structurally analogous to the disaggregated model, with the bus functioning as a simplified processing hub for the hyperspectral payload.

How the Processing Module Fuses Data Across Sensors

The most operationally significant capability of a disaggregated system is sensor fusion: the combination of data products from multiple sensor types into intelligence products that are more informative than any single sensor could produce alone. A SAR image of a port facility reveals the location and general dimensions of vessels at berth but cannot identify vessel types reliably. An EO image of the same port reveals fine structural details sufficient for type classification but requires daylight and clear weather. An AIS log from the same timeframe identifies which vessels were transmitting their identity and allows cross-referencing with the radar and optical images. A hyperspectral pass over the same area reveals the chemical composition of surface residues that might indicate what cargo a vessel recently carried.

Each of those data products individually is valuable. Combined and co-registered against a common spatial reference framework, they form a multi-spectral, multi-modal understanding of the scene that approaches what a team of specialized analysts would struggle to compile manually. The processing module’s role in sensor fusion is to maintain a common time and position reference for all sensor modules, so that data collected by each module can be co-registered at the pixel level. It must also manage the sequencing of observations to ensure that all sensor types collect data over the same area within a timeframe short enough that scene conditions have not changed significantly between collections.

Data fusion at the processing module level has been made more capable by recent advances in onboard artificial intelligence inference. Neural network models optimized for satellite hardware can run change detection, object classification, and anomaly identification directly on the satellite, generating derived intelligence products rather than raw data. ICEYE has been developing AI-based techniques for onboard processing of SAR data precisely because running analysis in space reduces the volume of data that needs to be transmitted to ground stations while delivering actionable results faster. In a disaggregated architecture, that onboard AI inference runs on the central processing module’s computing resources, with access to all sensor data streams, enabling cross-modal analysis that a SAR-dedicated or EO-dedicated spacecraft could never perform.

The question of when fusion happens is architecturally significant and not fully resolved as of 2026. Fusing data before downlink, onboard the processing module, saves downlink bandwidth and reduces latency in delivering actionable products. However, it limits what analysts can do with the raw data after the fact. Onboard algorithms may discard information that later proves relevant. Fusing data after downlink, in a cloud-based ground analytics platform, preserves all raw data but requires substantial downlink bandwidth and introduces latency. Most operational systems today use a hybrid approach: onboard processing produces quick-look products and priority alerts, while full raw data is queued for downlink and reprocessed on the ground when bandwidth allows.

The Disaggregated Architecture and National Defense Requirements

The strategic appeal of disaggregated satellite architectures for defense applications became much clearer after the Space Development Agency (SDA) was established by the US Department of Defense in 2019. The SDA’s mission was to replace high-value monolithic intelligence satellites, which represent single points of failure in any conflict scenario involving counter-space capabilities, with a proliferated architecture of smaller, cheaper, and more numerous spacecraft. If an adversary destroys one or two nodes in a constellation of hundreds, the overall capability degradation is minimal. If an adversary destroys a single billion-dollar monolithic satellite, an irreplaceable national intelligence asset is lost.

The SDA’s Proliferated Warfighter Space Architecture (PWSA) implements this logic at the constellation level. The PWSA separates communication functions into a Transport Layer of 300 to more than 500 satellites and sensing functions into a Tracking Layer equipped with infrared sensors for missile detection and tracking. On December 19, 2025, the SDA awarded four agreements worth approximately $3.5 billion to teams led by Lockheed Martin, Rocket Lab, Northrop Grumman, and L3Harris Technologies to build 72 Tracking Layer Tranche 3 satellites for launch in fiscal year 2029. Each vendor’s contract covered 18 satellites carrying infrared sensors for missile warning and tracking, with some satellites also carrying sensors for missile defense fire control.

While the PWSA separates communication and sensing layers across different satellites rather than implementing a single processing module with multiple sensor payloads, the underlying logic is architecturally related to full disaggregation. Both approaches reject the premise that a single large spacecraft must carry all mission functions. The PWSA demonstrates at operational scale that a constellation of smaller specialized nodes can deliver capabilities that previously required much larger and more expensive spacecraft. The disaggregated single-spacecraft model extends this principle one level deeper, separating sensing, processing, and communications functions within a single platform.

Johns Hopkins University Applied Physics Laboratory released Revision 6.1 of the Modular Payload Design Standard in October 2025, developed in collaboration with government and industry partners. The standard defines how electronic warfare, signals intelligence, and communications payloads should be designed for modular integration across unmanned and dismounted platforms. While directed primarily at airborne platforms, the standard’s core requirements for modular interfaces, standardized power and data connectors, and plug-and-play integration protocols are directly applicable to satellite systems and represent the kind of published standard that could accelerate disaggregated satellite payload development if applied more broadly.

The survivability argument for disaggregation goes beyond simply reducing the cost of each individual node. Disaggregated systems can be designed to degrade gracefully. If a sensor module fails or is disabled, the processing module continues operating with the remaining sensors. The processing module itself can be made more survivable through redundancy within its own architecture, carrying duplicate processing elements that activate if the primary elements fail. The modular design also allows rapid reconstitution: a replacement sensor module can be launched on a small vehicle, rendezvous with the processing module in orbit, and restore full capability without replacing the entire spacecraft.

Rollout Sequencing: Phasing Capabilities Across Module Generations

One of the most operationally significant advantages of the disaggregated model is the ability to phase capabilities across time rather than committing to a fixed sensor complement at the moment of launch. A processing and control module can be launched first, establishing the orbital position, communication links, and computing infrastructure needed for the full system. Sensor modules are then added sequentially as they become available, allowing the system to begin delivering value with partial capability while the remaining modules are manufactured and tested.

This phased approach mirrors the SDA’s spiral development model, which delivers a minimum viable product on a two-year cycle and adds capabilities with each subsequent tranche. The SDA stated explicitly that resilience is built through proliferation, with refreshed capabilities delivered approximately every two years. In a disaggregated single-spacecraft context, that cycle applies at the module level. An EO module launched with the processing hub in year one can be joined by a SAR module in year two, a hyperspectral module in year three, and an upgraded EO module with higher resolution in year four. The processing hub remains in place throughout, accumulating capability as new modules are integrated.

The operational challenge of phased rollout is interface compatibility across generations. If the processing module launched in year one defines a specific electrical connector standard, power envelope, and data protocol, then every sensor module launched in subsequent years must conform to that standard. A module developed two years after launch cannot change the connector pinout or the power delivery specification without requiring a hardware modification to the processing module as well. This means the interface standard must be defined conservatively enough to accommodate sensor modules not yet designed at the time the processing module is built.

Managing interface compatibility across module generations has no completely satisfying solution. The DARPA F6 program addressed it through the F6 Developer’s Kit, which published the complete interface specification as an open standard so that any developer could build conformant modules. The SDA Transport Layer addressed it by standardizing on optical inter-satellite link specifications so that satellites from different manufacturers can communicate. The Aerospace Corporation’s Slingshot Handle interface addressed it at the small satellite level by separating the mechanical and electrical connector standard from the data protocol, allowing payload developers to use the same physical connector regardless of the protocol running over it.

Sensor technology evolves faster than orbital hardware. A hyperspectral sensor module designed in 2026 will be superseded by a more capable design by 2030, but the processing module launched in 2026 may have a design life of seven to 10 years. The processing module’s computing architecture must be designed with enough performance headroom to handle the data rates and processing requirements of sensor modules that do not yet exist. That design requirement creates uncertainty and tends to push processing module designs toward higher-than-necessary capability at launch to provide that headroom, which increases cost and mass. Whether the cost of over-engineering the processing module is justified by the operational benefits of later sensor upgrades is a question that reasonable engineers disagree on, and it probably depends on how aggressively the mission requirements are expected to evolve.

Data Architecture: From Raw Sensor Data to Ground Users

Raw sensor data has limited operational value. An analyst who receives a multi-gigabyte hyperspectral data cube covering a 40-kilometre swath has, in principle, everything needed to identify every surface material in the scene. But extracting that identification requires calibration processing to remove atmospheric effects, geometric correction to register the image to a map coordinate system, spectral unmixing to separate mixed-material pixels, and classification algorithms to assign material labels. Each of these processing steps requires both computational resources and reference data that is not available onboard the satellite.

The processing module handles the first stages of the data pipeline: raw data buffering, compression, and formatting for downlink. After the data reaches the ground, the pipeline continues through ground segment infrastructure. Ground processing chains for disaggregated multi-sensor systems are more complex than those for single-sensor satellites because they must handle diverse data types, apply different processing algorithms for each sensor type, and eventually merge products from different sensors into fused analysis outputs.

Commercial cloud platforms have transformed this ground processing architecture. Data analytics companies like Planet Labs and Maxar now deliver processed and analysis-ready data products through API interfaces that allow software systems to query satellite imagery the same way they query a database. A disaggregated satellite system feeding data to a similar cloud platform would allow users to access not just raw imagery from individual sensor modules but pre-fused products that combine SAR, EO, and hyperspectral observations of the same area across multiple collection times.

Downlink capacity is a persistent constraint. A typical low Earth orbit satellite has roughly five to ten minutes of ground station contact per orbit pass, and contact is only possible when the satellite is above the horizon at a ground station location. During that contact window, the processing module transmits queued data at the highest rate the link supports. A high-frequency X-band downlink can transmit several hundred megabits per second, but a multi-sensor disaggregated system collecting data across multiple sensor modules can generate several gigabits of data per orbit pass. The gap between collection rate and downlink rate creates a data backlog that the processing module must manage through intelligent prioritization.

The use of inter-satellite optical links, an approach the SDA has demonstrated operationally with its Transport Layer satellites, provides a path to higher effective downlink capacity by routing data through a relay constellation to multiple ground stations simultaneously. An SDA Transport Layer node in view of a ground station serves as a relay for other satellites not in ground contact, extending the downlink window and increasing aggregate data throughput. A disaggregated multi-sensor satellite connected to this kind of relay mesh could offload data continuously during an orbit rather than waiting for direct ground contact.

Thermal Management Across Dissimilar Sensor Modules

The thermal engineering of a disaggregated satellite system is significantly more complex than that of a single-purpose platform. Different sensor modules have radically different thermal requirements. A cryogenically cooled infrared focal plane array must be maintained below 150 Kelvin to operate at its specified sensitivity. The SAR antenna’s transmit-receive modules generate substantial heat during radar pulse transmission and must be kept cool enough to prevent performance degradation. The processing module’s computing hardware generates continuous heat that must be dissipated to space. All of these heat sources and sinks exist within a shared mechanical structure exposed to the same orbital thermal environment.

In a disaggregated system with physically connected modules, thermal interfaces between modules require careful design. Heat generated in one module can conduct through structural connections into adjacent modules. A SAR module conducting heat into the processing module’s enclosure could push the processing module’s electronics outside their rated temperature range. The solution involves thermal isolation joints that minimize conduction across module interfaces, combined with thermal management systems within each module that route waste heat to dedicated radiators.

Passive radiators, which are panels coated with high-emissivity finishes and oriented away from the sun and Earth, are the primary means of rejecting waste heat in space. In a disaggregated architecture, each module may have its own dedicated radiator panel sized for that module’s peak heat dissipation requirement. The processing module’s radiator handles the continuous heat load from computing hardware. The SAR module’s radiator handles peak heat from transmission events. The hyperspectral module’s radiator must be designed to maintain the focal plane array at operating temperature without contaminating the optical path.

Cybersecurity and Encryption in a Multi-Module Architecture

A disaggregated satellite system’s multi-module architecture creates cybersecurity challenges that do not exist in a monolithic spacecraft. Every inter-module interface is a potential entry point for unauthorized access or interference. If one sensor module in a cluster is compromised, an attacker who can inject commands into that module’s data stream might be able to influence the processing module’s behavior or extract intelligence data before it is encrypted for downlink.

The F6 program addressed this through multi-level security architecture requirements that treated the inter-module network as an untrusted transport layer. Encryption was applied end-to-end between the command source and the target module, so that the network itself could not read or modify the command content. Authentication protocols verified that commands came from authorized sources before execution. Data transmitted from sensor modules to the processing module was similarly encrypted and authenticated, preventing injection of false data into the processing pipeline.

Ground-to-space command links add another layer of complexity. The processing module serves as the single command uplink destination for the entire cluster, receiving and authenticating commands before distributing them to sensor modules. This centralizes security but also creates a single point where command encryption must be maintained. The US Space Force’s standards for protected tactical waveform communications, applied to ground-to-satellite links, provide a baseline for how command and control security should be architected in operational systems. Compliance with those standards also requires that each sensor module’s firmware be authenticated before execution, preventing an adversary from uploading malicious code through the inter-module network even if they have access to the physical RF uplink frequency.

The Aerospace Corporation’s Slingshot Approach as a Near-Term Model

Among the most instructive recent examples of plug-and-play modular architecture in practice is the Aerospace Corporation’s Slingshot 1 mission. The satellite, a 12U CubeSat launched in 2023, carried 19 payloads on a single bus using the Handle interface standard. Payloads on Slingshot 1 included Vertigo, an attitude control system that could enable the satellite to point at Earth targets; Blinker, a GPS transponder for space traffic management; Hyper, a hydrogen peroxide thruster; and LaserComm, a next-generation space-to-ground laser communication downlink system. Each payload conformed to the Handle electrical interface and could be integrated into the satellite bus without modification to the bus hardware.

Hannah Weiher, Engineering Manager in the Aerospace Corporation’s iLab and program manager for Slingshot, said “Customization has traditionally played a major role in payload development and turnaround time. With Slingshot, payloads conform to a basic standard in which they can plug in and work even if they’re late in the development timeline.” That statement captures the core value proposition of the disaggregated approach more precisely than any systems architecture diagram. A late-arriving payload that meets the interface standard simply plugs in. A monolithic design would require rework at the system level.

The Slingshot model shows where the near-term implementation of disaggregated multi-sensor satellite architectures is most practical. Small satellite programs operating on commercial schedules with multiple payload sponsors are natural candidates for modular architectures because the benefits of late payload integration, independent payload development, and mixed payload types directly address the coordination problems that multi-stakeholder programs encounter. A program with one government agency providing an EO module, a second providing a hyperspectral module, and a commercial company providing a SAR module would benefit enormously from a standardized interface that allows each organization to deliver its module independently.

Data Processing Hardware Trends Supporting the Architecture

The viability of the central processing module concept depends on the availability of computing hardware capable of handling multiple high-data-rate sensor streams simultaneously while operating within the power and radiation constraints of the space environment. For most of satellite history, spaceborne computing capability lagged commercial terrestrial capability by a decade or more because of the long development cycles required to produce radiation-hardened processors.

That gap has narrowed significantly. Field-programmable gate arrays from Xilinx (now AMD), Microsemi (now Microchip), and European Cooperation for Space Standardization-compliant processors provide substantial compute capacity in radiation-tolerant or radiation-hardened versions. The European Space Agency’s PATTERN project, conducted in 2024 with ESA funding and Frontgrade Gaisler as an industry partner, extended AI inference support to space-qualified processors including LEON4, LEON5, NOEL-V RISC-V implementations, and Microchip’s PolarFire FPGA family. The project also established a path toward compliance with ECSS software standards ECSS-E-ST-40 and ECSS-Q-ST-80, meaning that AI inference software running on a central processing module can now plausibly claim compliance with the same standards applied to other flight software.

Neuromorphic and low-power AI accelerator chips are entering space qualification pipelines as well. These chips, designed to perform neural network inference at very low power consumption, are well suited to the always-on analysis tasks that a central processing module would need to perform: scanning AIS data streams for anomalous vessel behavior, monitoring SAR imagery for changes in facility activity, or running hyperspectral classification models on incoming sensor data. Running this inference onboard rather than on the ground reduces the latency between collection and actionable output from hours to minutes in most scenarios.

Mass storage for the processing module is evolving alongside computing hardware. High-density NAND flash storage qualified for the space radiation environment now supports capacities in the terabyte range, sufficient to buffer several orbit passes of multi-sensor data before a downlink opportunity. Combined with intelligent data management software that prioritizes which data to downlink first based on predicted analyst interest, a processing module with terabyte-scale storage can handle the data volume generated by even a heavily loaded multi-sensor cluster.

Limitations and the Uncertainties That Still Exist

Any honest assessment of disaggregated multi-sensor satellite architectures has to acknowledge that the concept has been more thoroughly studied than deployed. DARPA’s System F6, the most ambitious attempt to demonstrate fractionated satellite operation in orbit, was cancelled before any hardware flew. The fractionated spacecraft concept was coined in academic papers in 2006, and nearly two decades later, no operational satellite system separates processing and sensing into swappable modular elements at the level of sophistication the concept envisions.

What does exist is a set of converging trends: the SDA’s proliferated architecture demonstrating that disaggregated multi-node networks deliver real-world capability; the Aerospace Corporation’s Slingshot mission demonstrating that modular interfaces work in practice; commercial companies like Pixxel, ICEYE, and Capella Space demonstrating that specialized sensor modules can be economically viable when paired with capable platforms. Whether these trends converge into a true plug-and-play disaggregated multi-sensor satellite within the next five years remains an open question. It depends on whether the interface standardization problem gets solved at an industry-wide level, whether launch economics make replacing individual sensor modules cheaper than replacing whole satellites, and whether defense customers with the resources to fund multi-sensor cluster development decide to commit to the architecture.

The data fusion problem adds another layer of uncertainty. Fusing data from three or four different sensor types into a single coherent intelligence product requires not just the right hardware but the right algorithms, reference databases, and validation frameworks. Building those capabilities to the reliability standard that defense and intelligence customers require is a long engineering program in its own right, independent of the spacecraft architecture choices that precede it. The National Geospatial-Intelligence Agency, which is the primary US government customer for fused multi-source imagery intelligence, sets demanding standards for co-registration accuracy, spectral calibration, and product delivery timelines that any disaggregated multi-sensor system would have to meet before its products could be integrated into operational workflows.

Formation Flying, Proximity Operations, and Cluster Geometry

The physical arrangement of modules within a disaggregated satellite cluster has a significant effect on both operational capability and system complexity. In a physically docked architecture, all modules are rigidly attached to the processing hub and move together as a single spacecraft. This arrangement is mechanically simpler, eliminates the need for formation maintenance propulsion in the sensor modules, and provides the lowest-latency data interface between modules through direct electrical connections. The tradeoff is that docked modules share a single attitude, meaning the SAR antenna, the EO telescope, and the hyperspectral spectrometer must all point in whatever direction the unified spacecraft points. If the SAR module requires the spacecraft to fly at a specific squint angle to the ground while the EO module works best in nadir-pointing mode, a compromise must be found that serves neither sensor optimally.

Loose cluster or formation flying architectures address this constraint by allowing sensor modules to fly as separate spacecraft in coordinated orbits, sharing data through inter-satellite links rather than physical connectors. Each module can be independently pointed to optimize its own collection geometry. A SAR module can maintain a right-looking squint attitude while an EO module in the same cluster points precisely nadir. This flexibility comes at the cost of formation maintenance. Each module requires a propulsion system to hold its position relative to the processing hub against the perturbing forces that gradually disperse formation elements, including atmospheric drag at low altitudes and differential gravitational gradients. Managing the propellant budget of multiple formation members adds mission planning complexity and eventually limits mission lifetime.

Semi-rigid docking architectures offer a middle path. The processing hub provides a mechanical attachment port and an electrical interface bus, but sensor modules are not fully rigidly connected. They attach through compliant mounts that allow small angular offsets between the hub and the module. This allows limited independent pointing of individual modules without the full propulsion complexity of loose formation flying, at the cost of more complex mechanical interface design. Whether the operational benefit justifies the engineering investment is a question that mission architects have approached differently depending on mission requirements and budget constraints.

The geometry of a physically docked cluster also affects the electromagnetic environment inside the system. A SAR module transmitting high-power microwave pulses can interfere with an EO module’s electronics if they are mounted in close proximity without adequate shielding. A hyperspectral spectrometer’s detector can be contaminated by light scattered from adjacent spacecraft structures if the physical layout does not provide adequate baffling. In a modular system where different sensor combinations may be assembled depending on mission requirements, the electromagnetic compatibility analysis must account for all possible module combinations rather than a single fixed configuration.

Commercial Earth Observation Markets and the Value of Multi-Sensor Platforms

Commercial demand for satellite imagery has grown substantially across multiple sectors over the past several years. Agriculture technology companies need regular multispectral and hyperspectral coverage to monitor crop health, estimate yields, and detect early signs of disease or water stress. Energy companies need SAR data to monitor infrastructure integrity, track surface deformation near pipelines and storage facilities, and detect hydrocarbon seeps. Financial intelligence services need persistent EO coverage of economic indicators including port activity, parking lot occupancy at retail facilities, and construction site progress. Environmental monitoring agencies need hyperspectral data to track deforestation, measure carbon stock changes, and identify pollution sources.

Each of these customer segments has traditionally purchased data from single-sensor satellites optimized for their specific application. Agriculture customers buy hyperspectral or multispectral data. Energy infrastructure customers buy SAR data. Financial intelligence customers buy high-resolution EO data. A disaggregated multi-sensor satellite platform that collects all these data types during the same orbit pass over the same geographic area could serve all of these customer segments simultaneously from a single platform, reducing the cost per bit of collected data and increasing revenue per orbit pass.

The economic logic is attractive but faces execution challenges. Each sensor type serves customers who have developed workflows, data formats, and analysis tools specific to that sensor. A SAR customer’s processing pipeline is different from a hyperspectral customer’s pipeline. Delivering multi-sensor data in a format that integrates smoothly with existing customer workflows requires either that the satellite operator builds separate data products for each customer type or that customers adapt their analysis tools to a new multi-sensor data format. Neither option is trivial.

Planet Labs demonstrated one approach to this problem by deploying a constellation of EO satellites with a consistent simple interface and a developer-friendly API that allowed customers to integrate Planet data into their own analysis platforms. The company’s SkySat satellites, which provide sub-metre resolution imagery, and its Dove satellites, which provide daily global coverage at lower resolution, serve different segments of the imagery market from a common operational platform. The disaggregated model could extend this approach by allowing the same ground infrastructure to serve SAR, hyperspectral, and EO customers simultaneously from clusters where the satellite bus is standardized but the sensor configuration varies.

Satellogic, the Argentine-American Earth observation company, built a constellation that integrates hyperspectral and multispectral capabilities on the same satellites, offering both data types from a unified platform. The company’s Aleph-1 satellites carry both a 1-metre multispectral imager and a 30-metre hyperspectral imager, allowing customers to receive spectrally rich hyperspectral data that is registered against higher-resolution context imagery from the same pass. This design implements a rudimentary form of multi-sensor integration on a small satellite and illustrates the commercial appeal of combining complementary sensor types without requiring separate satellite purchases.

Regulatory and Licensing Considerations for Multi-Sensor Disaggregated Systems

Operating a satellite that carries multiple sensor types introduces regulatory complexity that a single-sensor satellite does not face. Commercial remote sensing satellites in the United States must be licensed by the National Oceanic and Atmospheric Administration (NOAA) under 15 CFR Part 960. The license covers the specific sensors, orbital parameters, and operational conditions for the licensed system. A disaggregated system that can change its sensor configuration by adding or removing modules raises the question of whether each possible configuration requires a separate license or whether a single license can cover the range of configurations the processing hub might support.

NOAA’s commercial remote sensing regulations, as revised in 2020, take a more flexible approach to licensing than the previous regime, allowing broader parameters for constellation operations and reducing the granularity of per-satellite licensing. However, specific provisions regarding modular or reconfigurable systems that can change their sensor payloads mid-mission had not been addressed directly as of early 2026. Operators planning to deploy disaggregated multi-sensor systems will need to work with NOAA’s licensing office to establish whether their architecture falls within existing license parameters or requires new regulatory guidance.

SAR-specific export control regulations add another dimension of complexity. SAR systems capable of certain performance thresholds in resolution and power are subject to International Traffic in Arms Regulations (ITAR) controls and may require State Department licensing for international sales or partnerships. A disaggregated satellite system where a US-built processing hub hosts a SAR module from a foreign manufacturer, or vice versa, involves a technical data exchange that may trigger ITAR requirements regardless of whether the physical hardware crosses national boundaries.

Hyperspectral systems with sufficiently fine spectral resolution in certain wavelength ranges are subject to export controls because of their potential applications in identifying materials of military or intelligence significance. A processing hub designed to interface with both hyperspectral and signals intelligence modules, even if the modules are not simultaneously present in the licensed configuration, may require the operator to secure licenses that account for the full range of potential module combinations.

Summary

Disaggregated satellite systems built around a central processing and control module represent a significant departure from how spacecraft have been designed and operated since the beginning of the space age. The processing module handles the functions that all sensor modules share: attitude control, power management, data routing, communications, and onboard computing. Sensor modules, whether electro-optical imagers, synthetic aperture radar payloads, hyperspectral spectrometers, or signals collection systems, attach to that core through standardized interfaces and operate as specialized nodes in a coordinated cluster.

The architectural logic is compelling. Sensor technology evolves faster than spacecraft buses. Modular designs allow the sensor complement to be updated without replacing the entire spacecraft. Multiple sensor types operating from a single processing hub can deliver fused intelligence products that no single-sensor satellite could produce. Defense applications benefit from the survivability and reconstitution advantages that distributed, replaceable modules offer. Commercial applications benefit from reduced development timelines and the ability to combine sensor types that serve different customer markets on a single platform.

Programs like the SDA’s PWSA, the Aerospace Corporation’s Slingshot, and DARPA’s historical System F6 work have advanced the enabling technologies considerably. Standardized electrical interfaces, modular payload design standards, onboard AI inference, high-density space-qualified storage, and optical inter-satellite links are all now sufficiently mature to support a next-generation disaggregated multi-sensor system. What remains is the organizational and programmatic commitment to build one, verify the interface standards through extensive ground testing, and fly the result at a scale sufficient to demonstrate the operational value that decades of analysis have predicted.

Commercial hyperspectral constellations like Pixxel’s Firefly network and multi-modal intelligence constellations like those operated under contract to the NRO are converging toward the point where a single processing hub supporting diverse swappable sensors makes more financial and operational sense than building separate satellites for each sensor type. The disaggregated model does not eliminate the engineering challenges. It relocates them, concentrating complexity in the interface standards and processing software rather than in bespoke integrated spacecraft designs. That relocation may not make the overall system simpler, but it does make individual modules cheaper to build, faster to replace, and more responsive to changes in mission requirements over time.

Appendix: Top 10 Questions Answered in This Article

What is a disaggregated satellite system?

A disaggregated satellite system distributes the functions of a traditional monolithic spacecraft across multiple modules that communicate through standardized interfaces. Rather than integrating all sensors, computing hardware, communications, and power management into a single structure, a disaggregated system assigns these functions to separate modules that operate together as a coordinated cluster. The architecture allows individual modules to be upgraded or replaced independently without rebuilding the entire spacecraft.

What does a processing and control module do in a disaggregated satellite?

The processing and control module serves as the central hub that manages all other modules in the cluster. It handles attitude determination and control for the entire system, manages power allocation across all sensor payloads, ingests and buffers data from each sensor module, applies compression and formatting for downlink, and routes commands from ground operators to specific modules. It also hosts onboard AI inference and data fusion capabilities that draw on the computing resources needed to process multiple high-data-rate sensor streams simultaneously.

What is the difference between electro-optical and synthetic aperture radar satellite modules?

An electro-optical module uses an optical telescope and focal plane array detector to capture visible and infrared light reflected from the Earth’s surface, producing photographic-style images that depend on daylight and clear weather conditions. A synthetic aperture radar module transmits microwave pulses and records the reflected signals, processing them into images using the spacecraft’s orbital motion to synthesize a large virtual antenna aperture. SAR can operate through clouds, rain, smoke, and darkness, making it complementary to electro-optical imaging.

Why do hyperspectral imaging modules produce such large volumes of data?

A hyperspectral module captures imagery across hundreds of narrow spectral bands simultaneously, often more than 100 bands each less than 10 nanometres wide. For every pixel in the image, the sensor records a complete spectral profile rather than just a few broad color values. A 300-band hyperspectral imager operating at 5-metre resolution over a 40-kilometre swath can generate data at rates exceeding 5 gigabits per second during collection, far exceeding the downlink capacity of most small satellite systems and requiring onboard compression or intelligent data selection before transmission.

What was DARPA’s System F6 program, and what did it demonstrate?

System F6, which stood for Future, Fast, Flexible, Fractionated, Free-Flying Spacecraft United by Information Exchange, was a DARPA program initiated in 2007 to develop and demonstrate fractionated satellite architecture. The program created the F6 Developer’s Kit, an open-source set of interface standards allowing any organization to build compatible modules, and the F6 Technology Package, hardware implementing wireless connectivity and encryption for cluster participation. DARPA cancelled the planned flight demonstration in May 2013 before hardware reached orbit, but the program’s interface standards and technical work informed subsequent modular satellite development.

How does sensor fusion work in a disaggregated multi-sensor satellite?

Sensor fusion combines data products from multiple sensor modules, such as electro-optical images, synthetic aperture radar imagery, hyperspectral spectral maps, and AIS vessel position data, into analysis products more informative than any single sensor could produce. The processing module maintains a common time and position reference for all sensor modules so that data from each can be co-registered to a common spatial coordinate system. Onboard AI inference algorithms then apply change detection, object classification, and anomaly identification across the combined data streams, producing derived intelligence products before the raw data is downlinked.

What role does the Space Development Agency play in advancing disaggregated satellite concepts?

The Space Development Agency builds and operates the Proliferated Warfighter Space Architecture, a constellation of hundreds of small satellites in low Earth orbit separated into a Transport Layer for communications and a Tracking Layer for missile warning and tracking. The PWSA implements disaggregation at the constellation level by separating communication and sensing functions across different satellites rather than combining them in one platform. In December 2025, the SDA awarded approximately $3.5 billion across contracts to Lockheed Martin, Rocket Lab, Northrop Grumman, and L3Harris for 72 Tracking Layer Tranche 3 satellites, demonstrating the architecture’s operational commitment.

What is plug-and-play satellite technology?

Plug-and-play satellite technology refers to standardized electrical and mechanical interfaces that allow a sensor module or payload to be integrated into a satellite bus without requiring customization of the bus hardware. The Aerospace Corporation’s Handle interface module, demonstrated on the Slingshot 1 mission, enables payloads to communicate with the host spacecraft bus and other payloads regardless of the bus manufacturer’s design specifics. Plug-and-play standards allow late manifest changes, reduce integration timelines, and make it economically viable to deploy diverse payload types from a common processing platform.

How does a SAR module’s power demand affect a disaggregated satellite’s design?

A synthetic aperture radar payload can consume 200 to 500 watts of RF output power during active transmission, with total system power substantially higher. The central processing module’s power management system must reserve sufficient battery capacity before a SAR collection event, restrict other power-consuming activities during radar transmission to stay within the total power budget, and schedule thermal dissipation time after collection ends before commanding another SAR pass. This power sequencing constraint means the processing module’s power management software must treat SAR collection as a time-limited, high-priority resource allocation event rather than a continuous background operation.

Can sensor modules in a disaggregated system be replaced or upgraded in orbit?

Current disaggregated satellite systems do not support routine in-orbit module exchange, but this capability is a recognized goal of several technology development programs. On-orbit servicing missions, such as those developed by Northrop Grumman’s Mission Extension Vehicle program and explored by the European Space Agency’s RISE program, demonstrate docking and mechanical manipulation in orbit. A disaggregated satellite system designed from the outset for module exchange would require standardized docking interfaces on the processing module, a capable servicing spacecraft to perform the exchange, and qualification of the replacement module for the mechanical and electrical interface standard. The DARPA Robotic Servicing of Geosynchronous Satellites (RSGS) program and commercial servicing initiatives by companies including Astroscale are steadily advancing the rendezvous, proximity operations, and mechanical attachment technologies that modular in-orbit servicing would require.