Satellite Collision Avoidance Tech 2026 — PatSnap Eureka
Satellite Constellation Collision Avoidance Technology Landscape 2026
LEO object counts exceeded 10,000 active satellites as of 2023, with projected growth to over 100,000 within a decade. This report maps the patent and literature landscape across space traffic management, onboard autonomous avoidance, AI-driven detection, and multi-operator coordination frameworks — spanning 60+ records from 2010 to 2026.
A Multi-Layered Technical Stack for Collision Avoidance
Satellite constellation collision avoidance technology encompasses four functional layers: (1) space situational awareness (SSA) data collection and cataloguing, (2) conjunction detection and collision probability calculation, (3) avoidance maneuver planning and optimization, and (4) multi-operator coordination and governance. The dominant focus across retrieved results is on mega-constellation environments of 100 or more satellites per operator, where the density of orbital objects fundamentally changes the calculus of traditional single-satellite avoidance.
The field bifurcates along a key architectural axis: ground-centralized systems that process SSA data and issue avoidance commands via uplink, versus onboard autonomous systems that execute collision avoidance logic directly on the satellite. A third axis — multi-operator coordination platforms — is rapidly emerging as constellations from competing operators now occupy overlapping orbital shells. Authoritative context from ESA, NASA, and ITU confirms that coordinated STM frameworks are a growing regulatory priority.
Studies modeling Starlink’s impact report that the lifetimes of target satellites can be shortened by up to 99.82% without early avoidance control, and that joining Starlink increased total conjunction risk significantly across all modeled scenarios. These quantitative findings underscore the systemic nature of the challenge now confronting the space industry. The PatSnap Analytics platform enables teams to monitor this IP landscape continuously.
From Foundational Autonomy to AI-Driven Edge Computing
The earliest relevant filing dates to 2013; a decisive acceleration cluster emerged in 2019–2022, coinciding with the first Starlink launches and the ESA–Starlink near-collision event.
Filing Activity by Period (2010–2026)
Active filings accelerated sharply after 2019 Starlink launches; 8 patents carry 2025–2026 publication dates, confirming the field remains in active development.
Technology Cluster Distribution
Ground-based STM with multi-operator coordination is the dominant cluster by filing volume; automated inter-operator agreement is the fastest-growing emerging cluster.
Four Clusters Defining the Collision Avoidance IP Landscape
From ground-centralized STM platforms to AI-enabled onboard autonomy, these clusters represent the core battlegrounds for IP and R&D investment through 2026.
Ground-Based STM with Multi-Operator Coordination
A central STM platform aggregates orbital data from SSA providers, mega-constellation operators, debris removal businesses, and rocket launch operators. The platform computes conjunction events, issues intrusion alerts, and coordinates avoidance actions through a shared Open Architecture Data Repository (OADR). A defining feature is structured role separation between SSA business devices, collision avoidance assistance business devices, and mega-constellation business devices, each with defined data sharing obligations and privacy-preserving secure communication channels. PatSnap Analytics can map this portfolio in depth.
Dominant cluster by filing volume — Mitsubishi Electric (18 patents)Onboard Autonomous Collision Avoidance
Satellites carry onboard sensors (radar, lidar, star/earth/sun sensors) and processors capable of independently calculating debris collision trajectories, computing minimum delta-V maneuvers, and firing onboard thrusters — without ground uplink. This architecture addresses the latency limitation of ground-centralized systems, where single-pass communication windows and multi-second uplink delays are incompatible with emergency avoidance timescales. Recent filings extend autonomy to neural-network and deep reinforcement learning implementations deployable on resource-constrained onboard computers. Foundational IP held by Robert Briskman spans CA, WO, EP, and IN jurisdictions.
Foundational IP: Briskman 2013 WO — no corporate assigneeConjunction Probability Calculation & Fast Screening
Mega-constellations involving thousands of satellites make brute-force pairwise orbit propagation computationally prohibitive. This cluster addresses rapid screening of which satellite pairs require detailed collision probability calculation. Approaches include relative orbital element-based filtering, improved spatial flux methods, apogee/perigee altitude pre-screening, GPU-parallelized SGP4 propagation, and second-order cone programming (SOCP) optimization for avoidance trajectory planning. Beijing Institute of Technology leads this sub-domain with 3 CN patents filed between 2022 and 2024. See also PatSnap’s computational analysis solutions.
Chinese institutions: 12 CN filings across 8 assigneesAutomated Multi-Operator Conjunction Agreement
When two maneuverable satellites from different operators face a conjunction, an automated platform generates, negotiates, and executes a bilateral “operative agreement” specifying which satellite maneuvers and by how much — removing human-in-the-loop delays. The system ingests Conjunction Data Messages (CDMs), checks maneuverability status of both spacecraft, identifies whether a pre-existing inter-operator agreement exists, and if not, auto-generates one. This directly addresses the gap where CDMs delivered via email introduce unacceptable response latency. Kayhan Space Corp. holds the key IP in this sub-domain with 3 US patents filed 2023–2026.
Kayhan Space: response time reduced from hours to secondsFrom Mega-Constellation LEO to Astronomical Observatory Protection
The technology stack serves five distinct application domains, ranging from commercial mega-constellation operations to civil science telescope scheduling.
Mitsubishi Electric Dominant; China Building Competing Portfolio
| Assignee | Country | Records (Dataset) | Jurisdictions | Focus Area |
|---|---|---|---|---|
| Mitsubishi Electric Corporation | Japan | 18+ patents | US, EP, JP, WO | Full STM stack: SSA, OADR, mega-constellation, debris removal, jamming avoidance |
| Beijing Institute of Technology | China | 3 CN patents | CN | Fast screening algorithms, SOCP trajectory optimization (2022–2024) |
| Beihang University | China | 2 CN patents | CN | Neural-network closed-loop onboard avoidance (2025) |
| Kayhan Space Corp. | United States | 3 US patents | US | Automated bilateral conjunction agreement & CDM processing (2023–2026) |
Five Signals Defining the 2025–2026 Frontier
Based on patents and literature with publication dates of 2024–2026, five clear emerging directions are identifiable in this dataset.
Satellite Edge Computing for Onboard Real-Time Avoidance
The National University of Defense Technology (2025, CN) proposes migrating collision avoidance computation from ground stations to satellite edge computing networks, citing that ground-centralized systems suffer from MB-scale data transfer latency and inability to maintain continuous communication with high-dynamic LEO satellites. This represents a fundamental architectural shift.
AI and Neural Networks on Onboard Computers
Two 2025 Beihang University patents describe neural-network-based closed-loop collision avoidance deployable on computationally constrained onboard processors, eliminating ground-uplink dependency. Haney’s 2025 US patent similarly employs deep reinforcement learning for onboard trajectory control. Both represent a convergence of AI capability with satellite hardware constraints.
Automated Inter-Operator Agreement Execution
Kayhan Space’s 2026 filing extends their automated conjunction assessment to include pre-negotiated operative agreements based on satellite characteristics and operator preferences, reducing human response time from hours to seconds. This directly addresses the operational inadequacy of CDM-via-email coordination as multi-operator conjunctions multiply.
IP Strategy Signals for R&D Teams and New Entrants
Mitsubishi Electric holds a structurally dominant IP position across the STM architecture stack, with sustained filings from 2020 to 2026 covering nearly every functional component — SSA, OADR, mega-constellation devices, debris removal coordination, and jamming avoidance. Any new entrant building an STM platform must either design around this portfolio or engage in licensing negotiations. The PatSnap customer success team has supported IP due diligence for similar landscape assessments.
The ground-to-onboard autonomy transition is a defining battleground. The field is converging on a hybrid architecture: ground systems for strategic conjunction planning, onboard AI for tactical emergency response. R&D teams should prioritize neural-network-based onboard avoidance for scenarios where uplink latency is prohibitive — a gap currently exploited by Beihang University and Haney.
Chinese institutions are building a competing, domestically-anchored IP portfolio. With at least 12 CN-jurisdiction filings across 8 assignees in this dataset, China is not relying on Western STM frameworks. R&D teams and IP strategists should monitor CN filings closely, particularly in fast-screening algorithms and edge-computing architectures, which represent genuine algorithmic advances. Resources from WIPO and the ITU provide complementary regulatory context. The PatSnap Analytics platform enables continuous CN filing monitoring.
Debris removal coordination is an underprotected IP zone relative to its operational importance. Only a handful of patents in this dataset directly address the ADR–constellation interoperability problem. The commercial urgency of active debris removal, combined with thin IP coverage, represents a white-space opportunity for new entrants in debris removal coordination platforms and real-time orbital data sharing protocols. See also PatSnap Open API for programmatic landscape monitoring.
- Mitsubishi Electric: 18+ patents spanning full STM stack — design-around analysis required for new STM entrants
- Hybrid ground + onboard architecture is the converging paradigm — IP strategies must reflect both layers
- Kayhan Space automated CDM agreement framework reduces inter-operator response time from hours to seconds
- 12 CN-jurisdiction filings across 8 Chinese assignees — domestically-anchored portfolio not reliant on Western frameworks
- ADR–constellation interoperability is a white-space IP opportunity with thin current coverage
- Robert Briskman’s foundational onboard autonomy IP (2013 WO) has no corporate assignee — licensing exposure for constellation operators
Satellite Constellation Collision Avoidance — key questions answered
LEO object counts exceeded 10,000 active satellites as of 2023, with projected growth to over 100,000 within a decade driven by mega-constellation deployments from Starlink, OneWeb, and Kuiper.
Ground-centralized systems process SSA data and issue avoidance commands via uplink, while onboard autonomous systems execute collision avoidance logic directly on the satellite using onboard sensors and computing. The key limitation of ground-centralized systems is latency: single-pass communication windows and multi-second uplink delays are incompatible with emergency avoidance timescales.
Mitsubishi Electric Corporation is the dominant single assignee by a substantial margin, with at least 18 distinct patent records spanning US, EP, JP, and WO jurisdictions, covering the full STM stack from SSA business devices to debris removal coordination and OADR architecture.
A Conjunction Data Message (CDM) is a standardized alert specifying when two spacecraft face a conjunction risk. CDMs delivered via email introduce unacceptable response latency for inter-operator coordination. Kayhan Space Corp.’s automated agreement framework addresses this gap by reducing human response time from hours to seconds.
Yes. Literature demonstrates that drag-based collision avoidance via ADCS attitude manipulation can achieve 7–106% deviation in vertical miss distances through cross-section area modulation alone, validating propulsionless maneuver strategies for 2U/3U/6U CubeSat form factors.
Studies modeling Starlink’s impact report that the lifetimes of target satellites can be shortened by up to 99.82% without early avoidance control, and that joining Starlink increased total conjunction risk significantly across all modeled scenarios.
PatSnap Eureka searches patents and research literature to answer instantly.