Space Mission Risk Assessment

As our reliance on satellites grows; guiding everything from our GPS systems to global communications; it’s becoming increasingly clear just how fragile our space infrastructure really is. Cyberattacks, Anti-Satellite (ASAT) weapons, and space debris are no longer just concerns for the future; they’re real threats that are already challenging the way we operate in space. Yet, most security conversations still focus on the risks we face on Earth, ignoring the vulnerabilities above us. I recently came across a paper that introduces a new way of thinking about space security: the Multi-Threat Risk Management Framework (MTRMF). This framework is all about resilience, not just defense. It’s about ensuring that even when space systems come under attack, they continue to function; by building in redundancies, diversifying communication methods, and having quick-response strategies ready to go. What really stood out to me was how the MTRMF uses real-time data; both open-source and classified; to continuously assess and prioritize risks. The framework emphasizes that space security isn’t just a national issue; it’s a global one. Protecting space requires collaboration across governments, private companies, and international organizations. The paper also takes a deep dive into the vulnerabilities of GPS, which is crucial to everything from air traffic control to financial transactions. Whether it’s cyberattacks on ground stations or the threat of ASAT weapons, disruptions to GPS could have far-reaching effects. The MTRMF’s resilience-based approach is key to keeping these critical systems safe, no matter what threats lie ahead. Full paper -> https://lnkd.in/d8PQZHX8 #SpaceSecurity #SatelliteProtection #MTRMF #Cybersecurity #SpaceRisks #GPS #Resilience #ASAT #SpaceDebris #GlobalCollaboration #TechInnovation #SpaceDefense

Three anomalies. One rocket. $4 billion per launch. Artemis II is now NET April 1, 2026 — and the helium flow failure that triggered last week’s VAB rollback is not the story. The story is what the anomaly pattern reveals about program architecture. The diagnostic: Since January 2026, NASA has logged three distinct technical failures during ground operations alone: → Liquid hydrogen leak (Feb. 2 wet dress rehearsal — aborted) → Orion hatch pressurization valve failure (Feb. 19) → ICPS helium flow interruption (Feb. 21 — triggered rollback) In program risk management terms, three pre-flight anomalies across three independent subsystems is not bad luck. It is a systemic signal. The structural issue most analysts miss: SLS is a cost-plus legacy architecture operating in a commercial-velocity environment. At $4B per launch, the program has no margin for iterative ground failures — financially or politically. The One Big Beautiful Bill locked in $4.1B for SLS through Artemis V. But every anomaly strengthens the case for commercial alternatives that cost a fraction of that figure. Meanwhile, NASA just officially restructured Artemis III. What was designed as humanity’s first crewed lunar landing since 1972 is now a low Earth orbit rendezvous and systems test — with the first actual surface landing reassigned to Artemis IV in 2028. That is a 3-year slip from original program baselines. In any capital allocation framework, that warrants a portfolio reassessment. The strategic read: SpaceX: Retains Starship HLS contract. Artemis III’s LEO pivot accelerates Starship’s centrality to the architecture. Net winner. Blue Origin: NASA reopened lunar lander competition in October 2025. Blue Moon is back in contention for Artemis IV. Competitive positioning strengthened. SLS/Boeing/Northrop: Institutionally protected by legislation. But the efficiency narrative is deteriorating with each rollback. The overlooked variable: New heliophysics research identifies an active solar superflare window running through mid-2026. Four astronauts operating beyond Earth’s magnetosphere in April face elevated radiation exposure risk. NASA’s risk modeling for this specific window has not been made public. The bottom line: Artemis II will fly. The Moon race is structurally real and geopolitically motivated. But the program is surfacing a fundamental tension that every serious space investor must price into their models: The gap between government program timelines and commercial execution velocity is not narrowing — it is widening. Capital flowing into the lunar economy needs to track that delta. Where do you see the architecture heading post-Artemis IV? I’d value the perspective of the operators and investors in this network.

𝗖𝗼𝗻𝗷𝘂𝗻𝗰𝘁𝗶𝗼𝗻 𝗮𝗹𝗲𝗿𝘁𝘀 𝗮𝗿𝗲 𝗻𝗼𝘁 𝗼𝗻𝗹𝘆 𝗮𝗯𝗼𝘂𝘁 𝗱𝗶𝘀𝘁𝗮𝗻𝗰𝗲. “It says 1km. We’re fine.” The real risk isn’t just how close you’ll get. It’s how 𝘶𝘯𝘤𝘦𝘳𝘵𝘢𝘪𝘯 we are about where you’ll be. Every object in space has uncertainty. That uncertainty is usually modeled as a 3D ellipsoid, which is a cloud of possible positions. Even if two satellites are predicted to pass 1 km apart, their uncertainty ellipsoids might still overlap. That’s what drives collision probability (Pc). Not raw distance. 𝗛𝗲𝗿𝗲’𝘀 𝘄𝗵𝗲𝗿𝗲 𝗶𝘁 𝗴𝗼𝗲𝘀 𝘄𝗿𝗼𝗻𝗴: • Acting on distance-only thresholds (e.g., "1 km = no action") • Relying on inappropriate orbital information, such as TLEs, which lack covariance info and aren’t updated fast enough • Issuing impulsive CAMs that move the satellite 𝘪𝘯𝘵𝘰 a riskier position 𝗪𝗵𝗮𝘁 𝘆𝗼𝘂 𝘀𝗵𝗼𝘂𝗹𝗱 𝗱𝗼 𝗶𝗻𝘀𝘁𝗲𝗮𝗱: • Use CDMs (Conjunction Data Messages) that include covariance and Pc • Run your own OD if you have GPS data and/or observations, and generate your uncertainty ellipsoids • Only act when the Pc crosses a known threshold (generally 1e-4) #SpaceStartups #Astrodynamics #ConjunctionAnalysis #CollisionAvoidance #LEO #SpaceDebris #SatelliteSafety #NewSpace

FEM/DEM Precision vs. Regolith Reality This week, I met with colleagues from the space industry to discuss the use and misuse of FEM/DEM modeling for ground design. One thing is clear: significant gaps exist due to a misunderstanding of basic soil mechanics. We are overlooking the fundamental lesson taught by our predecessors, Terzaghi, Peck, and Casagrande, that even on Earth, we must prioritize judgment and risk-accounting over blindly rational deduction due to the uncertainties inherent in soil. One of the primary objectives of the Artemis program is to construct habitats and infrastructure on the Moon, indicating a move from short-term exploration to long-term design and building efforts. The key question is whether space agencies, private companies, and academic institutions are fully prepared for this change in focus, or if they are still primarily focused on resource extraction rather than sustainable development. Perhaps this dichotomy is the core reason for the reliance on FEM/DEM models to simulate regolith and its physical parameters. These models clearly provide precise, high-resolution outputs; no doubt about it. However, when we feed them single, deterministic values for parameters like cohesion or friction angle, inputs often extrapolated from flawed simulant tests, we generate a false sense of security. My takeaway is: The model's numerical convergence must never be mistaken for the design's geotechnical reliability. This represents an unacceptable level of risk for any human-rated missions. We urgently need more Geotechnical Engineering SMEs at the design table. I will continue advocating for a shift to Reliability-Based Design (RBD) as the only responsible path forward. Incorporate risk mitigation into your design by preparing for multiple scenarios. Use probabilistic methods to quantify the risk of failure explicitly (Pf) rather than merely satisfying a single FoS. What critical piece of geotechnical information are you over-simplifying, and therefore mismanaging the risk? Note: The image below is AI-generated to provoke fruitful discussion. #TheMoonBuilders #SpaceGeotech #LessonsLearned #SpaceEconomy #GeotechRisks #MoonConstruction #Design #FEM #DEM

The space systems differ significantly from terrestrial networks. They are often built with older technology and designed to last for decades, making mid-life upgrades difficult or impossible. These characteristics create vulnerabilities, particularly in areas like command systems, uplinks, and user devices. For example, attackers can exploit weak encryption or outdated software to disrupt communications or hijack control of satellites. The guide recommends applying frameworks like the NIST Cybersecurity Framework (#CSF) to understand and mitigate risks across different segments of space systems—ground, space, uplink/downlink, and user devices. Each segment faces distinct challenges, from #malware threats in ground stations to #jamming and #spoofing attacks on uplinks. For user devices, weak #cybersecurity measures can allow attackers to spoof signals or cause denial of service. One of the standout messages is the importance of integrating cybersecurity measures into the design phase of space systems. This includes encryption, network segmentation, and supply chain security. The guide also stresses the need for continuous monitoring and regular software updates to defend against evolving threats.

When people talk about New Space, they usually highlight cost curves: smaller spacecraft, faster production, shorter integration cycles, reusable rockets. True. Modern space hardware should be cheaper to build, easier to deploy, and simple to replace. But there’s a blind spot: you can’t design next-generation systems using yesterday’s environmental data. Many legacy baselines were built on limited or outdated #spaceweather observations. Engineering based on historical averages has a long record of failing in dynamic environments. A classic example: Titanic sailed using iceberg climatology—historical ice-field patterns—rather than real-time surveillance. Space is no different. #Radiation levels, plasma density, drag variability, and surface charging conditions change rapidly and are not captured by historic averages. Designing a mission as if the environment is static leads to over-engineering or under-protection. A few examples that illustrate why real measurements matter: • Starlink’s February 2022 loss SpaceX lost 38 satellites to increased atmospheric drag after a geomagnetic storm sharply expanded the thermosphere; historical drag models did not predict the magnitude. • GPS degradation in #aviation Radiation-driven scintillation events disrupt VHF and GNSS navigation in polar routes. Airlines now adjust operations based on space-weather alerts because averages can’t capture fast ionospheric disturbances. Source: https://lnkd.in/eVr3yVG7 • Airbus A320 flight-control data corruption Regulators recently required urgent software and hardware patches after strong radiation corrupted flight-control inputs on an A320. The vulnerability emerged under modern solar-cycle conditions. • Lunar surface charging Apollo astronauts reported unexpected static buildup. Modern NASA research shows that surface potential swings can exceed hundreds of volts during terminator crossings. No historical dataset captures the full risk; new missions require fresh measurements. Source: https://lnkd.in/e3aKqgqx All of this points to one conclusion: you can’t guess the environment. You have to measure it. Shielding alone won’t solve it. Redundant comm channels won’t solve it. Digging deeper into lunar regolith won’t solve it. Those are static fixes for a dynamic problem. China already dominates manufacturing throughput. If we want to lead, the win won’t come from cheaper factories; it will come from better R&D, situational awareness, and environmental intelligence. Space systems built on real, high-temporal-resolution data will outperform those built on historical assumptions. That’s the actual frontier of New Space. Also, yes — we have observation missions. ACE, DSCOVR, GOES, POES, SWARM, TIMED, Cluster, THEMIS, and others have shaped everything we know about the near-Earth environment. But having missions and having the data required for modern engineering are not the same thing.

AI Safety Beyond Earth: How RideScan's ‘Teacher Brain’ Could Safeguard Space Missions   Edinburgh-based deep-tech startup RideScan is pioneering an innovative 'teacher brain' approach that could revolutionize safety in autonomous space missions 🚀🤖   The start-up inspired by the PhD work of founder and CEO Shivoh Chirayil Nandakumar from The National Robotarium - the Robotics and Artificial Intelligence Research Centre at Heriot-Watt University – designed this innovative solution initially for terrestrial autonomous systems. Its implications could extend far beyond Earth's boundaries.   RideScan's core innovation is what they call a "teacher brain" - an external intelligent monitor that observes and evaluates autonomous systems without interfering with their proprietary technology. This approach is particularly intriguing because it mirrors human cognitive processes, especially our ability to detect anomalies in familiar patterns. This teacher brain will observe multiple "children" or autonomous robots, objectively evaluate their performance, and report any abnormal behaviors including risk scores to relevant stakeholders.   Aleatoric uncertainty is a type of uncertainty that describes outcomes that are random and cannot be predicted. The RideScan CaaS (Cognition as a Service) model detects aleatoric uncertainties and anomalies, enabling proactive risk mitigation and preventing catastrophic failures.   The space sector presents unique challenges where this technology could prove transformative. Consider autonomous spacecraft operations or future Mars rovers. These systems must operate in extreme conditions, far away from Earth, where traditional human real-time control is unfeasible due to the long communication delay.   RideScan's technology could be adapted to create an additional layer of safety for space operations by:   1. Monitoring autonomous space systems for behavioral anomalies without requiring access to their core programming 2. Providing real-time risk assessment for critical operations 3. Detecting potential malfunctions before they become catastrophic 4. Offering valuable feedback for improving autonomous space systems' reliability   As we rely more heavily on autonomous systems, technologies like RideScan's could become crucial in ensuring mission success and safety. The company's work at the National Robotarium, testing their system across various platforms from quadruped robots to robotic arms demonstrates the versatility that could make it ideal for space applications. The future of space exploration will depend not just on our ability to build autonomous systems, but on our capability to make them consistently reliable and safe. Image Credit: RideScan - The relationship of anomalies and faults in the context of autonomous robotic missions.   #AISafety #RoboticSafety #SafetyInSpace #AIinSpace #SpaceSafety #SpaceAutonomy

🛡️When a satellite or space-based service is disrupted, the mission depends on knowing why - before ambiguity dictates the response. Detecting that something happened in #space is not the same as understanding what happened. In a contested domain, that difference can determine whether decision-makers respond appropriately, hesitate - or unintentionally escalate. This month, two independent signals from the #UK and #US national-security space communities converged around the exact capability gap Space DOTS® has been building to address. The UK’s newly updated Cross-Government Space Domain Awareness Requirements by UK Ministry of Defence x UK Space Agency define #SDA as more than tracking objects in orbit. The value of our approach lies in three key priority areas: • Persistent intelligence monitoring: identifying whether disruptions stem from environmental effects, system faults, electromagnetic or cyber interference, or hostile activity. • Predictive Space Intelligence: combining in-orbit sensing, external data and decision-support tools to assess ambiguous events and enable timely responses. • Resilience in contested environments: complementing UK and allied SDA data to strengthen multi-domain awareness and redundancy. At the same time, the The Mitchell Institute for Aerospace Studies’ newly published report, 'Space Superiority Through the Spectrum of Conflict', reached an equally direct conclusion: “Attribution is essential for effective response options.” But attribution in space is not simply about identifying an actor. It requires understanding the nature and scope of the effect, which systems were involved, whether the outcome was intended - and whether an apparent attack could instead be explained by a spacecraft fault or the natural space environment. Without that context, ambiguity delays decisions, weakens coalition alignment and increases the risk of both underreaction and unintended escalation. This is precisely why we built SKY-I, our early anomaly-intelligence platform, designed to combine proprietary in-orbit environmental measurements, external datasets and analytical models to help determine whether the disruption of a satellite or space-enabled service is: environmental, technical, caused by #Electronicwarfare or #cyber interference, potentially hostile, or still genuinely unresolved. For #defense, ambiguity delays response and complicates escalation control. For satellite operators, it delays recovery and threatens mission continuity. For financial institutions and insurers, it weakens the evidence needed to assess #PNT disruption, quantify exposure and understand what actually caused a service failure. And attribution is only one part of the mission. The objective is not merely to detect anomalies when they occur. It is to make invisible space risks predictable - so organisations can anticipate disruption, protect critical services and act on evidence rather than assumption. ▶️ SKY-I demo teaser in comments.

Modelling future launch traffic and the associated risk to new missions Due to the recent increase in the rate of launches to space, there is a renewed interest in modelling the evolution of the space environment. These models aim to predict how objects, including active spacecraft and debris, will be distributed in the environment in future. This work focusses on two components of an overall environment model: the launch traffic model and the risk model. Many space environment models that require an estimate of launch traffic will either use no launches as a baseline or repeat historical data. Recent shifts in launch traffic have shown that this approach is unrealistic and does not account for potential further changes in launch traffic. Models of risk in the space environment aim to quantify the risk posed to a mission by other objects in the environment and the additional risk a new mission generates to other objects in the environment as a result of potential collisions. This is difficult to calculate over the lifetime of a mission due to the number of encounters and the uncertainty in how active satellites are distributed in the environment. Our proposed approach to modelling launch traffic is a parametric model that can simulate various future launch scenarios. This model calculates the total number of objects launched per year using an exponential-logistic curve to capture the sharp increase in launch traffic. To determine the location and physical characteristics of launched objects, the model fits probability distributions to data from previous launches. These distributions can vary over time to simulate other changes in launch trends aside from the number of objects. The launch traffic model can be used along with other debris environment modelling tools to simulate the evolution of the space environment and accordingly to calculate the risk of collision for new missions. Our proposed model of risk estimates the likelihood and severity separately, where risk likelihood is a measure of the probability of collision and severity quantifies the consequences of a collision. The likelihood estimate comes from the intersection of distributions of orbital parameters, which is suitably fast to calculate over the lifetime of a mission. The model of severity in the environment uses a simple model to approximate the effects of a collision without needing to propagate all fragments. In addition, we model the severity of collisions for the mission of interest based on the predicted effect of a collision on the spacecraft. This work used different launch scenarios to generate varying evolutions of the space environment. Results are shown from applying the risk model to example missions across these scenarios Callum Wilson Francesca Letizia Stijn Lemmens Jinglang Feng Keiran McNally Andrew Ratcliffe Robert Garner Aerospace Centre of Excellence Institute on AI for Space Sustainability #spacedebris #spacesustainability