The Architecture of Automated Interception: A Systems Analysis of Iron Dome and the Human-Machine Defense Chain

Full report

On this page

The proliferation of unguided rockets, artillery, and mortars (RAM), alongside the increasing integration of unmanned aerial vehicles (UAVs) and cruise missiles by state and non-state actors, has fundamentally altered the geometry and tempo of modern air defense. The contemporary battlespace is characterized by saturation tactics: high-density salvos designed specifically to overwhelm the computational bandwidth of defensive radar systems and the cognitive bandwidth of human operators. In this environment, the traditional manual execution of air defense is not just inefficient; it is mathematically and operationally unviable. Developed by Rafael Advanced Defense Systems in collaboration with Israel Aerospace Industries (IAI) and mPrest Systems, the Iron Dome missile defense system entered operational service in 2011 as a direct response to this asymmetric threat landscape1. While initially conceived to counter short-range unguided rockets fired from the Gaza Strip and southern Lebanon, the system has since evolved into a multi-mission architecture capable of intercepting cruise missiles, precision-guided munitions (PGMs), and drone swarms3. However, the system's most revolutionary contribution to modern warfare is not strictly its kinetic interceptor, but rather its automated battle management logic. Iron Dome represents a paradigm shift in human-machine teaming, heavily relying on algorithmic automation to execute a highly complex, micro-second interception chain while elevating the human operator to a supervisory, macro-decision-making role5.

The Anatomy of the Automated Interception Chain

To understand the efficacy of the Iron Dome, one must deconstruct its interception chain. The timeline from the detection of an enemy launch to the kinetic destruction of the threat spans mere seconds. In a fully manual process, an operator would be required to inspect each individual radar track, estimate its trajectory, determine whether it threatens a protected area, select an available launcher, check interceptor availability, avoid conflicting assignments, and authorize the launch8. When multiplied by a barrage containing dozens or hundreds of tracks, human execution inevitably collapses under cognitive overload. Automation accelerates and optimizes this chain through a sequence of discrete, machine-speed operations encompassing eight critical stages.

1. Detect Launch

The engagement sequence begins with the EL/M-2084 Multi-Mission Radar (MMR), developed by IAI's Elta Systems1. Operating in the S-band, the EL/M-2084 is a 3D Active Electronically Scanned Array (AESA) radar capable of detecting and tracking up to 1,100 targets simultaneously for air surveillance purposes3. Unlike legacy mechanically steered radars, the AESA architecture utilizes small solid-state transmit/receive modules to scan the horizon electronically, providing exceptionally fast update rates and high-resolution tracking1. The moment a projectile breaches the horizon, the MMR detects the thermal or kinetic signature of the launch.

2. Establish Track

Once the initial detection occurs, the radar must separate the actual threat from environmental clutter, such as flocks of birds, weather anomalies, or electromagnetic interference12. The EL/M-2084 applies advanced signal processing and active identification algorithms to establish a definitive, continuous track of the object in three-dimensional space, instantly beginning to feed raw spatial data (azimuth, elevation, and range) to the command center13.

3. Calculate Trajectory

Raw radar tracks are instantly transmitted to the Battle Management & Weapon Control (BMC) system, developed by the software firm mPrest Systems2. Utilizing advanced non-linear filtering algorithms, the BMC calculates the projectile's velocity vector, aerodynamic drag, and ballistic arc15. The microservices architecture of the BMC allows it to process millions of data points from the radar network, applying real-time correlative analytics to determine the exact aerodynamic behavior of the object in flight15.

4. Predict Impact Area

Within fractions of a second, the software extrapolates the remainder of the calculated trajectory to predict the precise point of impact on the ground7. The BMC overlays the predicted impact points onto a digital geographic information system (GIS) map of the defended sector, mapping the trajectory against known topographies and infrastructure locations18.

5. Prioritize Threats (Selective Defense)

The most critical automated feature of the Iron Dome is selective defense4. Asymmetric warfare inherently favors the attacker economically; an unguided Qassam or Grad rocket may cost a few hundred dollars, while a single Iron Dome Tamir interceptor costs between $40,000 and $50,0003. Attempting to intercept every projectile in a barrage would rapidly deplete the battery's magazine and impose ruinous financial costs. Therefore, the system autonomously categorizes each track: projectiles predicted to land in open fields, the sea, or unpopulated areas are actively ignored3. The system flags only those projectiles on a trajectory to strike designated population centers, strategic military bases, or critical infrastructure.

6. Assign Interceptor

For threats prioritized for destruction, the BMC evaluates the readiness state of the battery's localized assets. A standard Iron Dome battery consists of three to four launchers, each holding up to 20 Tamir interceptors2. The system calculates the optimal intercept geometry, evaluates which launcher has the best kinematic advantage, ensures interceptor availability, and automatically assigns the engagement to a specific launcher2.

7. Intercept

Upon assignment, the launcher fires a Tamir missile. The Tamir, weighing 90 kg and measuring 3 meters in length, relies on an asymmetric aerodynamic design with steering fins for high maneuverability, capable of reaching speeds of Mach 2.22. The interceptor does not fly autonomously from launch; it receives midcourse guidance updates from the EL/M-2084 radar via a secure command datalink, continuously adjusting its flight path as the target's trajectory updates3. In the terminal phase, the Tamir's onboard active radar seeker acquires the target, and a proximity fuze detonates its high-explosive blast-fragmentation warhead, destroying the incoming threat kinetically or via shrapnel cloud2.

8. Update the Defended Picture

Following the kinetic engagement, the radar conducts a kill assessment. The BMC updates the air situational picture, clearing destroyed tracks from the interface and instantly reallocating computational and kinetic resources to the next wave of incoming threats.

Mitigating Cognitive Overload: The Operational Imperative of Automation

The transition from manual air defense to automated engagement is not merely a technological luxury; it is an absolute tactical necessity dictated by the neurological limitations of human operators. During conflicts such as Operation Protective Edge (2014) or Operation Swords of Iron (2023), adversaries frequently launched high-density salvos specifically to saturate the defense network21. In these scenarios, the automation of the interception chain provides several distinct, force-multiplying advantages.

Parallelism

The human brain processes complex analytical tasks serially, bottlenecking when presented with simultaneous, high-stakes trajectory estimations. Conversely, the BMC processes information in parallel. The system can evaluate, update, and manage hundreds of trajectories at the exact same time without experiencing degradation in processing speed or accuracy7. When a high-density salvo fills the radar screen, parallelism ensures that no single track is neglected due to an operator's divided attention.

Resource Discipline

Automated impact prediction enforces strict resource discipline, ensuring that limited interceptors are not expended on projectiles assessed to be heading toward open areas4. An operator subjected to the extreme psychological stress of a rocket barrage might instinctively fire at a large cluster of radar tracks out of panic or a desire to guarantee safety. The algorithm, unburdened by emotional stress, dispassionately calculates that a high percentage of those tracks will fall harmlessly and allocates interceptors only to genuine threats, thereby preserving the magazine depth of the battery3.

Continuous Replanning

The battlespace is highly dynamic. Rockets alter their trajectories due to atmospheric drag, wind shear, and manufacturing imperfections17. Interceptors may fail to launch, or a launcher may deplete its magazine. The automated chain engages in continuous replanning, ensuring that tracks, predicted impact points, and interceptor availability are recalculated seamlessly throughout the engagement7. If one interceptor fails, the system instantly recognizes the anomaly and can reassign a secondary interceptor from another launcher before the threat breaches the defended zone.

Reduced Cognitive Overload

When faced with an overwhelming number of flashing tracks, auditory alarms, and radio chatter, human operators can experience cognitive tunneling, fatigue, or complete decision paralysis5. Automation abstracts the routine, repetitive calculations away from the operator. By handling the micro-decisions of trajectory estimation and launcher assignment, the system allows operators to supervise the broader defensive picture and manage exceptional conditions, drastically reducing the cognitive overhead required to operate the battery.

Local Resilience

The modular, distributed architecture of the Iron Dome allows a single battery to perform its core detection, tracking, prioritization, and interception functions locally2. The battery does not depend on a distant headquarters for every firing decision. This local resilience minimizes latency—crucial when intercepting threats with flight times measured in seconds—and insulates the battery from centralized command-and-control decapitation strikes or wide-area communication jamming.

The Mathematics of Machine Prediction: Trajectory Estimation Algorithms

The success of the selective defense mechanism, and the justification for removing humans from the micro-decision loop, relies entirely on the mathematical accuracy of the impact point prediction. Estimating the state of a fast-moving, high-altitude projectile from noisy radar data requires the application of advanced non-linear filtering algorithms. In standard radar systems, measurements such as range, azimuth, and elevation contain inherent noise and uncertainty16. The Kalman Filter is the foundational recursive algorithm used to estimate the hidden variables of a system based on these inaccurate measurements, predicting future states based on past estimations16. However, the standard Kalman Filter assumes linear system dynamics, which is insufficient for the highly non-linear kinematics of a ballistic missile subject to varying atmospheric drag, wind profiles, and Earth's gravity16. To address this, modern battle management systems rely on advanced variations of non-linear filters:

• Extended Kalman Filter (EKF): The EKF linearizes the non-linear equations of motion around the current estimate using a first-order Taylor series expansion, generating a Jacobian matrix16. While computationally efficient and widely utilized in legacy systems processing radar returns, the EKF can suffer from severe linearization errors. If the target's non-linearities are extreme, the EKF can experience filter divergence, leading to highly inaccurate estimations24.
• Unscented Kalman Filter (UKF): Developed to overcome the flaws of the EKF, the UKF avoids derivative-based linearization entirely. Instead, it approximates the probability distribution of the target's state using a deterministic set of carefully chosen "sigma points." These points capture the true mean and covariance of the state accurately to the third order, providing significantly better tracking stability for highly dynamic targets26.
• Cubature Kalman Filter (CKF): A more recent evolution in trajectory estimation, the CKF provides enhanced numerical stability and accuracy by utilizing spherical-radial cubature rules to compute multivariate integrals, making it highly robust against the severe measurement noise common in dense electronic warfare environments27.

While ballistic trajectories are relatively predictable once the boost phase concludes, the proliferation of UAVs and cruise missiles introduces profound new complexities. These air-breathing threats do not follow a parabolic arc; they can maneuver unpredictably, change altitude, and exploit terrain masking to avoid detection29. To counter this, modern trajectory prediction is moving beyond pure kinematic filtering toward Physics-Augmented Deep Learning (PADL) and behavioral intention recognition29. Algorithms utilizing Support Vector Machines (SVM) and Bidirectional Gated Recurrent Units (Bi-GRU) are being implemented to analyze spatial-temporal correlation features, predicting a UAV's future flight path based on its recognized behavioral "intent" rather than solely its current velocity vector29.

Where the Humans Move: Human-On-The-Loop (HOTL) Architecture

Public discourse frequently mischaracterizes highly automated weapon systems as entirely devoid of human agency, assuming that "automated" is synonymous with "autonomous." In reality, the integration of automation into lethal systems operates on a well-defined spectrum of control: Human-in-the-Loop (HITL), Human-on-the-Loop (HOTL), and Human-out-of-the-Loop (HOOL)6. Because public descriptions do not consistently disclose the exact human-authorization configuration used in every deployment, Iron Dome should be understood as a heavily automated defensive chain, not automatically as proof that every operational interception occurs without human authorization. The system operates primarily under a Human-on-the-Loop (HOTL) or "supervised autonomy" paradigm5. Inserting a human into the micro-decision loop (HITL) for a rocket traveling at Mach 2 toward a city situated 10 kilometers away would induce fatal latency6. Therefore, a safer and vastly more effective division of responsibility has evolved, changing where the humans move within the defensive chain. The machine operates at tactical machine-speed; the human operates at the strategic, macro-level. Humans establish the parameters of the engagement before the first rocket is even fired. Operators designate the protected zones on the BMC interface, set geographical exclusions, dictate the battery's readiness state, and input the broader engagement doctrine—such as determining the acceptable threshold of risk or the maximum number of interceptors to allocate per incoming threat9. Once a barrage begins, the automation processes individual projectiles at machine speed. The human operators transition to a supervisory role, monitoring the system for anomalies, assessing the broader geopolitical context of the engagement, and, crucially, retaining shutdown or override authority9. The system is profoundly more effective because people do not have to manually rediscover the same trajectory calculations for every single projectile. Human judgment remains strictly reserved for deciding what must be defended, under what circumstances, and with what broader consequences.

The Pathology of Automation Bias: Lessons from the Patriot Fratricides

While automation mitigates cognitive overload, the transition to HOTL architectures introduces new, subtle vulnerabilities into the human-machine team. The primary risk is "automation bias"—the unwarranted, uncritical over-reliance by human operators on algorithmic outputs, leading to a degradation of situational awareness and the ceding of control responsibility to the machine8. The catastrophic potential of automation bias was starkly illustrated during the initial stages of Operation Iraqi Freedom (OIF) in 2003, when U.S. Army Patriot air defense batteries were involved in two high-profile fratricide incidents, mistakenly engaging and destroying a British RAF Tornado ZG710 and a U.S. Navy F/A-18 Hornet8. Post-incident assessments by the Army Research Laboratory revealed that the fratricides were rooted in a complex interplay of "undisciplined automation" in the system's design and "automation misuse" by the crews8. The Patriot system, heavily upgraded after the 1991 Gulf War, was originally designed to counter massive, high-speed Soviet ballistic missile salvos at the edge of space37. This design ethos featured an automatic mode that essentially relegated operators to passive monitors. During the 2003 conflict, the system misclassified the friendly aircraft as anti-radiation missiles and tactical ballistic missiles37. The human operators, conditioned by a command climate that emphasized rapid rote battle drills over critical thinking, accepted the machine's flawed classifications and authorized the lethal engagements8. These incidents highlight the "MABA-MABA trap" (Men Are Better At / Machines Are Better At), a flawed engineering philosophy that assigns tasks to either humans or machines based on isolated capabilities, ignoring the fact that a human-machine hybrid creates an entirely new cognitive entity5. When humans are relegated to passive monitoring roles for systems that operate perfectly 99% of the time, their vigilance inevitably degrades. When the 1% anomaly occurs—such as a misclassified radar track—the human is too disconnected from the situational context to intervene effectively9. For systems like Iron Dome, preventing automation bias requires rigorous, expertise-focused training. Operators must maintain a deep functional understanding of how the algorithms make targeting decisions, including the known weaknesses of the radar's track classification in specific weather anomalies or dense clutter environments8. The human must not merely monitor the screen; they must be trained to actively interrogate the machine's logic, retaining meaningful human control over the application of lethal force.

Operational Efficacy and the Postol Critique

Since its deployment, the Iron Dome has intercepted thousands of rockets, achieving a reported interception success rate of approximately 90% against targets assessed as threats to populated areas1. The system's operational tempo has been severely tested across multiple high-intensity conflicts. During Operation Pillar of Defense (2012), the system intercepted over 400 rockets1. During Operation Protective Edge (2014), militant groups launched roughly 4,500 rockets and mortars; the system successfully intercepted 735 of the approximately 800 projectiles threatening populated areas3. In the May 2021 Operation Guardian of the Walls, despite facing unprecedented barrage densities, the system intercepted over 1,400 rockets, maintaining a kill rate between 90% and 95%3.

Despite these operational claims, the Iron Dome has faced intense scrutiny regarding its actual efficacy. The most prominent critique originates from Theodore Postol, a physicist and missile defense analyst at the Massachusetts Institute of Technology (MIT)40. Postol argues that the Israeli government's claims of a 90% success rate are vastly inflated, suggesting the actual rate of successful warhead destruction is likely 5% or less41. Postol bases his analysis on the visual geometry of interceptor contrails observed in open-source videos, alongside the behavior of the Tamir's laser proximity fuze. He posits that the Tamir interceptors frequently engage the incoming rockets from the side or from behind, rather than head-on40. According to his calculations, unless the Tamir achieves a direct, antiparallel head-on collision, the shrapnel from the blast-fragmentation warhead will only damage the expended rocket motor tube of the target. This would leave the lethal warhead intact, allowing it to fall and detonate on the ground43. Postol attributes the exceptionally low Israeli casualty rates not to the Iron Dome, but to Israel's robust early warning systems, such as the "Color Red" app, and the country's extensive network of reinforced bomb shelters42. However, defense analysts, systems engineers, and empirical ground-damage data provide robust rebuttals to this critique. The primary counter-argument focuses on the specific kinematics of unguided rockets. Hamas and Hezbollah predominantly utilize unguided, aerodynamically fragile rockets21. A side-impact or a proximity detonation that shears the tail fins off an unguided rocket traveling at high velocity destabilizes its aerodynamic profile entirely, causing the projectile to tumble and vastly altering its ballistic trajectory44. Because these rockets are aimed at broad area targets (entire cities) rather than precise point targets, deflecting the rocket out of the sky or causing it to shatter under aerodynamic stress achieves a mission kill, even if the primary warhead is not instantly vaporized in the air44. Furthermore, independent assessments of insurance claim data and the drastically reduced rates of property damage during major rocket barrages strongly corroborate the system's high interception rates, aligning much closer to the official 90% figure than Postol's 5% estimate42.

Multi-Domain Evolution and Transnational Architectures

The success of the Iron Dome architecture has driven its evolution beyond stationary land-based defense, resulting in widespread multi-domain adaptations and international procurement initiatives designed to counter a broader spectrum of threats.

Naval Integration: The C-Dome

As Israel's strategic focus expanded to include the protection of offshore exclusive economic zones (EEZ) and vital natural gas platforms, such as the Leviathan field in the Mediterranean, Rafael adapted the system into a maritime variant known as C-Dome47. First declared operational in late 2022 aboard the Israeli Navy's German-built Sa'ar 6-class corvettes, C-Dome provides full hemispheric protection against surface-to-surface cruise missiles, UAVs, and rockets47. The engineering brilliance of the C-Dome lies in its modular integration. Rather than requiring massive structural redesigns or deep deck penetration, the system utilizes a compact, 40-cell Vertical Launch Unit (VLU) holding the Tamir interceptors19. Furthermore, it completely dispenses with a dedicated fire-control radar, instead seamlessly interfacing with the Sa'ar 6's organic EL/M-2248 MF-STAR AESA radar and the ship's existing Combat Management System (CMS)19. This deep software integration demonstrates the versatility of the mPrest BMC logic, which can seamlessly adapt its trajectory and selective defense algorithms to a moving, pitching naval platform in a highly corrosive marine environment. In 2024, the C-Dome successfully intercepted hostile UAVs over the Red Sea, proving its combat efficacy beyond test environments50.

Transnational Architecture: USMC MRIC

The modularity of the Iron Dome components has also facilitated integration into foreign defense architectures. The United States Marine Corps (USMC), recognizing a critical capability gap in its medium-range air defense against cruise missiles and drones, developed the Medium-Range Intercept Capability (MRIC)51. Rather than purchasing the Iron Dome as a monolithic, closed system, the USMC decoupled its components. MRIC integrates the Israeli Tamir interceptor (adapted in the U.S. as the "SkyHunter") and a highly mobile, trailer-based Iron Dome launcher with indigenous U.S. sensors and command networks51. Specifically, MRIC utilizes the USMC's AN/TPS-80 Ground/Air Task-Oriented Radar (G/ATOR) for target acquisition and tracking, linking the entire architecture through the Common Aviation Command and Control System (CAC2S) and the highly advanced Integrated Battle Command System (IBCS)51. This "plug-and-fight" integration proves that the Iron Dome's kinematic effector (the Tamir) can successfully execute midcourse updates and terminal intercepts even when the targeting data is calculated, prioritized, and broadcast by entirely foreign sensors and command logic54.

The Future of the Defense Chain: Directed Energy Integration

As adversaries adapt to the Iron Dome by exponentially increasing the sheer volume of their barrages, the economic asymmetry of kinetic interception remains a strategic vulnerability. Firing a $50,000 interceptor at a $500 commercial drone or a $1,000 unguided rocket places a definitive upper limit on how long a defense network can sustain operations before facing catastrophic magazine depletion58. To permanently solve the cost-per-kill equation and alleviate the logistical burden of interceptor resupply, Rafael is currently integrating Directed Energy (DE) into the Iron Dome architecture through a system designated "Iron Beam"61. Iron Beam is a 100kW-class High Energy Laser Weapon System (HELWS) designed to operate alongside and complement the kinetic Tamir interceptors. Operating within the same overarching mPrest BMC software architecture, the system will autonomously evaluate incoming threats and apply a hybrid engagement doctrine15. If a threat is detected at very short range, or if the system calculates that a massive drone swarm poses a high risk of magazine depletion, the BMC will assign the Iron Beam laser to engage. The laser intercepts at the speed of light, boasts an unlimited magazine depth (constrained only by the platform's electrical power generation capabilities), and operates at a negligible cost per interception61. Conversely, if weather conditions (such as heavy fog or dust) degrade the laser's effectiveness, or if a heavier ballistic target requires a kinetic kill, the system seamlessly prioritizes a Tamir launch. This hybrid kinetic-directed energy architecture represents the next necessary evolution in automated resource discipline.

Conclusion

The Iron Dome is significantly more than a localized point-defense system; it is a pioneering case study in the successful application of automated, supervisory control in high-stakes combat environments. By fusing advanced AESA radar tracking with microservices-based, non-linear trajectory algorithms, the system conquers the mathematical complexities of ballistic interception. More importantly, through the algorithmic doctrine of selective defense, it solves the operational paradox of cognitive overload and resource depletion during mass saturation attacks. The evolution of modern warfare dictates that human operators can no longer function effectively inside the micro-decision loop of individual weapon engagements. The speeds are too high, the data volume too massive, and the cognitive toll too severe. The Iron Dome proves that the future of defense lies in the Human-on-the-Loop paradigm—where machines execute the flawless, hyper-fast logic of detection and geometry, and humans retain the ultimate authority over strategy, doctrine, and moral consequence. As aerial threats continue to evolve from predictable ballistic arcs to swarming, maneuvering anomalies, the algorithms powering these systems will inevitably shift from statistical filtering to intent-based deep learning. Yet, as the hard-learned lessons of past fratricides demonstrate, the ultimate success of these automated shields will always depend not just on the brilliance of the code, but on the rigorous, critical vigilance of the humans observing the loop.

Works cited

1. Iron Dome: Israel's Air Defense System | PDF | Nozzle | Aerospace Engineering \- Scribd, https://www.scribd.com/document/246875443/Iron-Dome

2. Iron Dome \- Wikipedia, https://en.wikipedia.org/wiki/Iron\_Dome

3. Iron Dome (Israel) | Missile Threat \- CSIS, https://missilethreat.csis.org/defsys/iron-dome/

4. Iron Dome \- RAFAEL Advanced Defense Systems, https://www.rafael.co.il/system/iron-dome/

5. Humans in the Loop \- wp0 | Vanderbilt University, https://wp0.vanderbilt.edu/lawreview/wp-content/uploads/sites/278/2023/03/Humans-in-the-Loop.pdf

6. Command and Control in the Age of Automated Decision Making Process, Challenges and Opportunities \- DTIC, https://apps.dtic.mil/sti/trecms/pdf/AD1177394.pdf

7. Iron Dome Israeli Missile Defense System \- ODIN \- OE Data Integration Network, https://odin.t2com.army.mil/WEG/Asset/Iron\_Dome\_Israeli\_Missile\_Defense\_System

8. Not by widgets alone \- Armed Forces Journal, https://armedforcesjournal.com/not-by-widgets-alone/

9. Lessons from air defence systems on meaningful human control for the debate on AWS Ingvild Bode and Tom Watts \- Drone Wars UK, https://dronewars.net/wp-content/uploads/2021/02/DW-Control-WEB.pdf

10. EL/M-2084 \- Wikipedia, https://en.wikipedia.org/wiki/EL/M-2084

11. ELM 2084 MMR (Israel) \- Missile Defense Advocacy Alliance, https://www.missiledefenseadvocacy.org/defense-systems/elm-2084-mmr-israel/

12. EL/M-2084, https://www.deagel.com/components/elm-2084/a002389

13. IAI Elta unveils next generation radar \- Globes English \- גלובס, https://en.globes.co.il/en/article-iai-elta-unveils-next-generation-radar-1001289528

14. 2084 | ELTA North America, https://eltanorthamerica.com/products/2084/

15. SMART MONITORING, CONTROL AND BIG DATA ANALYTICS FOR INDUSTRIAL IOT \- mPrest, https://mprest.com/wp-content/uploads/2020/08/mPrest-Company-Brochure4.pdf

16. KALMAN FILTER DESIGN FOR BALLISTIC MISSILE DEFENCE APPLICATIONS \- IJCRT.org, https://ijcrt.org/papers/IJCRT2208222.pdf

17. WO2013105093A1 \- Anti-rocket system \- Google Patents, https://patents.google.com/patent/WO2013105093A1/en

18. DEFENSE \- mprest, https://mprest.com/mprest-industries/defense/

19. Israel Adapts Iron Dome Missile Defense to Navy Corvettes \- Seapower Magazine, https://seapowermagazine.org/israel-adapts-iron-dome-missile-defense-to-navy-corvettes/?print=print

20. Iron Dome Explained: Inside Israel's $100000-a-Shot Answer to $300 Rockets \- The Defense Watch, https://thedefensewatch.com/military-ordnance/iron-dome-israel-air-defense-explained/

21. Palestinian rocket attacks on Israel \- Wikipedia, https://en.wikipedia.org/wiki/Palestinian\_rocket\_attacks\_on\_Israel

22. "Swords of Iron" war, https://www.idf.il/media/if4fnlaz/swords-of-iron-war-%D7%91%D7%99%D7%9F-%D7%94%D7%A7%D7%98%D7%91%D7%99%D7%9D-%D7%9C%D7%90%D7%AA%D7%A8-28-10-24-%D7%A1%D7%95%D7%A4%D7%99-1.pdf

23. Autonomous weapons systems and military ro \- Army University Press, https://www.armyupress.army.mil/Portals/7/military-review/Archives/English/pros-and-cons-of-autonomous-weapons-systems.pdf

24. An Improved Extended Kalman Filter for Radar Tracking of Satellite Trajectories \- MDPI, https://www.mdpi.com/2411-9660/5/3/54

25. Extended Kalman Filter for Missile Trajectories | PDF \- Scribd, https://www.scribd.com/document/166200501/875083

26. Ballistic Missile Tracking using Unscented Kalman Filter | Request PDF \- ResearchGate, https://www.researchgate.net/publication/264189321\_Ballistic\_Missile\_Tracking\_using\_Unscented\_Kalman\_Filter

27. Enhanced Ballistic Trajectory Estimation from Radar Data Using the Cubature Kalman Filter \- Biblioteca da SBrT, https://biblioteca.sbrt.org.br/articlefile/4791.pdf

28. State Dependent Multiple Model-Based Particle Filtering for Ballistic Missile Tracking in a Low-Observable Environment \- University Defence Research Collaboration (UDRC), https://udrc.eng.ed.ac.uk/sites/udrc.eng.ed.ac.uk/files/publications/WP2.1\_PUB9.pdf

29. Research on Trajectory Prediction Algorithm Based on Unmanned Aerial Vehicles Behavioral Intentions \- MDPI, https://www.mdpi.com/2504-446X/9/9/640

30. (PDF) Research on Trajectory Prediction Algorithm Based on Unmanned Aerial Vehicles Behavioral Intentions \- ResearchGate, https://www.researchgate.net/publication/395574533\_Research\_on\_Trajectory\_Prediction\_Algorithm\_Based\_on\_Unmanned\_Aerial\_Vehicles\_Behavioral\_Intentions

31. Trajectory Prediction Algorithm of Missile Based on Physics-Augmented Deep Learning | Request PDF \- ResearchGate, https://www.researchgate.net/publication/396962568\_Trajectory\_Prediction\_Algorithm\_of\_Missile\_Based\_on\_Physics-Augmented\_Deep\_Learning

32. A Cryptographic Trust Framework for Agentic Autonomy in Defence \- NATO, https://publications.sto.nato.int/publications/STO%20Meeting%20Proceedings/STO-MP-IST-210/MP-IST-210-2A.03.pdf

33. Evolution of Methods for Countering Drone-Based Airborne Threats, https://jatm.com.br/jatm/article/download/1436/1095/6413

34. Autonomous Weapons Systems: Technical, Military, Legal and Humanitarian Aspects \- ICRC, https://www.icrc.org/en/download/file/1707/4221-002-autonomous-weapons-systems-full-report.pdf

35. The Israeli Occupation Industry \- mPrest Systems \- Who Profits, https://www.whoprofits.org/companies/company/6611?mprest-systems

36. AI Safety and Automation Bias \- CSET, https://cset.georgetown.edu/wp-content/uploads/CSET-AI-Safety-and-Automation-Bias.pdf

37. PATRIOT WARS \- Amazon S3, https://s3.amazonaws.com/files.cnas.org/documents/CNAS-Report-EthicalAutonomy5-PatriotWars-FINAL.pdf

38. Lessons Learned from Blue on Blue for Air Defence Capabilities | The Cove, https://cove.army.gov.au/article/lessons-learned-blue-blue-air-defence-capabilities

39. The Palestine Laboratory: How Israel Exports the Technology of Occupation around the World \- Rah's Open Lid, https://www.rahs-open-lid.com/wp-content/uploads/2024/01/Loewenstein-Antony-The-Palestine-Laboratory\_-How-Israel-Exports-the-Technology-of-Occupation-around-the-World-Verso-2023.pdf

40. How successful was Israel's Iron Dome? | Investigative News \- Al Jazeera, https://www.aljazeera.com/news/2014/9/8/how-successful-was-israels-iron-dome

41. Theodore Postol \- Wikipedia, https://en.wikipedia.org/wiki/Theodore\_Postol

42. MIT Expert Says Israel's Much-Vaunted Iron Dome Only 5% Effective \- The Forward, https://forward.com/fast-forward/202642/mit-expert-says-israels-much-vaunted-iron-dome-onl/

43. This Scientist Explains Why Israel's Iron Dome Is Overrated \- Defense One, https://www.defenseone.com/technology/2014/07/this-scientist-explains-why-israels-iron-dome-overrated/89132/

44. "The Evidence That Shows Iron Dome Is Not Working." Analysis by MIT physicist Theodore Postol \- Reddit, https://www.reddit.com/r/geopolitics/comments/2ca28x/the\_evidence\_that\_shows\_iron\_dome\_is\_not\_working/

45. 7 October Parliamentary Commission Report \- Squarespace, https://static1.squarespace.com/static/67bf0490d422da027d74c55c/t/69ba6cbe190f9704cd62d455/1773825214807/Second+Edition+7+October+Parliamentary+Commission+Report.pdf

46. Iron Dome Software Developer Creates System to Predict Natural Disasters | The Tower, http://www.thetower.org/7290oc-iron-dome-software-developer-creates-system-to-predict-natural-disasters/

47. Israel Just Used Its 'C-Dome' for the First Time. Here's What to Know About the Defense System, https://time.com/6965118/israel-c-dome-ship-defense-system/

48. C-DOME™: Maritime Application of the Combat-Proven IRON DOME \- RAFAEL Advanced Defense Systems, https://www.rafael.co.il/system/c-dome-2/

49. Sa'ar 6-class corvette \- Wikipedia, https://en.wikipedia.org/wiki/Sa%27ar\_6-class\_corvette

50. Watch Israel's Sa'ar 6-Class Corvette In Action | Naval Iron Dome Intercepts Drone In The Red Sea \- YouTube, https://www.youtube.com/watch?v=rtKUpHrpOi4

51. Program Acquisition Costs by Weapons System \- Office of the Under Secretary of Defense (Comptroller), https://comptroller.war.gov/Portals/45/Documents/defbudget/FY2027/FY2027\_Weapons.pdf

52. U.S. Army Advances Next-Generation Interceptor for Enhanced Air Defense \- https://debuglies.com, https://debuglies.com/2024/08/03/u-s-army-advances-next-generation-interceptor-for-enhanced-air-defense/

53. Second defender: Army to pick batch of cruise missile interceptors in 2025 for new competition \- Breaking Defense, https://breakingdefense.com/2024/01/second-defender-army-to-pick-batch-of-cruise-missile-interceptors-in-2025-for-new-competition/

54. Defending the skies – US style \- European Security & Defence, https://euro-sd.com/2025/02/articles/42609/defending-the-skies-us-style/

55. AN/TPS-80 Tested Against Cruise Missiles \- European Security & Defence, https://euro-sd.com/2022/08/news/27015/an-tps-80/

56. Raytheon Technologies USA | Page 2 | DefenceHub | Global Military & Security Forum, https://defencehub.live/threads/raytheon-technologies-usa.16345/page-2

57. MISSILE, ARTILLERY, HYPERSONICS, BALLISTICS AND, https://battle-updates.com/update/missile-artillery-hypersonics-ballistics-and-soldier-systems-update-6/

58. Tag: iron dome \- Missile Threat \- CSIS, https://missilethreat.csis.org/tag/iron-dome/

59. Two years since 7 October: Israel's forever wars across the Middle East | ACLED, https://acleddata.com/report/two-years-7-october-israels-forever-wars-across-middle-east

60. Israel's Operation Swords of Iron Update October 18, 2023 | JINSA, https://jinsa.org/wp-content/uploads/2023/10/Israels-Operation-Swords-of-Iron-Update-10-18-23-3.pdf

61. IRON BEAM \- Rafael USA Inc., https://www.rafael-usa.com/programs/iron-beam/