As missile defense systems and autonomous aerial platforms grow more complex, the next era of national-security technology will depend on tighter data integration, smarter coordination and AI that can handle unseen scenarios. New technical architectures, improved policies and adaptive wireless networks are emerging as central requirements for resilience in space and the air domain.
Both missile defense networks and autonomous UAV systems face parallel challenges: they rely on diverse sensors, must operate in contested environments and depend increasingly on intelligent software to interpret data and guide decisions.
Integrating Today’s Missile Defense Systems
Across the defense sector, stakeholders are grappling with architectures that combine legacy platforms, new sensors and multinational assets. The result is an environment where interoperability and translation across diverse systems are becoming foundational requirements, Robin Dickey, director of policy and government affairs at Slingshot Aerospace, told Constellations.
“Integrating missile defense system is like conducting an orchestra over Zoom.” -Robin Dickey, Slingshot Aerospace
“The analogy that I like to use is the idea that integrating a missile defense system is like conducting an orchestra over Zoom, and it’s because there’s both the technical problems, there’s policy challenges, there’s players that were not trained or initially designed to do the exact purpose that they’re being integrated for,” said Dickey. “There’s always some level of translation that you might have to do if you’re using diverse pieces of the architecture.”
Allied sensing will follow the same pattern: More partners improve coverage but make integration harder, said Dickey, returning to the Zoom-orchestra analogy. Deepening collaboration will require more joint exercises and realistic training, she added. AI-enabled, physics-based simulations allow partners to practice together without risking on-orbit assets—essential for building trust and operational cohesion.
Integrating different types of sensors and data will require a mix of contracting and acquisition models, including systems that may remain contractor-owned and operated as long as they offer interfaces that allow their data to interoperate with defense systems, Dickey said. Commercial sensors and military sensors both add value, but a major opportunity lies in the fusion and command-and-control layers, where software helps interpret data rather than just collect it.
Safeguarding the Space Domain
The industry is also placing greater emphasis on protecting the architectures themselves, not just the threats they track. This reflects a broader shift toward space domain awareness and defensive cybersecurity for space assets, Dickey noted.
“It’s not just enough to have a system that’s integrated, that is doing its job. If there’s a bad guy who wants to mess with our capability to defend against missiles, we need to be able to protect that capability,” Dickey said. “And so that’s everything from maintaining the intelligence of what’s happening in space – not just tracking the missiles, but tracking who’s approaching our missile tracking satellites and figuring out when do we need to maneuver, when is there an anomaly or a threat that needs to be responded to? And so there’s more than just the data around an individual missile threat that all needs to be integrated into this bigger picture that is the defense mission.”
National missile defense depends on being able to constantly assess risk, including identifying objects in orbit that could pose a danger and figuring out how to protect or reposition parts of the architecture to keep operations running smoothly.
Threats to space-based defense architectures aren’t limited to adversaries, Dickey said on a panel during SATShow Week. National missile defense depends on being able to constantly assess risk, including identifying objects in orbit that could pose a danger and figuring out how to protect or reposition parts of the architecture to keep operations running smoothly, she said. This requires integrating and accounting for every aspect of the environment—from weather on Earth to conditions in space—so the system can function in real time, Dickey said. This becomes even more important when adversaries are aware of these capabilities and may try to interfere, she said.
AI’s Expanding Role in Data and Threat Detection
The rapid growth of space-based data is also pushing agencies to adopt AI tools that can process and triage information at scale and support real-time decision-making—a trend seen widely across intelligence, surveillance and reconnaissance missions, according to Dickey. For example, AI and machine-learning tools can scan massive datasets to spot unusual behavior in satellite constellations or quickly fingerprint newly launched objects. With the volume of space data growing, the real advantage commercial providers can offer is turning that data into usable intelligence that supports decisions, not simply delivering raw information.
Information overload from interpreting data from multiple sensors is another challenge that AI and machine learning can help mitigate, Dickey noted. “We’re still a long way off from and maybe never will be at the point where humans should be out of the loop making decisions,” she said. “They absolutely should be in and on the loop for high-stakes decisions, but you need to lighten the load on what decisions they’re making and so being able to do machine learning in which you’re processing and identifying [patterns] from large amounts of information, or even agentic artificial intelligence.”
Next-generation missile tracking will depend on better autonomy in how satellites detect, hand off and maintain custody of fast-moving threats, said Dickey. No single sensor can cover an entire missile timeline, so the challenge is integrating multiple sensors and “knowing which sensors to use” as part of a full kill chain. That requires stronger fusion and AI-driven processing, Dickey said.
Aligning Policy With Rapid Innovation
But even as analytic capabilities improve, the industry faces a different obstacle: the speed of policy and acquisition. Despite advances in policy, the industry continues to face procedural bottlenecks during system fielding. This challenge is common across defense modernization programs, where authorities, approvals and cultural adoption often lag behind available technology, Dickey said. Standardization is another recurring challenge across defense programs as more commercial and international partners become involved.
“It’s always a cycle of catch-up. You can always want to go faster, try to go faster, and therefore, it’s also never too late to try to get these standards in place,” said Dickey. “So I wouldn’t say necessarily that there’s a point of no return. It’s more that this is something that we’ll just have to keep working and developing, and maybe you find some of the standards aren’t perfect the first time you try them, so you’ve got to iterate and make those better through constant effort.”
AI-Enabled Positioning and Navigation for UAVs
Parallel to these space-focused challenges, airborne autonomy faces its own push for more robust sensing, network support and resilience in contested environments, said Dr. Walid Saad, professor at the Virginia Tech Department of Electrical and Computer Engineering and Next-G wireless lead at the Virginia Tech Institute for Advanced Computing.
Precise UAV navigation increasingly relies on two advances, according to Saad: integrated sensing and communications (ISAC) for RF-based positioning and fast AI algorithms that interpret the environment and adjust flight paths in real time.
“The wireless network has a broader coverage of the physical world around autonomous systems and can help in providing more precise command and control.” -Walid Saad, Virginia Tech
“The wireless network has a broader coverage of the physical world around autonomous systems and can help in providing more precise command and control, in defense and space-related missions,” Saad said. “The network can act as an ‘intelligence’ provider, guiding the decision making of the devices.”
Three approaches look especially promising for making localization more resilient to jamming, spoofing, and other GNSS disruptions, according to Saad:
- Multi-modal positioning that blends GNSS with network signals, inertial sensors, and vision systems to add redundancy and reduce single-point failures;
- ISAC techniques that extract positioning and environmental data directly from communication signals, which are harder to spoof.
- AI methods that detect anomalies, identify spoofing, and fuse data from multiple sources to keep localization reliable under attack.
In the longer term, quantum sensors may also offer reliable guidance in GPS-denied environments, Saad said.
The scalability and reliability of UAV swarm operations are still constrained by communication, coordination and robustness challenges, Saad said.
“As swarm size grows, maintaining reliable, low-latency communication becomes increasingly difficult, especially in contested or spectrum-constrained environments. This directly impacts the ability of UAVs to coordinate decisions in real time,” he said. “In addition, decentralized control remains challenging, as each UAV has only partial and potentially noisy information about the global state, which can lead to instability or suboptimal collective behavior.”
Vulnerabilities to interference, jamming and dynamic environmental conditions further limit the reliability of a UAV’s decisioning, Saad said.
“The biggest opportunities lie in tighter integration between sensing, communication and control, supported by AI,” Saad said.
Network-enabled intelligence that offers wider situational awareness can improve swarm coordination, while advances in distributed AI let UAVs make effective decisions with limited data and still achieve efficient overall behavior, he said.
“By using sensing information, the wireless network can now act as an additional intelligence component for the UAV,” said Saad. “Since the wireless network has a broader coverage/perspective on the world around the UAV, it can provide guidelines for enhancing navigation and control decisions.”
Additionally, the wireless network can leverage readily available communication signals to supplement UAV-level capabilities in terms of positioning, but only through a joint integration of capabilities at both the UAV device level and network level, he said.
Multi-Layer Connectivity, Semantic Communications and Future Autonomy
Across all domains, defense modernization is now converging around multi-layered networking. The trend is toward architectures where terrestrial, airborne and space-based links function as a single adaptive system.
“The most critical capability is the ability to leverage network-level sensing to help autonomous systems adapt to unforeseen scenarios – arguably one of the central challenges in AI,” Saad said.
The shift toward “world-model” reasoning is also becoming widespread across autonomous systems, allowing platforms to anticipate rather than merely react.
The shift toward “world-model” reasoning is also becoming widespread across autonomous systems, allowing platforms to anticipate rather than merely react.
With information from the network, an autonomous agent can build a broader world model that lets it understand its environment and anticipate possible future scenarios, instead of relying only on past data, Saad said. This kind of predictive reasoning helps systems handle conditions they weren’t trained for, which is crucial in fast-changing or contested settings, he said.
For example, if a UAV experiences intermittent jamming, the network can indicate whether it’s a local issue or a wider problem, helping the UAV adjust its route or communications, said Saad. And in completely new situations—like unusual interference or unexpected obstacles—the network-informed world model lets the UAV consider different outcomes and choose resilient actions even without prior training data, he said.
Distributed autonomy depends on networks that can dynamically complement each other, a major focus of current 5G-Advanced and 6G advancements.
Operators should approach these multi-layered networks as a unified wireless system rather than a collection of independent network links, Saad said.
“The key is to tightly integrate terrestrial, airborne and non-terrestrial networks so that they can dynamically complement each other in terms of coverage, reliability and latency,” he said.
Each network layer offers distinct advantages, according to Saad: terrestrial networks deliver high capacity and low latency, airborne platforms add flexible coverage and non-terrestrial systems like satellites provide wide-area reach – though sometimes intermittently. The challenge is coordinating these layers smoothly, which requires intelligent link selection, adaptive spectrum use and coordinated resource management across the network, he said.
“Task-aware communication, or more broadly what we call semantic communication, will be critical for enabling autonomous systems to operate under tight bandwidth constraints,” Saad said. “Rather than transmitting raw data, the goal is to communicate only the information that is relevant to the task at hand, such as intent, key situational cues or decision variables. This is the core idea behind semantic communications, where the focus shifts from transmitting bits to conveying meaning, allowing systems to ‘do more with less’ by leveraging compact semantic representations.”
Semantic communication further helps systems stay functional through jamming or coverage gaps by sending compact “semantic showers” that AI can reconstruct into full meaning, reducing network load and preserving performance even with partial data, he said.
Navigation resilience remains a central driver behind these innovations across defense aviation, said Saad.
A key problem faced by today’s autonomous airborne systems is loss or degradation of reliable positioning in GPS-denied or contested environments, which can result in navigation errors or mission failure, Saad said.
“Emerging capabilities such as ISAC, multi-source localization and network-assisted positioning can provide alternative and complementary positioning signals,” Saad said. “When combined with AI-based data fusion and anomaly detection, these approaches can maintain accurate and resilient navigation even under jamming or spoofing conditions.”
Present AI systems can also be prone to generalizing unforeseen scenarios, Saad noted.
Adaptability to unknown situations is another major industry focus as more autonomy moves to the edge, Saad said.
“Today’s AI systems when embedded in physical AI devices like drones, rely extensively on what they saw in the training data. Once they encounter unforeseen scenarios, they may fail or require significant retraining,” he said. “By leveraging network-level world models, wireless systems can help those autonomous devices navigate and adapt to such unforeseen conditions.”
Powering the Shift Toward Unified Architectures
Overall, the direction of the industry is toward architectures that merge sensing, communication and control into a single intelligent fabric, said Saad.
“The most impactful decisions will be those that treat communication, sensing and control as a unified system rather than separate functions,”
“Architectures that support ISAC, semantic communications and AI-driven distributed decision-making will enable tighter integration between air and ground assets,” Saad said. “In addition, designing systems to be adaptive through flexible spectrum use, modular AI frameworks, and multi-layer connectivity will ensure long-term scalability and robustness as environments and mission requirements evolve.”
The next generation of UAV and high-altitude systems will be defined by tighter integration across radios, compute, sensing and timing to enable fully adaptive and resilient autonomy.
The next generation of UAV and high-altitude systems will be defined by tighter integration across radios, compute, sensing and timing to enable fully adaptive and resilient autonomy, Saad said.
Wireless capabilities such as integrated sensing and communications, multi-layered connectivity and semantic communications will allow systems to extract information from their environments while efficiently sharing intent and situational awareness with other systems, he said.
Integrating multi-source positioning and communication that combines GNSS with network-based sensing, ISAC and onboard sensors can reduce single points of failure, said Saad. Shifting toward multi-layer, hybrid connectivity architectures can also enhance redundancy and support resilient operations, he said.
Finally, Operators should also deploy AI-driven frameworks for real-time data fusion and anomaly detection so that their systems are more adaptable to interference and environmental changes, he said.
Embedded AI that can build and update world models in real time will enable autonomous systems to make predictive and proactive decisions under conditions of uncertainty, while improvements in distributed sensing along with highly accurate timing and synchronization will enable multi-agent systems to coordinate more reliably, Saad said.
“AI will play a central role in learning how to switch between links, aggregate them when needed and maintain robust connectivity under mobility, interference and varying channel conditions,” Saad said. “Ultimately, the goal is to provide a continuous and resilient communication fabric that supports real-time control, sensing and decision-making for autonomous UAVs, even in highly dynamic or contested environments.”
“Collectively, these capabilities will enable autonomous systems to operate robustly in dynamic, contested and previously unseen environments,” he said.
Explore More:
Interceptors ‘Only 1 Piece’ of the Puzzle: Missile Defense Begins in Space
Countering Hypersonics from Space: Multi-Orbit Missile Warning/Missile Tracking
DoD’s Increasing Demand for Data Fusion in Space
