By the time a human operator recognizes a problem, debates the options, and issues new instructions, the mission may already be lost.
That reality underlies the recent shift in focus to mission-level autonomy as an operational necessity rather than a future aspiration. As contested environments, degraded communications and time-compressed decision cycles become the norm, autonomy can no longer be limited to executing predefined tasks or optimizing individual platforms. Autonomous systems must understand intent, reason about goals and adapt when plans inevitably collide with reality.
“At the mission level, we’re concerned with what the human wants, why they want it and ideally not having humans need to specify every little thing,” said Dr. Matt Molineaux, Director of Autonomy at Parallax Advanced Research. “You want something that’s assisting you—not something that has to be handheld through every step of the way, but also not something that violates operator intent.”
From Tasks to Missions
Much of today’s autonomy performs well in controlled conditions. Task automation focuses on executing predefined actions. Platform autonomy optimizes how a vehicle or sensor operates. Both answer important questions—but neither answers the most important one: What should be done next to achieve the mission objective?
Mission-level autonomy operates at the level of goals and outcomes. It translates commander or operator intent into objectives and enables systems to plan, adapt and coordinate actions across multiple assets and domains without continuous human direction.
“Mission-level autonomy answers, ‘what should be done next,’ not just ‘how do I do this step,’” Molineaux said. “It’s about handling abstraction and ambiguity—the things humans are actually good at—and giving machines the ability to reason about those things in a trustworthy, explainable way.”
That distinction becomes critical when missions deviate from plan—as they almost always do.
Why Mission-Level Autonomy Is Required Now
The pace and complexity of modern operations have outgrown centralized, human-centric control models. Adversaries move faster. Communications are increasingly contested. Decisions must often be made at the edge, where waiting for guidance can mean failure.
“We’ve reached a point where our tools are extremely powerful but not always well directed,” Molineaux said. “Being able to understand what humans really mean and what they really want before automation goes off and does something has become very important.”
In this environment, autonomy is no longer about efficiency gains. It is about resilience—maintaining operational effectiveness when people and networks cannot keep up.
Dynamic Replanning in an Uncertain World
Static autonomy assumes the world will cooperate. Dynamic replanning assumes it will not.
“Missions rarely go as planned,” Molineaux said. “Historically, we put a lot of effort into developing a really great plan at the start, but not enough into thinking about what happens when that plan doesn’t work.”
Dynamic replanning allows autonomous systems to adjust objectives, priorities and actions as conditions change—lost assets, new threats or shifting goals—rather than blindly executing an outdated plan or defaulting to failure.
“With replanning, the system can say, ‘I had a plan, but the world has changed under me. How do I adapt from where I am right now?’” he said. “That ability can turn what would have been a mission failure into a partial or even fully successful outcome.”
Designing for that adaptability is not trivial. Systems must reason under uncertainty, balance speed with safety and make decisions with incomplete or conflicting information—all while remaining aligned with mission intent.
Goal Reasoning: Understanding Intent, Not Just Instructions
Goal reasoning is what allows mission-level autonomy to function when the unexpected occurs.
“If you just give instructions—A, B, C and D—the system doesn’t know what to do when C becomes impossible,” Molineaux said. “If you give it a goal, then C is just one way of accomplishing it. If something else achieves the goal, that’s fine.”
By reasoning about goals rather than rigid task lists, autonomous systems can handle “unknown unknowns” – contingencies that cannot be anticipated – more effectively. They can assess whether new conditions threaten mission objectives, determine what options remain and adapt proportionally—without requiring exhaustive programming for every edge case.
“Goal reasoning gives the system a way to think about its future and its ability to handle the mission,” he said. “Without requiring lots and lots of special-case programming for situations that might never come up.”
Humans as Teammates, Not Overseers
A persistent concern across the autonomy community is that increased autonomy removes humans from the loop. Parallax and the Ohio Aerospace Institute take this concern to heart and build systems that cooperate with humans.
“Autonomy should extend human decision-making, not replace it,” Molineaux said. “The system absorbs complexity and speed, while humans focus on judgment, intent and boundaries.”
In practical terms, that means machines independently handle time-critical, high-volume decisions—such as sensing, routing and local resource allocation—while humans retain decisive control over intent-setting, ethical constraints, risk acceptance and mission-level tradeoffs.
Trust is central to that partnership.
“Explainability helps humans make better decisions about whether to trust,” Molineaux said. “If operators understand why the system made a recommendation or took an action, they can assess whether it’s aligned with intent. Over time, that transparency turns autonomy from something people monitor skeptically into something they rely on.”
From Research to Real Operations
Moving mission-level autonomy from the lab into operational environments requires more than technical breakthroughs. Systems must work with imperfect data, degraded networks and real-world workflows. Verification and human alignment are key.
“Autonomy must be designed for imperfect data, unreliable communications and human variability from the very beginning,” Molineaux said. “Systems that perform well in ideal conditions often fail when they encounter the friction and ambiguity of real operations.”
Parallax and OAI address that gap by starting with real mission problems and constraints—not abstract benchmarks—and iteratively testing capabilities alongside end users.
“Autonomy should be judged by whether it helps achieve mission outcomes under constraints and aligns to human needs. Technical elegance is not enough,” he said.
The Path Forward
Five to ten years from now, the most effective human–machine teams will operate less like control hierarchies and more like collaborative staffs. Humans will spend less time directing individual actions and more time shaping objectives, constraints and intent. Autonomous systems will coordinate across domains, adapt continuously and keep people informed at the right level.
The takeaway for decision-makers is clear.
“The most important principle is to invest in autonomy that understands and serves mission intent,” Molineaux said. “Prioritize adaptability, trust and integration over narrowly optimized performance. Autonomy must prove itself under uncertainty—not ideal conditions.”
Mission-level autonomy is not about removing humans from operations. It is about ensuring that when the environment is working against us, missions can still succeed.
###
About Parallax Advanced Research and the Ohio Aerospace Institute
Parallax Advanced Research is a research institute that tackles global challenges through strategic partnerships with government, industry, and academia. It accelerates innovation, addresses critical global issues, and develops groundbreaking ideas with its partners. With offices in Ohio and Virginia, Parallax aims to deliver new solutions and speed them to market. In 2023, Parallax and the Ohio Aerospace Institute formed a collaborative affiliation to drive innovation and technological advancements in Ohio and for the nation. The Ohio Aerospace Institute plays a pivotal role in advancing the aerospace industry in Ohio and the nation by fostering collaborations between universities, aerospace industries, and government organizations, and managing aerospace research, education, and workforce development projects.