What Does a GNC Engineer Do?

If you have ever watched a spacecraft land on Mars, rendezvous with another vehicle, hold an antenna steady on a target millions of kilometres away, or point a telescope at something faint and distant, a GNC engineer had something to do with it.

GNC stands for guidance, navigation and control. In many European programmes you will also hear AOCS, attitude and orbit control system. The name changes between organisations. The underlying problem does not: work out where the spacecraft is, work out where it needs to go, and make it do that reliably, with no one able to reach out and correct it by hand.

We spend a lot of our time talking to GNC engineers about exactly this, because it is a discipline that is genuinely hard to explain to anyone outside it. On a whiteboard the problem sounds simple. The spacecraft needs to point here. It needs to move there. It needs to stay within these limits. In practice you are solving a tightly coupled problem involving physics, sensors, actuators, software, uncertainty, hardware limitations and mission operations, all at once, often from hundreds of thousands or millions of kilometres away.

The three questions GNC has to answer

At its simplest, GNC answers three questions. Guidance asks where the vehicle should go. Navigation asks where it is now. Control asks what commands will make it behave as intended. ESA describes GNC and AOCS as systems that estimate the spacecraft’s state from sensor measurements, work out the control action required, and use actuators to move the spacecraft toward the desired state.

A spacecraft cannot look out of a window and steer the way an aircraft does. Depending on the mission, it may have no conventional GPS coverage, limited communications, delayed commands, imperfect sensors, and only a narrow window to correct an error. In deep space, position can be determined through ground tracking; in Earth orbit, a spacecraft may rely on GNSS, onboard orbit propagation, or other techniques entirely. The GNC system has to build an estimate of the spacecraft’s true state from imperfect measurements, then act on that estimate with confidence.

Guidance, navigation and control, in practice

A GNC engineer might work on almost any part of a spacecraft’s motion, pointing or autonomous behaviour. Day to day, that can mean developing guidance algorithms for orbital manoeuvres or atmospheric flight, designing attitude control laws, building navigation and state estimation algorithms, or working directly with the hardware that makes it all real: IMUs, star trackers, sun sensors, GNSS receivers, optical navigation systems, reaction wheels, thrusters, control moment gyros and magnetic actuators.

• Building high fidelity six degree of freedom simulations
• Analysing stability, robustness and performance
• Running large Monte Carlo campaigns
• Supporting hardware in the loop testing
• Defining requirements and verification methods
• Investigating anomalies during commissioning and operations

The important thing to understand is that GNC was never just about writing elegant equations. A theoretically beautiful controller that cannot tolerate sensor latency, actuator saturation, processor limits or unexpected spacecraft behaviour is not a good flight controller. The engineers we rate most highly are the ones who understand the whole chain, not just their corner of it.

Guidance

Guidance determines the desired trajectory or manoeuvre: an orbital transfer, a rendezvous path, an entry corridor, the commands for a precision landing. It becomes genuinely interesting once the environment turns uncertain. Fuel may be limited. The target may be moving. The atmosphere may not behave exactly as modelled. Navigation errors may already be significant. Guidance has to balance performance against uncertainty and whatever control authority is actually available.

Navigation

Navigation determines the spacecraft’s state: position, velocity, attitude, angular velocity, and other mission specific parameters. Measurements rarely arrive perfectly. Sensors carry noise, biases, scale factor errors, timing errors and occasional outright failures, so navigation engineers work extensively in state estimation, which is where techniques such as Kalman filtering become fundamental. A filter combines a model of how the spacecraft should behave with what the sensors actually report. The equations matter less than the assumptions underneath them: a filter can be mathematically correct and still perform badly if the process model, covariance assumptions, timing, coordinate frames or sensor characteristics are wrong.

Control

Control compares the desired state with the estimated one and works out how to close the gap without opening another problem. Perhaps the reaction wheels are already spinning fast. Perhaps a thruster has a minimum impulse bit. Perhaps a sensitive payload cannot tolerate the correction you would otherwise choose. Perhaps the whole manoeuvre has to land inside a narrow time window. This is where classical, modern, nonlinear and robust control all become relevant, applied to hardware that includes reaction wheels, thrusters and magnetic torquers, guided by sensors such as gyroscopes, star trackers and magnetometers.

Orbital mechanics still matters

A GNC engineer needs a strong physical understanding of spacecraft motion: two body dynamics, perturbations, orbital transfers, rendezvous, attitude dynamics, and the effects of drag, solar radiation pressure and gravity harmonics. The equations are elegant. The spacecraft is not. A real vehicle has flexible structures, imperfect actuators, sensor noise, thermal effects, delays, quantisation, software limits and propellant constraints, and hardware that behaves slightly differently from the model that described it on paper. Good GNC engineering means understanding both worlds at once, and knowing where they disagree.

What this looks like on a real mission

Mars was roughly 200 million kilometres from Earth around the time of Perseverance’s landing, and the critical descent sequence had to run autonomously. There was no possibility of real time control from the ground. The spacecraft’s Terrain Relative Navigation system solved a genuinely hard problem: it knew roughly where it was, but that was not good enough for the hazardous terrain around Jezero Crater.

That is what GNC engineering is actually for. Not simply making a spacecraft move, but changing what a mission is able to attempt in the first place.

ESA’s Rosetta mission offers a different lesson. Operating around comet 67P/Churyumov-Gerasimenko, its star trackers were used to determine orientation, but dust from the comet could appear to the tracker like a star. In 2016, when Rosetta was roughly five kilometres from the surface, its star trackers locked onto a false star caused by cometary dust, producing pointing errors that eventually triggered a safe mode event.

Your sensor can be working exactly as designed. Your algorithm can be mathematically correct. Your simulation can look excellent. And the environment can still surprise you. A good GNC engineer thinks about that possibility before flight.

The GNC loop

It helps to think of GNC as a closed loop, running continuously and sometimes many times a second: sensors feed an estimator, the estimator combines those measurements with a model to work out what is probably happening, guidance decides what should happen, the controller works out what commands are required, actuators create the physical response, and the sensors measure the result before the cycle runs again. Every part of that loop has limitations, which is why GNC engineers spend so much time thinking about interfaces, timing, failure modes and edge cases rather than just the algorithm at the centre of it.

What a GNC career tends to look like

A graduate typically starts by supporting simulation, analysis, testing or algorithm development, learning how the models behave, how the sensors actually perform, and how requirements get verified in practice. With experience, engineers take ownership of a subsystem, controller, estimator or simulation environment. Senior engineers make architectural calls, review technical work, and carry responsibility for performance and verification. At principal or lead level, the job becomes less about whether the controller works, and more about whether it is the right architecture for the mission, what assumptions sit underneath it, where the risk lives, and how you prove it will work when the spacecraft is too far away to fix.

What we look for when we recruit for GNC

The strongest GNC engineers we place combine mathematical depth with engineering pragmatism. They can derive a controller, and explain how it could fail. They understand a Kalman filter, and understand sensor timing and data interfaces just as well. They can build a simulation, and know it is only as good as its assumptions. They are comfortable with theory and comfortable close to the hardware. And, almost without exception, they are curious about what the spacecraft is actually doing, not just what the model says it should be doing.

That combination is not common, which is exactly why an experienced GNC engineer can have an outsized effect on a mission. The discipline sits at the point where mathematics becomes physical reality, and when it works, a set of equations, thousands of lines of software, a handful of sensors and actuators, and years of engineering effort make a spacecraft do something that would otherwise be impossible.