Beneath the curves of the airframe, beneath the roll and climb, there is a deeper shape that governs how the aircraft behaves.
Not just a function. Not just a system of equations. But a geometry of response—partly linear, but with a built-in shift.
This is the domain of the affine formulation.
To understand how aircraft move—especially under control—we often start with linear models. They give clarity. They give structure. But in the real world, control forces don’t always emerge from a clean origin. Disturbances, biases, and gravity itself create a permanent offset, a baseline motion even when inputs are zero.
And this is where affine models become essential.
An affine system is like a linear system, but translated. It includes a constant term—a vector offset—that captures the behavior of the system when no input is applied. In formal terms, it takes the form:
ẋ = A(x)x + B(x)u + c(x)
Here, the term c(x) is the shift. It reflects things like gravitational pull, aerodynamic trim conditions, or baseline torque—forces that act even when the control input u is zero. The system doesn’t pass through the origin. It drifts, and the controller must correct for that drift.
In aircraft modeling, this is crucial.
Imagine an aircraft trimmed for steady cruise. Its velocity and orientation are nonzero, but unchanging. The forces holding it in balance—lift, drag, thrust, weight—don’t cancel at zero input. They cancel at some operating point. Affine models let us expand the dynamics around that point, acknowledging that flight isn’t born from stillness, but from a constant dance of offset forces.
By working with affine formulations, control designers can:
– Linearize around a realistic equilibrium, rather than the origin.
– Incorporate trim conditions directly into the model.
– Design controllers that respect persistent dynamics, such as constant drag or lift-related offset forces.
Affine systems preserve much of the structure of linear systems. They can still be stabilized, observed, and controlled using linear tools—with added care to handle the offset. In Model Predictive Control, in particular, affine formulations are often used to build prediction models that reflect the true resting behavior of the aircraft.
And in doing so, they give the control system something precious: a clearer picture of reality.
Because real aircraft don’t rest at zero.
They fly balanced on edges—between thrust and drag, between lift and fall.
Affine models let us speak to them in their language.
They are not more complex for the sake of complexity.
They are more honest.
And when we model truthfully,
our control becomes not only accurate—but aligned with what flight really is.