FANAT is not an evolution of existing solutions. It's a new flight control architecture built from scratch on physical modelling instead of empirical approximation.
Most flight controllers use averaged coefficients for sensor groups and generic dynamics models for different frames. It works. But it's a trade-off between compute speed and model accuracy.
FANAT builds individual matrices for each sensor. Models frame geometry as part of the dynamic system. Computes the full equation set in real time. No simplifications. With code optimization.
Every accelerometer has its transfer coefficient. Every gyroscope has its temperature drift. Every magnetometer its nonlinearity. Ignoring that means losing accuracy before filtering even starts.
FANAT calibrates each sensor separately. Builds an individual transfer function. Accounts for temperature. Result: the system works with real data, not a "average sensor" approximation.
Frame geometry defines the matrix that turns control signals into moments. Distance from CoG to motor. Arm angle. Mass distribution. That's not "tuning". It's parameters of the physical model.
Not a set of modules. A single system where every element follows one philosophy: physics instead of approximation.
Flight controller. Processor, IMU, magnetometer, barometer. Per-sensor factory calibration.
Remote inertial unit. Extra accelerometers and gyros on frame arms. Deformation and vibration measurement.
Telemetry. Two-way link. Sensor data downlink. Command uplink.
Navigation module. GPS, GLONASS, Galileo, BeiDou. RTK for centimetre accuracy.
Motor ESCs. CAN bus to controller. Telemetry: current, voltage, temperature, RPM.
Power system. Battery monitoring. Load distribution. Overload protection.
Operating system. Manages all modules. Mission planning. Data logging.