Introduction to aircraft controls
Many researchers would agree that the first steps into experimental aircraft control analysis are sometimes the hardest because young researchers must close the gap between the theoretical knowledge and the know-to-how on the praxis.
The main difficulties in a practical set-up involve determining the required instrumentation, programming suitable communication protocols with the data acquisition system, and calibrating the integrated sensors, as well as the aircraft dynamic modeling prior to the control design. The following presents a brief summary on how to handle the most common technical difficulties faced when taking the first steps into fixed-wing aircraft controls.
Test bench
Test benches are designed to simulate aircraft motion within controlled environments. This should be the first step rather than facing the uncertainties of real-world operation (modeling uncertainties, wind gusts, etc). With this purpose, wind tunnels represent the ideal base for aircraft control test-bench development.
Consider Figure 1 as an example of a wind-tunnel test bench for aircraft controls, the platform allows simulating the pitch dynamics of a fixed-wing aircraft by taking in account the horizontal stabilizer deflection. An important remark is that the pivoting point should always be located near the aircraft’s center of gravity. The deflection of the horizontal stabilizer is used to maintain flight conditions or perform maneuvers; it induces the aircraft to climb, descend and obtain sufficient lift from the wings to keep the aircraft in level flight at various speeds. Thus, allowing to simulate aircraft motion during characteristic operation.
Avionics
In aviation, whether it corresponds to commercial, civil, or military applications, most aircraft must accomplish stability and performance relying only on the sensors aboard, usually comprised by inertial measurement units. A popular solution to measure both acceleration and angular velocity is the MPU6050, which has been tested with success for many applications. This IMU is compatible with the I2C communication protocol, easily achievable with low-cost processing boards such as Arduino, Raspberry Pi, and Beaglebone. Moreover, these devices can be programmed through a simple Matlab interface with favorable results, achieving a simple but powerful solution to the avionics determination. However, IMU measurements inherently include different errors for attitude estimation, such as unbounded steady-state bias, high-density noise spectrum, and sensitivity to external vibrations. These situations may hardly degrade the system’s performance. That is:
“The controller performance and robustness will only be as good as the information provided by the sensor/estimator scheme”
Sensor fusion
In order to improve the attitude estimates, sensor fusion algorithms are implemented. Experimental tests have demonstrated that a properly tunned Luenberger observer can provide estimations as fast as a Kalman filter, with the same robustness and fewer computational requirements. Moreover, these sensor fusion algorithms are compatible with well-known linear control and stability analysis tools. Thereafter, the problem relies on calibrating the Luenberger observer (or commonly called complementary filter in these applications).
The article [1] presents a reformulation that allows designing the Luenberger observer gain (L) as a classical sensitivity (S) and complementary sensitivity (T) perturbation and noise rejection, achieving optimal estimates by tunning the pole of a transfer function. A recommended value is also provided within the article.
[1] Valderrama, J. F. V., Takano, L., Liceaga-Castro, E., Hernandez-Alcantara, D., Zambrano-Robledo, P. D. C., & Amezquita-Brooks, L. (2020). An integral approach for aircraft pitch control and instrumentation in a wind-tunnel. Aircraft Engineering and Aerospace Technology.
Original link: https://www.emerald.com/insight/content/doi/10.1108/AEAT-10-2019-0193/full/html
