Task 5

The actual heading is subtracted from the commanded one. The perturbation affects both the rudder and the commanded bank angle via proportional gains. The actual bank angle is subtracted from the commanded one and commands an aileron input via an additional proportional gain. Together, the adjusted rudder input and bank angle change the heading to bring the aircraft back towards its commanded value.

The numeric values for the gains chosen are:

- K_R = 1 (degrees of rudder deflection per degree away from desired heading)

- K_Phi_c = -0.7 (degrees of bank angle desired per degree away from desired heading)

- K_a = 1 (degrees of aileron deflection per degree away from desired bank)

Supporting Experimental Data Using the Bixler 2

In the following, three of the autonomous test flights are presented and analyzed. As can be seen below, steady state errors are present due to improper steady state trim values. The controls team was not severely concerned by this as the main goal was to be able to maintain heading and altitude over time, rather than worry about minimizing steady state errors as trims can be adjusted easily and do not have a large effect the robustness of the control law.

The plots show altitude, heading, airspeed, and when the autopilot was engaged, in that order.

In general, what can be seen is an approximately 5m rise in altitude from the initial start of the auto-flight (likely due to too large of a level flight trim throttle). Once this initial rise to the steady state settling position is completed, the altitude is maintained within about +/- 3m.

For heading, a similar error to altitude can be noted. Typically the steady state settling value is increased by about 10 degrees from the initial autopilot engagement, then once settled it holds to within +/- 10 degrees.

Airspeed was indirectly determined by the desired pitch angle, and since it was not closed separately was not held to the value at autopilot engagement. Rather, it typically converged to approximately 8 m/s +/- 1.5 m/s for every flight.

It should be noted that the testing of the autopilot was done in imperfect conditions, causing numerous perturbations. It is quite likely that the range of error after steady state is achieved could be significantly reduced by flying in more ideal conditions. Another attempt was planned but ended up being canceled due to scheduling conflicts and the equipment being moved to the Sparrow 1.

Controls Test 1

Controls Test 2

Controls Test 3

This data proves that our basic control law is successful in maintaining a steady direction and altitude with no human input, and serves as a baseline for our further development. Once the airspeed loop is closed and more accurate trims determined on the Sparrow 1, the controls team believes that this basic control law which can be developed almost entirely with experimental data will allow for rapid closure of navigation loops.