Piloted Simulation and Flight Testing

Abstract

The control law went through several periods of flight simulation testing by both trained and untrained pilots. Piloted simulation proved very important in the evolution of the control law, and showed up effects not discernible from non-piloted simulation. Some of these effects are as follows:

• 1. Pilot in the loop. When the pilot is flying the controlled aircraft he closes the loop between his inceptors and what he is attempting to control. For most tasks the pilots’attention is primarily directed towards controlling flight path angle (or height rate if in the hover) which is fed back to the pilot by the symbol in the centre of the head-up display. It is actually possible for the pilot to destabilise the system if the lag in the controlled system is too high. This effect is commonly referred to as pilot-induced oscillation (PIO).

• 2. Head-up display format. The form of the head-up display (HUD) and the information presented have a large impact on the assessment of the control law. For example, the original HUD used for jet-borne flight caused problems in that both pitch attitude and flight path are shown simultaneously. This meant that it was very easy for the pilot to pick up the wrong symbol when changing from the wing-borne flight HUD in a deceleration. The pitch attitude symbol was removed to resolve this problem, it being superfluous given that the control law sets the attitude for landing automatically. The precision to which variables are displayed was also found to be very important; if airspeed is displayed to the nearest knot, then there is a tendency to try and set the speed to this accuracy, even though this would not be normal practice.

• 3. Lateral manoeuvres. Lateral manoeuvres led to three compensation terms being added to the longitudinal control law. Feedforward was added from bank angle to airspeed demand in wing-borne flight to ensure no airspeed is lost. Feedforward was also added from bank angle to fuel flow demand when in the hover so as to prevent any height loss following a rapid bank demand. The prioritized desaturation scheme was also adjusted for engine saturation at large bank angles, backing off vertical speed demand if necessary.

• 4. Stick dynamics. The stick and hydraulic feel system have a relatively large lag which affected the perceived response of the control law. In particular the lack of a well defined centered stick position required a large dead-band to be implemented which could lead to difficult height rate capture in the hover.

• 5. Blend region. The speed region over which the blend between control modes is effected was altered several times. There was a trade-off between allowing the pilot to use aerodynamic lift down to as low a speed as possible, and ensuring that the aircraft stays within safe operating limits in terms of angle of attack and maximum descent rates on approach to hover.

• 6. Engine non-linearity. In the course of flight simulation a small engine limit cycle was found around the knee in the throttle-fanspeed characteristic. The size of this can be minimised by adjusting the model of this characteristic within the control law i.e. by adjusting ENJB and ENJD. However, the effect cannot be completely removed since the point at which the engine governor cuts in cannot be precisely predicted.

• 7. Flight path quickening. The HUD used for piloted simulation had flight path quickening. Whilst beneficial in terms of improving the perceived tracking performance with the pilot in the loop, it has a detrimental affect on the flight path hold facility. For instance, if the pilot tries to capture 5 degrees up flight path he will centre the stick when he sees the flight path symbol on the HUD reach 5 degrees up. With the flight path quickening the HUD symbol will be ahead of the actual flight path which in turn means the captured flight path will be some value less than 5 degrees. In piloted simulation the effect is that the flight path symbol drifts down to a lower value. A solution to this could be to use the quickened value of flight path to drive the flight path hold.

• 8. Speed hold authority. The authority of the speed hold is on the low side in wingborne flight. This is partly due to the limitations imposed by the engine non-linearity on bandwidth. An extra integrator might also help, but this was not done as the order of the controller would increase by 1, and the available processor power is insufficient to cope with this. Hence feedforward terms are used from bank angle to speed demand to help maintain airspeed in steep turns.