REAL-TIME PARAMETER-IDENTIFICATION APPLIED TO FLIGHT SIMULATION

In-flight simulations are normally accomplished using model-following control laws which depend on accurate knowledge of the stability deriv atives of the host aircraft. Degraded simulation results if the stabil ity derivatives deviate considerably from their presumed values. Gain scheduling is often employed to compensate for plant parameter variati ons, but this form of open-loop compensation usually requires extensiv e flight testing for proper fine tuning. An adaptive, fast-sampling co ntrol law to compensate for changing aircraft parameters is described. The step-response matrix required for implementation is identified re cursively using a technique which does not need special ''test'' signa ls and which automatically discounts old data depending on the input e xcitation detected. Tracking fidelity is maintained despite parameter changes which occur either abruptly or slowly, and actuator position a nd rate limiting are discussed. The performance of the resulting syste m is excellent and demonstrates the relative advantages of adaptive co ntrollers for in-flight simulation.