Efficient Reconfiguration and Recovery from Damage for Air Vehicles

Abstract

In order to increase survivability and maximize performance, autonomous vehicles will require the development of algorithms that fulfill the role of an adaptive human pilot in response to failures, damage, or uncertain vehicle dynamics. The guidance and control algorithms implemented on autonomous vehicles must therefore be able to react and compensate, whenever possible, for failures so that the impact of a failure can be minimized. A great deal of progress has been made over the past decade in the development of trajectory reshaping algorithms, adaptive guidance and reconfigurable control; however, much of this work is dependent upon future developments in integrated vehicle health management (IVHM). This manuscript will highlight how some recent guidance, control, and trajectory reshaping methods rely on fault detection and isolation (FDI) and IVHM capabilities to respond to failures or damage to the vehicle. These requirements are outlined here to communicate the needs of the GN&C community to the IVHM/FDI community. In this work, a number of methods are discussed that are designed to enable an autonomous air vehicle to recover from failures/damage. The requirements imposed on fault detection and isolation (FDI) systems or integrated vehicle health management systems (IVHM) are highlighted as each method is discussed. The first technique described accounts for control effector failures and utilizes three layers of reconfiguration. The starting point is a reconfigurable inner-loop control law, where control effectors are re-mixed to account for a locked or floating effector. The control effector failure information must supplied by an FDI or IVHM system. The reconfigurable inner-loop control law is based on dynamic inversion with explicit model following coupled with an optimization based control allocator. The dynamic inversion control law 1 requires the use of a control effector allocation algorithm, if the number of control effectors exceeds the number of controlled variables or if actuator rate and position limits must be enforced. Illustrative examples of the methods will be used throughout the paper and will make use of a vehicle that has 6 control surfaces, namely, left and right ruddervators, left and right flaperons, speedbrake, and bodyflap. Because there are 6 control surfaces and only 3 axes to control, it is possible that the desired control variable rate commands can be achieved in many different ways and so a control allocation algorithm is used to provide a unique solution to such problems. 2, 3 To complete the inner-loop, prefilter blocks are designed to produce desired closed-loop dynamics. An explicit model-following prefilter scheme is used as an example to show how the inner-loop bandwidth can be adjusted by modifying the bandwidth of the explicit model when all control power is exhausted in one or more axes. The second layer of compensation for a failure, namely, guidance adaptation, becomes active when this situation occurs (called axis saturation or control power deficiency). When control power deficiency is detected, the bandwidth of the reference model in the explicit model following ∗ This material is declared a work of the U.S. Government and is not subject to copyright protection in the United States. † Electronics Engineer, 2210 Eighth Street, Bldg. 146, Rm. 305, Ph. 937-255-8490, Email