Adaptive maneuver control of hypersonic re-entry flight via self-organizing recurrent functional link network

摘要

The maneuver control of hypersonic vehicles (HSVs) during re-entry is a challenging work due to the object features of serious nonlinearity, strong uncertainty and fast time variation. In this paper, we design the maneuver control architecture of lateral turning for the HSV first, and then a new self-organizing recurrent functional link network (SORFLN) is proposed to estimate the dynamical uncertainties/disturbances in flight. The SORFLN-based nonlinear controller is presented to compensate for the effect of uncertainties. The training algorithm to grow the SORFLN and adapt the parameters is derived from Lyapunov theory. The pruning strategy to keep the SORFLN size as small as possible is put forward based on the output features of the nodes. Finally, the simulation results show that the presented method can achieve satisfactory control performance and the uncertainty/disturbance rejection is successfully accomplished with small size of the network.