PATH PLANNING FOR LOW-ALTITUDE FLIGHT BY FUZZY-REASONING

This paper discusses flight-path planning for an aircraft flying at lo w altitude while avoiding terrain crashes. In a past study, several al gorithms based on exploratory procedures have been proposed; however, they do not generate practical paths that take into account aircraft i nertia. This paper proposes a new algorithm to generate a flight path taking account of the aircraft's inertia; this algorithm does not gene rate the flight path directly, but determines the aircraft's steering acceleration at every point along the way and then generates the fligh t path based on the steering acceleration. This algorithm extracts fea ture parameters representing a relation between the aircraft and the t errain, determines steering acceleration by fuzzy reasoning using the feature parameters, and updates state variables of the aircraft. Furth ermore, this paper also proposes a learning algorithm to design the me mbership values in the fuzzy reasoning from reference flights since it is difficult to optimize them by trial and error. The flight-path pla nning by the proposed algorithm is compared with that based on the exp loratory procedure. The learning algorithm of the membership values is validated through computer simulations.