ردیابی مسیر فرود هواپیمای مدیریت عملیات با استفاده از سیستم استنتاج فازی عصبی تطبیقی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|7892||2008||6 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Computers & Mathematics with Applications, Volume 56, Issue 5, September 2008, Pages 1322–1327
The adaptive neural fuzzy inference system is used to simulate trajectory tracking in aircraft landing operations management. The advantage of the approach is that by using the linguistic representation ability of fuzzy sets and the learning ability of neural networks, the approximate linguistic representations can be improved or updated as more data become available. This approach is illustrated by the use of both zero and first order Takagi–Sugeno inference systems [T. Takagi, M. Sugeno, Fuzzy identification of systems and its application to modeling and control, IEEE Transactions on Systems, Man, and Cybernetics 15 (1) (1985) 116–132] with auto-landing flight path data.
A major challenge in modern aircraft design is the controllability of the flight vehicle. When an airplane is in the air, not all atmospheric factors acting on it can be controlled by human efforts. The trajectory tracking of a landing plane is a useful approach but is an unpredictable and sophisticated process. It is difficult to devise an approach that allows modern control theory to deal with this nonlinear control problem in a systematic way. Various investigators have studied this control problem. For example, Al-Hiddabi and McClamroch  used nonlinear control theory in developing a controller for conventional aircraft take-off and landing based on trajectory tracking control. Liu and Harmon  presented a real-time control and simulation investigation for aircraft landing requiring stability and robustness with respect to variations in speed, weight, center-of-gravity and time delays. Zou and Devasia  demonstrated the use of a previous-based stable-inversion technique in online output-tracking. Shan et al.  proposed a synchronized trajectory-tracking control strategy using a feed forward term and a PD feedback control term for multiple experimental three-degrees-of-freedom helicopters. Fuzzy logic and/or neural network systems have also been applied to solve aircraft control problems. Pistauer and Bernhardi-Griasson  proposed a method for the design and implementation of a helicopter flight mechanic model with a specific fuzzy system structure. Wood and Schneider  used fuzzy expert system tools to control anti-submarine helicopters. Gharieb and Nagib  presented a general hierarchical fuzzy control design for a multi-variable helicopter system. Jorgensen and Schley  described a simplified neural network baseline model for aircraft control. Juang and Chio  presented an aircraft landing control based on fuzzy networks. McMichael et al.  examined the combined application of fuzzy methods and genetic algorithms in flow control. Melin and Castillo  described a hybrid method for adaptive intelligent control of aircraft systems. The hybrid approach was obtained by the combined use of neural network, fuzzy logic and fractal theory. In this paper, the adaptive neural fuzzy inference system (ANFIS) is proposed for trajectory tracking in the nonlinear control of aircraft landing. Specifically, only longitudinal stability and control analysis are considered; and the open-loop pitch angle or the pitch angle of the elevator (PAE) is considered as the output of the system. The fuzzy neural inference system will be summarized in Section 2 and the application of this ANFIS system to trajectory tracking with some numerical results will be presented in Section 3. The numerical auto-landing flight path data used are obtained from Shen et al.  and . Finally, Section 4 presents some discussion.
نتیجه گیری انگلیسی
The adaptive neural fuzzy inference system has been shown to be very useful in many different areas where the data are vague and may be even linguistic. This is because of the combined advantages of both the linguistic representation ability of fuzzy sets and the learning ability of neural networks. Thus, at the beginning, the data can be represented approximately by fuzzy membership functions and these approximate representations can be improved or updated as more data become available. It should be noted that many useful systems such as in manufacturing or in control are vague and frequently need human experts to operate. These experts can only give data linguistically. Using auto-landing flight path data, trajectory tracking of aircraft landing was simulated by the use of the adaptive neural fuzzy inference network. Since during trajectory tracking, many of the atmospheric influencing factors are uncertain and changing rapidly, especially under severe weather conditions, nonlinear adaptive control with the ability of learning appears to be one of the promising approaches. Even with these limited results due to limited data, this simulation approach by the use of the fuzzy adaptive network appears to be a promising approach. The convergence rates are fairly fast even with a very small number–3 to 6–of membership functions.