دانلود مقاله منطق فازی، شبکه عصبی-فازی و الگوریتم زنبور عسل برای حرکت رباتهای هوایی
چکیده
مقدمه
مروری بر مطالعات پیشین
راهنمای جدید برای روبات های هوایی مستقل
کاهش عدم قطعیت توسط شبکه عصبی فازی
عدم قطعیت ها
معادلات سینماتیک
بهبود موقعیت ربات های هوایی از طریق منطق فازی
نتیجه گیری
منابع
Abstract
Introduction
Literature review
Novel guidance for autonomous aerial robots
Decreasing uncertainties by neural‑fuzzy network
Uncertainties
Kinematic equations
Enhancing position of aerial robots by fuzzy logic
Conclusion
References
چکیده
در این مطالعه، یک حرکت خودکار الهام‌گرفته از طبیعت جدید با استفاده از الگوریتم زنبور عسل برای روبات‌های هوایی مورد بررسی قرار گرفت. ایده اصلی مربوط به یک قیاس جدید بین زنبورهای عسل بهینه و حرکت ربات های هوایی در ارائه راهنمایی مستقل بود. شبیه‌سازی‌های سه‌بعدی برای روبات‌های هوایی برای نشان دادن عملکرد کارآمد هدایت خودکار در نظر گرفته شد. یک سیستم معادله جدید نیز بر اساس کنترل زاویه انحراف برای ساده کردن محاسبات پرواز پویا توسعه داده شد. علاوه بر این، عدم قطعیت های مختلف مانند جریان باد جانبی و نویز ناوبری دقیقاً با استفاده از یک شبکه عصبی-فازی برای افزایش قابلیت اطمینان هدایت خودکار در نظر گرفته شد و بررسی شد. بر این اساس، حرکات خودمختار رباتهای هوایی با منطق فازی برای غلبه بر ارتباطات داده‌ای با کیفیت پایین بین روبات‌های هوایی توسعه داده شد. نتایج این مطالعه نشان داد که هدایت یکپارچه الهام گرفته از طبیعت توسط منطق فازی برای ربات های هوایی به ترتیب گذر کلی کمتر و زمان نهایی 24.64% و 21.87% داشت.
توجه! این متن ترجمه ماشینی بوده و توسط مترجمین ای ترجمه، ترجمه نشده است.
Abstract
In this study, a novel nature-inspired autonomous motion was investigated using the honey-bee algorithm for aerial robots. The main idea belonged to a novel analogy between optimal honey bees and aerial robots’ motion in proposing autonomous guidance. Three-dimensional simulations for aerial robots were considered to show the efficient performance of autonomous guidance. A new equation system was also developed based on the yaw angle control to simplify dynamic flight calculations. Moreover, different uncertainties such as lateral wind current and navigation’ noise were considered and checked precisely using a neural-fuzzy network to enhance autonomous guidance reliability. Accordingly, aerial robots’ autonomous motions were developed by fuzzy logic to overcome low-quality data linkages between aerial robots. The results of this study illustrated that the integrated nature-inspired guidance by fuzzy logic had a lower total passing and the final time of 24.64% and 21.87% for aerial robots, respectively.
Introduction
Autonomous aerial robots with closed-loop guidance are of paramount importance for civil missions. Further research and ideas are required for conventional guidance methods to develop the next generations of aerial robots. The usual open-loop guidance methods are underpinned by the ground station; therefore, the reduction of ground station commands by novel intelligent methods is essential for the next aerial robots’ development. Moreover, many studies have examined closed-loop guidance using artifcial intelligence methods to remove uncertainties (Al-Rabayah and Malaney 2012; Babaei et al. 2011; Bernsen and Manivannan 2012; Bitam et al. 2013; Chen et al. 2016, 2017; Dadkhah and Mettler 2012; Eng et al. 2010). The problem statement, the main contributions and the motivation of this paper are as follow. The main problem of this paper is belonged to propose a novel nature-inspired autonomous motion of arial robots with artifcial intelligence. The idea of this paper represented a novel analogy between optimal honey bees and aerial robots’ motion in proposing autonomous guidance. Moreover, to explain the motivation of this paper, new equation system was also developed based on the yaw angle control to simplify dynamic fight calculations. Furthermore, diferent uncertainties such as lateral wind current and navigation’ noise were considered using a neural-fuzzy by training a network to develop autonomous guidance reliability. Also, aerial robots’ autonomous motions were enhanced by fuzzy logic to overcome low-quality data linkages between aerial robots as a closed-loop guidance method.
Conclusion
This study focused on a nature-inspired autonomous algorithm for multiple aerial robots. This algorithm was considered not only as a novel nature-based autonomous guidance but also as an autonomous method. Considering a decrease in fight mechanics’ calculations, a new equation system was proposed based on the derivatives by the yaw control angle. Moreover, the neural-fuzzy network as one of the powerful methods was applied to decrease diferent uncertainties such as lateral wind current and navigation’ noise. The neural-fuzzy network was proved to be highly efcient in estimating, approximating, and increasing reliability. To implement the intelligent neural-fuzzy system, several fight scenarios are selected. Accordingly, three fight scenarios were selected with close boundary conditions here. The yaw control angle accomplished the dynamic system training by artifcial intelligence as the neural-fuzzy network output. The proposed neural-fuzzy network could deliver the aerial robots by high reliability and consider uncertainties from the initial point to the path’s endpoint. According to the results, aerial robots’ precision in arriving at the target increased from 31.62 to 67.32% for diferent lateral wind currents and navigation’ noise. Moreover, aerial robots’ autonomous motions were improved by the fuzzy logic to develop information linkages between aerial robots with a lower total passing and fnal times of 24.64% and 21.87%, respectively. In sum, autonomous aerial robots could be designed intelligently using a novel combination of the bee algorithm, neural-fuzzy network, and fuzzy logic.