International Journal of Engineering and Technical Research (IJETR) ISSN: 2321-0869, Volume-2, Issue-12, December 2014 185 www.erpublication.org Abstract__In this research, novel algorithm based on the optimal control theory is proposed for Samarai micro aerial vehicle optimal guidance policy. Considering wind effect in the system dynamic equation increase the robustness of this optimal guidance. Open-loop optimal control is obtained regarding to novel and new proposed method. In this way, intelligent methods such as GA-PSO optimizer and neural-fuzzy are utilized in the proposed new algorithm. Results of new algorithm are compared with pseudo-spectral optimal control solver and show high accuracy. Closed-loop guidance not only damped noises but also simplify controller performance for this unstable vehicle. Next, closed-loop optimal guidance base on neural-fuzzy method is proposed to achieve autonomous guidance to increase the stability of this unstable micro vehicle versus wind effects. Keywords__ Optimal Control; Wind Effect; Autonomous Guidance; GA-PSO Optimization; Neural-Fuzzy I. INTRODUCTION Rescue vehicles for earthquake to get information from damages or generally civil applications of unmanned air vehicles have increased in recent years. Small unmanned air vehicles (SUAV) are used in civil projects where the focus is on autonomous vehicles; however, MAVs like Samarai make use of radio controls [1]. Lockheed Martin's Intelligent Robotics Laboratories has spent the last five years to develop an unmanned micro aerial vehicle to replicate the motion. The idea was based on maple seed; the seeds that drop from maple trees, whirling softly to the ground like silent one-winged helicopters, Therefor, these air vehicles are the inspiration for a new kind of flying machine that could be useful for military and civil information-gathering missions. Lockheed Martin Advanced Technology Laboratories (ATL) developed the Samarai MAV, a 30- centimeter-radius maple seed like aircraft that can take-off/land vertically and fly laterally (like a helicopter) to the intrinsic stability of nature’s maple seeds [2]. Dynamic systems solutions in the optimal control framework can be classified as the two main categories of open-loop and closed-loop. Open-loop solution is proper, if there are no un-modeled disturbances and/or process noises. Unlike open-loop optimal controls, closed-loop optimal controls are considered as functions of states. Manuscript received December 21, 2014. M. Tafreshi, Department of Aerospace engineering, K. N. Toosi University of Technology, Tehran, Iran, +989124862457. I. Shafieenejad, Department of Aerospace engineering, K. N. Toosi University of Technology, Tehran, Iran, +989123465637. A. A. Nikkhah, Department of Aerospace engineering, K. N. Toosi University of Technology, Tehran, Iran, +989123875213. Hence, closed-loop optimal controls increase robustness of the vehicle against noises and/or undesirable disturbances exerted on the system. Of course, it is usually difficult to achieve closed-loop optimal controls. Therefore, finding a method with robust characteristics to overcome this difficulty is highly desirable. Neural-fuzzy combined with the highly transparent and compact form of fuzzy rules makes dynamic systems candidate against disturbances and/or noises as closed-loop method for autonomous vehicles [3, 4]. Millary and Zachari applied the theory of partially Markov decision processes to design guidance algorithms for the motion of unmanned aerial vehicles. They used on-board sensors for tracking ground targets [5]. Han and Bang investigated proportional navigation guidance to avoid collision based on the optimal method [6]. In 2006, autonomous operation was demonstrated for unmanned air vehicle by Ma and Stepangan [7]. Also, autonomous guidance system based on receding horizon optimization was described by Mettler and Dadkhan [8]. In 2010, Paw and Balas presented an integrated framework for small unmanned aerial vehicles’ flight control development. Moreover in reference [9], software-in-the-loop and flight testing are conducted with a synthesized controller [9]. There has not been complete investigation into MAVs like Samarai with one wing and radio controller that made by Lockheed Martin's Intelligent Robotics Laboratories [2]. Therefore, Kellas investigated the design and development of controllable single-blade autorotation vehicles in 2007. Simulation results were examined to provide insight into selecting the best control concept and hardware for the final guidance of Samarai design. The free-flight simulation results predicted approximately 10% of experimentally observed performance while the coning angle was predicted to be 25% of the observed angle [1]. An efficient strategy was proposed by Babaei and Mortazavi to design the autopilot for a UAV which was non-minimum phase, and its model included both parametric uncertainties and unmodeled nonlinear dynamics. Babaei’s work had been motivated by the challenge of developing and implementing an autopilot that was robust with respect to these uncertainties. By combination of classic controller as the principal section of the autopilot and the fuzzy logic controller to increase the robustness a new methodology was developed [10]. Challenges uniquely associated with developing this type of vehicle are identified and a dynamic modeling and control synthesis procedure [11]. Therefore, this paper focuses on the optimal control and designing closed-loop trajectories. Open-Loop and Closed-Loop Optimal Guidance Policy for Samarai Aerial Vehicle with Novel Algorithm to Consider Wind Effects M. Tafreshi, I. Shafieenejad, A. A. Nikkhah