Leonid S. Zhiteckii et at. Int. Journal of Engineering Research and Application www.ijera.com ISSN : 2248-9622, Vol. 6, Issue 10, ( Part -4) October 2016, pp.70-79 www.ijera.com 70 | Page Design of Digital Autopilot for Lateral Motion Control of an Aircraft Based on l 1 -Optimization Approach Leonid S. Zhiteckii, Valerii N. Azarskov, Andriy Yu. Pilchevsky, Klavdia Yu. Solovchuk ABSTRACT The optimal digital autopilot needed to control of the roll for an aircraft in the presence of an arbitrary unmeasured disturbances is addressed in this paper. Such autopilot has to achieve a desired lateral motion control of this aircraft via minimizing the upper bound on the absolute value of the difference between the given and true roll angles. It is ensured by means of the two digital controllers. The inner controller is designed as the discrete-time PI controller in order to stabilize a given roll rate. This variable is formed by the external P controller. The necessary and sufficient conditions under which the two-circuit feedback discrete-time control system will be stable are derived. To optimize this control system, the controller parameters are derived utilizing the so-called l 1 -optimization approach advanced in modern control theory. A numerical example demonstrating the l 1 -optimization technique and results of some simulation experiments are presented to illustrate the performance of the l 1 -optimal controller. The robustness properties of this controller are established. Index Terms-Aircraft, Lateral Dynamics, Digital Control System, Discrete Time, Stability, l 1 -Optimization, Random Search Algorithm, Robustness. I. INTRODUCTION The problem of efficiently controlling the motion of an aircraft in a non-stationary environment capable to ensure its high performance index is important enough from the practical point of view and remains actual up to now [1], [2]. To solve this problem, different approaches based on achievements of the modern control theory, including adaptive and robust control, neural networks, etc., have been reported by many researches (see, for example, [3] – [9]). Recently, new results in this research direction have been presented, in particular, in [10]. Unfortunately, most of these works dealt with an ideal case when there are no external persistently exciting disturbances. Nevertheless, they are always present in reality. To implement approaches advanced in modern control theory, digital technique is appropriate. Point is that, by the end of the twentieth century, digital control has become a highly developed technology in control applications [11], [12]. Digital control systems have some features associated with sampling [12], [13]. Namely, it leads to arising the discrete-time system description. It turns out that accurate discrete-time models can be derived for sampled continuous-time systems under digital control [14]. One of the efficient methods devised in the modern control theory for rejecting any unmeasured disturbance is based on the so-called l 1 - optimization concept [15] – [17] applicable to discrete-time control systems. This concept has been utilized in [18] to design the digital lateral autopilot for an UAV capable to cope with a gust.In order to implement the l 1 -optimization of any digital controller, the information with respect to the dynamics model of a plant to be controlled including its structure and parameters is required. In practice, however, it may not be available in full detail. In this real situation, the following question naturally arises: is the l 1 -optimal PI controller designed via the use of a priori knowledge of the so-called nominal lateral dynamics model robust? This paper extends the approach which we have first reported among other authors in [18] to deal with a digital autopilot for the lateral motion control. Its aim is to synthesize a digital autopilot which is able to maintain a given roll orientation of an aircraft with a desired accuracy and to reject an arbitrary external disturbance (in particular, the gust). Again, this controller needs to be robust with respect to parametric and nonparametric uncertainties. As in traditional continuous-time (analogue) control systems [5], the digital control system is designed as the two-circuit closed-loop system having the inner feedback loop and the external feedback loop. The distinguishing feature of the digital autopilot is that it is designed as the so-called l 1 -optimal controller containing the discrete-time PI and P controllers. The novel contribution of this paper includes the following: 1. The l 1 -optimal PI and P controller parameters of the digital autopilot are calculated simultaneously (in contrast with [18]). 2. The aileron servo dynamics is taken into account to derive the stability condition for l 1 - optimizing the controller parameters. RESEARCH ARTICLE OPEN ACCESS