High Performance Disturbance Observer based Control of the Nonlinear 2DOF Helicopter System Almir Salihbegovic, Emir Sokic, Nedim Osmic and Mujo Hebibovic Department for Automatic Control and Electronics Faculty of Electrical Engineering, University of Sarajevo 71000 Sarajevo, Bosnia and Herzegovina almir.salihbegovic@etf.unsa.ba Abstract— This paper addresses the challenges of the distur- bance observer (DOB) algorithms faced with highly nonlinear electromechanical systems which are dealing with high resolu- tion and high speed operations. It describes the synthesis of ro- bust and stable controllers and their applications in controlling azimuth and elevation angles of the helicopter model CE 150 supplied by Humosoft. Description of the helicopter, including its mechanical characteristics and mathematical model, is given in the paper. Tracking error, transient performances, power consumption and motor strains are used for the validation of control quality. Implementation of the control system on the experimental setup is also explained. MATLAB and Simulink are used as tools for developing the simulation model of the helicopter system. Obtained simulations are showing that developed controllers provide significantly improved results even in the presence of unknown and unpredictable inputs (disturbance and noise), unpredictable and unknown dynamics, external forces (torques) and change of the system parameters. I. I NTRODUCTION Modern electromechanical systems are often required to operate at high speeds to yield high productivity. Precision and accuracy requirements are becoming more and more strict at the same time. Advanced control plays a significant role while meeting these challenges. The helicopters are widely used in transportation, air surveillance and as combat aerial vehicles. They are very interesting from the control point of view due to nonlinearity, instability in the open-loop and high cross-coupling effects. Main difficulties in controlling such systems are nonlinear friction, uncertainties of the systems parameters, unmodelled dynamics and external disturbances. In the last two decades nonlinear control methods for the the nonlinear systems have been intensively developing [1]. For highly nonlinear systems usage of the classical control theory (PI, PD and PID controllers) is not recommended. These controllers give satisfactory results only in a very small area around the set point. The control performance could be improved using an output tracking based on approximate linearisation [2], [3]. This approach neglects cross-coupling effects of the helicopter. The model of a small unmanned helicopter has been linearised at different set points along the elevation axis and a gain scheduling control has been implemented in [4]. Many authors have been studied methods for controlling azimuth and elevation angles of the helicopter model CE 150 supplied by Humosoft [5]–[9]. The performance comparison of three optimal control techniques to a helicopter system: model predictive control (MPC), linear quadratic optimal control combined with a state estimator (LQG), and optimal linear quadratic output control (PLQ), is discussed in [8]. These three schemes obtained significantly improved results over classical control algorithms. However, the application of these algorithms is not trivial due to demand for frequent model linearisation and significant random disturbances. MPC has substantial inter-sample computation demands and the largest memory requirements [9]. LQG and PLQ algo- rithms proved satisfactory results only below the horizontal line ψ =0. Above this angle the instability of the open loop plant and increasing model-plant mismatch leads to poor tracking results. Papers [5]–[7] describe capabilities of the inteligent meth- ods for controlling 2DOF nonlinear helicopter model. The fuzzy control in dealing with helicopter uncertainties is described in [5]. The capability of neural networks scheme to control laboratory helicopter model CE 150 was discussed in [6]. Performance of these controllers are slightly degraded due to inability for precise estimation of the helicopter para- metric uncertainties, the dynamic of actuators, nonlinear fric- tion forces, external disturbances and cross-coupling effects. Because of the computational burden of these algorithms, it is not efficient to implement these algorithms for the more complex realistic systems. Design of disturbance observer (DOB) based controller is one of the most popular methods in the field of high performance positioning systems. DOB based techniques appeared in the late 1980s. In [10] dynamics for each elbow of the robot manipulator is decoupled, and DOB controller is designed for each part independently. The nonlinear distur- bance observer based design, with assumed upper and lower bounds of the disturbance to be known, is discussed in [11]. The DOB based control has been widely used in industry [12]–[15]. This paper focuses on high performance tracking control for electrical driven helicopter body with unmodelled and unknown uncertainties. The cross-coupling effects of the elevation and azimuth dynamics can be treated as external disturbances. The DOB control is used to compensate these effects, thus the interference terms can be decoupled and the desired dynamics performances can be obtained. Hence,