In this paper a Crazyflie 2.0 nano quadrotor helicopter (quadcopter) as an open source experimental platform for research and education in robotics and control engineering has been presented. This low cost, easily expandable and upgradeable flying robot is here characterized in terms of hardware and software. Three aspects, which demonstrate the potential of broad use of this unmanned aerial vehicle (UAV) by researchers and students, are discussed in the paper. The first one is an acquisition of measurement data from test flights by the proposed, freely available “black-box” software. The second is the use of a new, advanced 4FLY Simulator in order to utilize the MATLAB®/Simulink environment to easily implement a mathematical model of Crazyflie 2.0 dynamics, as well as for a synthesis of various types of controllers with support of OpenGL cross-language in the visualization of simulations results. The 4FLYSimulator allows to test autonomous flights (and landings) with obstacles avoidance and to conduct learning and teaching the basics of Crazyflie 2.0 piloting. In the third aspect the authors outlined promising, preliminary results obtained in control of flying robot by pointing device (positioner) and with the support of a vision system, which basis only on a single Kinect sensor.

This paper presents a novel tuning method for the extended state observer (ESO), which is applied in the active disturbance rejection control (ADRC) algorithm operating in a stochastic environment. Instead of the traditional pole placement (PP) method, the selection of ESO gains based on the noise variances of the Kalman filter (KF) is proposed. Also, a simple parametrization of ESO gains for the particular control process based on the observer bandwidth is introduced. A root locus and frequency analysis is conducted for the KF-based observer and presented with regard to the proposed tuning method. The presented results come from experiments carried out on the ball balancing table (BBT) real plant for various measurement noise levels. The possibilities of rejecting measurement noise by the estimation algorithm were investigated to ensure effective control and minimize the control signal energy. Based on the conducted experiments, one can conclude that the presented tuning method provides better results than the traditional PP algorithm in the stochastic environment in terms of control quality and reduction in measurement noise.

This paper presents a new adaptive modification of active disturbance rejection control (ADRC) with parameter estimation based on a recursive least-squares (RLS) method. The common ADRC used in many applications relies on the simple approach, which assumes the simplification of the object into an integral chain form. However, this model-free ADRC does not guarantee the stability of a closed-loop system in the presence of noticeable modeling uncertainties, so it is compared in this paper to another approach, in which the linear part of the system is included in the ADRC framework (generalized ADRC). This incorporation of the model is examined in the paper for a wide range of model and controller parameters, considering also the presence of external disturbances as well as parameter uncertainties, pointing out the limitations of fixed-gain algorithms. Then, the adaptive modification of the model-based ADRC is proposed, which is equipped with a real-time estimation of model parameters by means of the RLS method in continuous time. The stability conditions of the proposed modification of the algorithm in the closed control loop are also analyzed. It can be concluded that, under appropriate conditions, the inclusion of information about known plant parameters into the ADRC can noticeably improve the conditions of the control system. The proposed adaptive model-based approach enables quality improvement during the control process even with initially unknown parameters, for time-varying parameters, and in the presence of parametric uncertainties and external disturbances. The tests were performed on a real plant—the task of controlling the angular velocity of the direct current (DC) motor was considered.

This paper presents a new particle filter algorithm (MultiPDF) for state estimation of nonlinear systems. The proposed method is a modification of the standard particle filter approach. Due to the strong need for the acceleration of calculations and an improvement in the estimation quality of state estimation, the authors propose a method which enables one to divide the main particle filter into smaller sub-filters with an accordingly smaller number of particles for each one of them. The algorithm has been implemented for various numbers of particles and subordinate parallel filters. Estimation quality has been checked for nine nonlinear objects (both one- and multidimensional) and evaluated through the quality index, average root-mean-squared error. The computation time of the particle filter algorithm for several hardware configurations has been compared. Based on the obtained results, it can be concluded that, besides the computation acceleration, the parallelization of the particle filter’s operation also improves the estimation quality.

In this paper, we present a new approach to mathematical modeling of the bicycle. It is based on the detailed nonlinear Whipple scientific description. We are focused on the state space representation which we use to solve the control law and we test the optimal linear quadratic control which finally gives satisfactory results. The article includes a several computer simulations of the single-track vehicle motion. We indicate when the bicycle is self-stable. The results are useful to make further research on the bicycle modelling.