This paper presents a methodology to design the flatness-based controller (FBC) for a nonlinear half-car hydraulic active suspension system. Vehicle dynamics is provided by using handling the trade-off of the ride comfort and safety. For this purpose, considered output variables are vertical acceleration and displacement of the car and vertical displacements of front and rear wheel. The main objective is to design a differential FBC which can isolate car body from the vibrations due to road disturbances acting and provide efficient control under model uncertainty. Extensive simulations are performed for different road profiles and different values of suspension parameters. The obtained results show that the proposed controller performed well in improving the ride comfort and road handling for the half-car model with a hydraulically actuated suspension system.

The purpose of modern wind energy conversion systems is to extract the aerodynamic power from the wind and convert it to electric power. The most popular today is variable-speed horizontal wind turbine (HWT) with three blades. The main objective of this paper is to design the controller that is capable to optimize the power output in the lower wind speed region and to maintain the rotor power limits within the high wind speed region simultaneously. In order to accomplish this task we designed a multiple switching model predictive controller (MPC) with three separate controllers for various wind speed operating regions. In order to design multiple MPC the nonlinear model of wind turbine is represented by a set of local linear models for different regions. The simulation was conducted for different controller settings to validate the effectiveness of the considered MPC in presence of different system constraints on control variables and turbulent winds acting.

In this paper, we present an algorithm for fully autonomous exploration and mapping of an unknown indoor robot environment. This algorithm is based on the active SLAM (simultaneous localization and mapping) approach. The mobile robot equipped with laser sensor builds a map of an environment, while keeping track of its current location. Autonomy is introduced to this system by automatically setting goal points so that either previously unknown space is mapped, or known landmarks are revisited in order to increase map accuracy. Final aim is to maximize both map coverage and accuracy. The proposed procedure is experimentally verified on Pioneer 3-DX mobile robot in real environment, using ROS framework for implementation.

This paper presents approximate methods to estimate iso-values in 3D thermal model reconstruction of indoor environments. The input data is represented in the form of registered point clouds with added temperature information obtained using 3D laser scanner and thermal imaging camera. They are then registered to the common coordinate system using SLAM6D. Marching cubes based algorithm for point cloud polygonalisation using standard Gaussian-like iso-values estimation, as well as proposed approximations, is utilized in order to produce a 3D mesh of an indoor environment with added temperature information. The effectiveness and quality performance of proposed approximative methods are verified through their comparisons with standard Gaussian-like estimation.

The paper proposes a wireless navigation mobile robot system for both path planning and trajectory execution within an indoor maze environment. This system consists of the mobile robot, trajectory planner, motion controller, visual sensor (CCD camera), ZigBee wireless communication device and a maze terrain. The camera is used to capture images of the mobile robot within the maze. Developed image processing and analyzing algorithms determine the robot's position and orientation based on color markers recognition. Markers are mounted on the top of the robot. Based on this data the implemented navigation system calculates a trajectory for the mobile robot from a starting point to a target point. The proposed navigation system is an upgrade to our previously developed system. Maze encryption and motion planning modules have been added to the previous system. Breadth First Search (BFS) and modified Depth First Search (DFS) algorithms were used for the trajectory calculation. A developed control algorithm calculates control signals in real time. These signals are sent to the robot via modules for wireless communication, causing robot motion along the calculated trajectory and eventually, the completion of the trajectory. The whole control system is realized and experimental results have been obtained. The experimental results confirm the robustness and effectiveness of the implemented control system.

This paper describes the model predictive control (MPC) in terms of an explicit controller without on-line optimization problem solving and its implementation in the programmable logic controller (PLC). For this purpose the multi-parametric quadratic programming (MPQP) approach with binary search tree (BST) is used. The optimal values are computed off-line for all possible system states and then stored in the form of appropriate table, while the computation of current system states and the search for proper control output are performed on-line. The task is to control the temperature in an incubator unit with the PLC, which executes an explicit MPC algorithm. Setting the reference value and monitoring of the processes are carried out by the human machine interface (HMI) which communicates with the PLC via Ethernet network. So, the objective is to show that the MPC control can adequately be adapted in PLC, which is commonly used for the control of industrial processes. The effectiveness and robustness of this approach are verified through experiments.

The paper uses Probabilistic Road Map (PRM) based path planning algorithm which composes Halton point sets to improve mobile robot navigation capabilities. The Halton set obtains a good coverage of robot environment with points that is better than using grid based methods. Within PRM learning phase Halton point sets are used to randomly generate the robot configurations which constitute probabilistic road maps. The shortest path between an initial and a final configuration was founded out using A* algorithm adopted for query phase of PRM. The influence of number of Halton points and a distance between adjacent nodes for the current configuration to the path planning are analyzed. For purpose of map building of an unstructured environment a new histogramic based method is applied. In order to implement PRM based planning algorithm the whole navigation system is designed and implemented. The effectiveness of the proposed navigation system was demonstrated in both simulation and experimental modes.

This paper presents a classical approach to model reference adaptive control for trajectory tracking problem. Gradient based or MIT rule adaptation technique was applied to the well known trajectory tracking controller and the mathematical model of that adaptation is presented. The proposed solution is compared to the controller without adaptation and to the feedback linearisation based controller. In both simulation and experimental environments the effectiveness of the proposed adaptive controller is shown and its use justified.

This paper deals with the problems of identification and control of nonlinear helicopter model. The helicopter model is achieved by closed loop identification based on grey-box structure through two steps. The first step considers experiment design to minimize the number of parameters to be estimated. It employs a genetic algorithm for parameter value estimation. In the second step, a criteria function was defined for evaluation of model fitness. After that, linearization of nonlinear model was performed and linear state-space model was obtained for use in controller synthesis. The main contribution of this paper is design of loop shaping controller which is capable to control a nonlinear helicopter model. Advantage of this controller is its simpler linear structure and simpler synthesis. The quality of obtained helicopter model and effectiveness of proposed controller are verified both in simulation and experimental modes.

Determining the position of a mobile robot in every time instant from sensor data is the fundamental problem in mobile robotics. This paper considers a localization of holonomous mobile robot solved in this paper using two different approaches: odometry localization and landmark based localization. In both cases the robot is placed in known environment with landmarks whose coordinates were also known. Detecting the landmarks was done by using the Microsoft Kinect camera. For odometry localization four encoders were used. Data acquired from encoders and camera is fused together employing extended Kalman filter in order to get more accurate estimation of position and orientation. Obtained experimental results prove that using encoders without any additional measurements is not enough for getting reliable estimation of robots position. Odometry localization produced an error that accumulates over time, while in the case of landmark based localization, the error is kept inside acceptable limits.

Recent developments in environment sensing and virtual modelling enabled the construction of virtual models of indoor environments with added thermal information. In this paper we present the methods for heat sources extraction from the 3D thermal model of an indoor environment. They are based on known image segmentation techniques but adjusted to work with 3D models with added thermal information. All data are acquired using an autonomous mobile robot platform equipped with 3D laser scanner and thermal imaging camera.