Luka Teslić

EKF-Based Localization of a Wheeled Mobile Robot in Structured Environments

This paper deals with the problem of mobile-robot localization in structured environments. The extended Kalman filter (EKF) is used to localize the four-wheeled mobile robot equipped with encoders for the wheels and a laser-range-finder (LRF) sensor. The LRF is used to scan the environment, which is described with line segments. A prediction step is performed by simulating the kinematic model of the robot. In the input noise covariance matrix of the EKF the standard deviation of each robot-wheel’s angular speed is estimated as being proportional to the wheel’s angular speed. A correction step is performed by minimizing the difference between the matched line segments from the local and global maps. If the overlapping rate between the most similar local and global line segments is below the threshold, the line segments are paired. The line parameters’ covariances, which arise from the LRF’s distance-measurement error, comprise the output noise covariance matrix of the EKF. The covariances are estimated with the method of classic least squares (LSQ). The performance of this method is tested within the localization experiment in an indoor structured environment. The good localization results prove the applicability of the method resulting from the classic LSQ for the purpose of an EKF-based localization of a mobile robot.

Supervised Hierarchical Clustering in Fuzzy Model Identification

This paper presents a new, supervised, hierarchical clustering algorithm (SUHICLUST) for fuzzy model identification. The presented algorithm solves the problem of global model accuracy, together with the interpretability of local models as valid linearizations of the modeled nonlinear system. The algorithm combines the advantages of supervised, hierarchical algorithms, which are based on heuristic tree-construction algorithms, together with the advantages of fuzzy product space clustering. The high flexibility of the validity functions that is obtained by fuzzy clustering combined with supervised learning results in an efficient partitioning algorithm, which is independent of initialization and results in a parsimonious fuzzy model. Furthermore, the usability of SUHICLUST is very undemanding, because it delivers, in contrast with many other methods, reproducible results. In order to get reasonable results, the user only has to set either a threshold for the maximum number of local models or a value for the maximum allowed global model error as a termination criterion. For fine-tuning, the interpolation smoothness controls the degree of regularization. The performance is illustrated on both analytical examples and benchmark problems from the literature.

Using a LRF sensor in the Kalman-filtering-based localization of a mobile robot.

This paper deals with the problem of estimating the output-noise covariance matrix that is involved in the localization of a mobile robot. The extended Kalman filter (EKF) is used to localize the mobile robot with a laser range finder (LRF) sensor in an environment described with line segments. The covariances of the observed environment lines, which compose the output-noise covariance matrix in the correction step of the EKF, are the result of the noise arising from a range-sensors (e.g., a LRF) distance and angle measurements. A method for estimating the covariances of the line parameters based on classic least squares (LSQ) is proposed. This method is compared with the method resulting from the orthogonal LSQ in terms of computational complexity. The results of a comparison show that the use of classic LSQ instead of orthogonal LSQ reduce the number of computations in a localization algorithm which is a part of a SLAM (simultaneous localization and mapping) algorithm. Statistical accuracy of both methods is also compared by simulating the LRFs measurements and the comparison proves the efficiency of the proposed approach.

Nonlinear System Identification by Gustafson–Kessel Fuzzy Clustering and Supervised Local Model Network Learning for the Drug Absorption Spectra Process

This paper deals with the problem of fuzzy nonlinear model identification in the framework of a local model network (LMN). A new iterative identification approach is proposed, where supervised and unsupervised learning are combined to optimize the structure of the LMN. For the purpose of fitting the cluster-centers to the process nonlinearity, the Gustafsson-Kessel (GK) fuzzy clustering, i.e., unsupervised learning, is applied. In combination with the LMN learning procedure, a new incremental method to define the number and the initial locations of the cluster centers for the GK clustering algorithm is proposed. Each data cluster corresponds to a local region of the process and is modeled with a local linear model. Since the validity functions are calculated from the fuzzy covariance matrices of the clusters, they are highly adaptable and thus the process can be described with a very sparse amount of local models, i.e., with a parsimonious LMN model. The proposed method for constructing the LMN is finally tested on a drug absorption spectral process and compared to two other methods, namely, Lolimot and Hilomot. The comparison between the experimental results when using each method shows the usefulness of the proposed identification algorithm.

Kalman-filtering-based localization of a mobile robot with a LRF in a simulated 2D environment

This article deals with the modelling, simulation and localization of a mobile robot using a laser range finder (LRF) in a 2D environment. The presented simulator is applied for a localization algorithm based on the extended Kalman filter (EKF) design. The prediction of the robotpsilas pose is performed by simulating the kinematic model of the robot. The input-noise covariance matrix is derived from the noise variances of the angular velocity measurements of both robot wheels. Correction of the robot pose is performed by matching the line parameters of the local environment map to the line parameters of a global map transformed to the robotpsilas coordinates.

Mobile-robot pose estimation and environment mapping using an extended Kalman filter

In this paper an extended Kalman filter (EKF) is used in the simultaneous localisation and mapping (SLAM) of a four-wheeled mobile robot in an indoor environment. The robot’s pose and environment map are estimated from incremental encoders and from laser-range-finder (LRF) sensor readings. The map of the environment consists of line segments, which are estimated from the LRF’s scans. A good state convergence of the EKF is obtained using the proposed methods for the input- and output-noise covariance matrices’ estimation. The output-noise covariance matrix, consisting of the observed-line-features’ covariances, is estimated from the LRF’s measurements using the least-squares method. The experimental results from the localisation and SLAM experiments in the indoor environment show the applicability of the proposed approach. The main paper contribution is the improvement of the SLAM algorithm convergence due to the noise covariance matrices’ estimation.

Simulation of a Mobile Robot with an LRF in a 2D Environment and Map Building

This article deals with the modelling and simulation of a mobile robot with a laser range finder in a 2D environment and map building. The simulator is built in the Matlab Simulink environment, thereby taking advantage of the powerful Matlab toolboxes for developing mapping, localization, SLAM, and navigation algorithms. A map-building algorithm is developed and tested in a simulation. The line segments, extracted from the LRF’s output in each scan, are made up of polylines, which are merged with the existing global map to form a new global map. The global map of the environment is represented by unions of line segments, where each union represents an object in the environment. Map building, localization and navigation are important issues in mobile robotics. To develop and test algorithms for executing tasks of this kind, it is useful to have a simulator of a mobile robot equipped with sensors in a static environment. Since a Laser Range Finder (LRF) is often used as the basic interaction between the robot and the environment, the represented mobile robot model also includes a model of the LRF. The problem of robotic mapping and localization has been widely studied. A robot has to know its own pose (localization problem) in order to build a map, and it also needs to know the environment map (mapping problem) to localize itself to its current pose. The problems of mapping and localization can be handled separately if the robot’s pose is given to the robot by a human or from using GPS and INU sensors (outdoor environments) when map building. The map of the environment can then be used to solve the localization problem.

Monitoring of waste-water treatment plant using Takagi-Sugeno fuzzy model

In this paper detection of sensor faults in waste-water treatment plant by Gustafson-Kessel fuzzy clustering algorithm is discussed and presented. The main idea in the case of process monitoring by the use of fuzzy clustering algorithms is comparison between the fuzzy clusters for a normal operation regime and the current behavior. The detection of sensor faults was applied to the simulation model of the waste-water treatment plant, where the following measurements were obtained: influent ammonia concentration, dissolved oxygen concentration in the first aerobic reactor tank, temperature, dissolved oxygen concentration and the ammonia concentration in the second aerobic reactor. The results of fault detection based on fuzzy model are shown and discussed.