Search

Ренер-Телеров ефекат представља вибрационо-електронску (вибронску) спрегу и последица jе нарушавања Борн-Опенхаjмерове апроксимациjе приликом савиjања линеарних молекула са просторно дегенерисаним електронским стањем. Услед савиjаjућих вибрациjа молекула, електронско стање коjе jе просторно дегенерисано при линеарноj геометриjи се цепа на два блиска електронска стања коjа међусобно интерагуjу. Вибронска спрега тада одређуjе ровибрациону структуру спектра молекула, као и могуће понашање молекула у хемиjским реакциjама. У овоj докторскоj дисертациjи проучаван jе Ренер-Телеров ефекат код вишеатомских молекула са произвољним броjем jезгара (више од три jезгра) у Π електронским стањима и линеарном равнотежном геометриjом применом вариjационе методе. За разлику од троатомских молекула где се поjављуjе класични Ренер-Телеров ефекат коjи jе описан у претходном пасусу, код четвороатомских молекула и молекула са већим броjем jезгара уочава се комбинациjа Ренер-Телеровог ефекта и поjаве коничних пресека потенциjалних површи, односно Jан-Телеровог ефекта. Детаљно jе представљен модел за третирање Ренер-Телеровог ефекта и његова поузданост jе тестирана на примеру молекула C2H+ 2 у X 2Π u електронском стању ab initio рачунањем електронских потенциjалних површи и неадиjабатских матричних елемената. Проучена jе топологиjа (избегнутих) пресецања потенциjалних површи и неадиjабатских матричних елемената. У програмском jезику Python написан jе програм за вариjационо рачунање ниско-енергетских вибронских нивоа вишеатомских молекула са произвољним броjем jезгара, линеарном равнотежном геометриjом, Π електронским стањима и произвољне спинске мултиплетности. Конкретно, проучаван jе четвороатомски молекул C2S− 2 у X 2Πu стању. Вариjационо су израчунати вибронски нивои овог молекула и упоређени са доступним експерименталним подацима, при чему jе предложена асигнациjа трака коjе нису биле експериментално асигниране. Даље, топологиjа (избегнутих) пресецања потенциjалних површи и неадиjабатски, The Renner-Teller effect represents а vibrational-electronic (vibronic) coupling and arises as a consequence of violation of the Born-Oppenheimer approximation in linear molecules with spatial degenerate electronic states upon bending vibrations. During bending vibrations of the molecule, a spatially degenerate electronic state at linear geometry splits into two energetically close electronic states that mutually interact. Vibronic coupling could lead to complicated rovibrational spectra of molecules, and also determine their reaction pathways. In this doctoral thesis, we have studied the Renner-Teller effect in linear polyatomic molecules with arbitrary number of nuclei (more than three) in Π electronic states and linear equilibrium geometry by using the variational method. In comparison to the classical Renner-Teller effect in triatomic molecules that was defined in the previous paragraph, in tetra- and any-atomic molecules a combination of classical Renner-Teller effect and appearance of conical intersections of potential energy surfaces occur, i.e. the Jahn-Teller effect. We describe in details a model for treatment of the Renner-Teller effect and test its plausibility on example of C2H+ 2 molecule in the X 2Πu electronic state by ab initio calculations of its electronic potential surface and nonadiabatic matrix elements. Topology structure of (avoided) crossings of potential surfaces and nonadiabatic matrix elements was studied. The program for variational calculation of low-lying vibrionic levels of polyatomic molecules with arbitrary number of nuclei and linear equilibrium geometry in Π electronic states and arbitrary spin multiplicity has been written in Python programming language. We have studied the tetra-atomic molecule C2S− 2 in its X 2Π u electronic state. Vibronic levels of this molecule have been variationally calculated and the obtained results were compared with available experimental data. The assignation of several previously unassigned spectral

In the present paper, the results of combined theoretical and experimental investigation of small non-stoichiometric bromine-doped potassium clusters are reported. Potassium-bromide clusters were obtained by the Knudsen cell combined with surface ionization in the temperature range of 1000–1600 K, and selected by a magnetic sector mass spectrometer. Positive ions of KnBr (n = 3–6) and KnBrn-1 (n = 3–5) clusters were detected for the first time in one set measurement. In order to reveal the geometrical structure of each type of detected cluster, the randomized search algorithm was employed to survey the (Born-Oppenheimer) potential energy surface of both the neutral and cationic KnBr(0,1+) (n = 1–6) and KnBrn-1(0,1+) (n=3–5) clusters, followed by Density functional theory geometry optimizations, and many lowest-energy conformational isomers are presented. From the total electronic energies of clusters computed by the ab initio RCCSD(T)/ECP10MDF(K),cc-pVTZ-PP(Br) method at obtained equilibrium nuclear geometries, the following stability parameters of clusters were computed: their relative energies, the adiabatic and vertical ionization energies, binding energies per atom, and dissociation energies. Both experimental and theoretical results have shown that the title clusters belong to the group of “superalkali” clusters.

Process planning and scheduling are two of the most important manufacturing functions which are usually performed sequentially in traditional approaches. Considering the fact that these functions are usually complementary, it is necessary to integrate them so as to improve performance of a manufacturing system. This paper presents implementation of novel nature-inspired Ant Lion Optimization (ALO) algorithm for solving this combinatorial optimization problem effectively. As the ALO algorithm mimics the intelligent behavior of antlions in hunting ants, the main steps of hunting prey, its mathematical modeling, and optimization procedure for integration of process planning and scheduling is proposed. The algorithm is implemented in Matlab environment and run on the 3.10 GHz processor with 2 GBs of RAM memory. Experimental results show applicability of the proposed approach in solving integrated process planning and scheduling problem.

Market growth and mass customization cause a need for a change in traditional manufacturing. Decentralized decision making and integration of process planning is necessary in order to become concurrent in the market. The paper presents decentralized decision making methodology using multi-agent systems. The model is used for integrated process planning and scheduling based on the minimum processing time under dynamic change of the environment. Two types of disturbance are used to represent the change: part arrival and machine breakdown. The proposed model comprises part agent, job agent, machine agent and optimization agent. Comparative analysis is conducted using simulation in AnyLogic software in order to verify the proposed approach.

The extended Kalman filter (EKF) has become a popular solution for the simultaneous localization and mapping (SLAM). This paper presents the implementation of the EKF coupled with a feedforward neural network for the monocular SLAM. The neural extended Kalman filter (NEKF) is applied online to approximate an error between the motion model of the mobile robot and the real system performance. Inadequate modeling of the robot motion can jeopardize the quality of estimation. The paper shows integration of EKF with feedforward neural network and simulation analysis of its consistency and implementation of the NEKF with a mobile robot, laboratory experimental environment, and a simple USB camera. The simulation and experimental results show that integration of neural network into EKF prediction-correction cycle results in improved consistency and accuracy.

Nowadays, one of the most desirable features of every robotic system is the ability to adapt to the real world changing conditions. Similarly, failure prediction is equally important in different manufacturing environments in which repairs are often infeasible and failures can have disastrous consequences. In industrial robotics, failure prediction is helpful in reduction of a system down-time by identifying and repairing faulty components. Also, the reliability of a product manufacturing and increased human safety is ensured by implementing fault tolerance and failure prediction unit in the robotic system. It is known that the supervision and learning of robotic executions is not a trivial problem. In the 21st century, robots must be able to tolerate and predict internal failures in order to successfully continue performing their tasks. This doctoral dissertation presents a novel approach for prediction of robot execution failures based on machine learning technique - neural networks (NNs). Real data consisting of robot forces and torques recorded immediately after the system failure are used for the NN training. Two types of neural networks are used: feedforward and recurrent (Elman) NNs. In total, 7 different learning algorithms and 24 NN architectures are implemented in order to find optimal solution for the problem of robot execution failures prediction. Each multilayer feedforward NN with different learning algorithm and architecture that consists of 1, 2, 3, or 4 hidden layers is evaluated several times, and the same NN architectures are trained using Elman recurrent NN. Experimental results indicate that Bayesian Regularization algorithm is the best choice for the prediction problem with prediction rate of 95.4545 percent, despite having the erroneous or otherwise incomplete sensor measurements invoked in the dataset. The experimental results show that the NN outperforms state-of-the-art algorithms, such as the Naïve Bayes, Decision Trees and Support Vector

In this paper we test Extended Information Filter (EIF) for sequential training of Hyper Basis Function Neural Networks with growing and pruning ability (HBF-GP). The HBF neuron allows different scaling of input dimensions to provide better generalization property when dealing with complex nonlinear problems in engineering practice. The main intuition behind HBF is in generalization of Gaussian type of neuron that applies Mahalanobis-like distance as a distance metrics between input training sample and prototype vector. We exploit concept of neuron’s significance and allow growing and pruning of HBF neurons during sequential learning process. From engineer’s perspective, EIF is attractive for training of neural networks because it allows a designer to have scarce initial knowledge of the system/problem. Extensive experimental study shows that HBF neural network trained with EIF achieves same prediction error and compactness of network topology when compared to EKF, but without the need to know initial state uncertainty, which is its main advantage over EKF.

Obtaining an optimal process plan according to all alternative manufacturing resources has become very important task in flexible process planning problem research. In this paper, we use a novel nature-inspired algorithm called Ant Lion Optimizer (ALO) to solve this NP-hard combinatorial optimization problem. The network representation is adopted to describe flexibilities in process planning and mathematical model for the minimization of the total production time and cost is presented. The algorithm is implemented in Matlab environment and run on the 3.10 GHz processor with 2 GBs of RAM memory. The presented experimental results show that the proposed algorithm performs better in comparison with other bio-inspired optimization algorithms.

This paper presents concise comparison study of newly developed bioinspired algorithms for global optimization problems. Three different metaheuristic techniques, namely Accelerated Particle Swarm Optimization (APSO), Firefly Algorithm (FA), and Grey Wolf Optimizer (GWO) are investigated and implemented in Matlab environment. These methods are compared on four unimodal and multimodal nonlinear functions in order to find global optimum values. Computational results indicate that GWO outperforms other intelligent techniques, and that all aforementioned algorithms can be successfully used for optimization of continuous functions.

Multi-agent systems have been used for modelling various problems in the social, biological and technical domain. When comes to technical systems, especially manufacturing systems, agents are most often applied in optimization and scheduling problems. Traditionally, scheduling is done after creation of process plans. In this paper, agent methodology is used for integration of these two functions. The proposed multi-agent architecture provides simultaneous performance of process planning and scheduling and it consists of four intelligent agents: part and job agents, machine agent, and optimization agent. Verification and feasibility of a proposed approach is conducted using agent based simulation in AnyLogic software.

The paper presents an approach based on the application of the Particle Swarm Intelligence algorithm for solving the combinatorial optimization problem of determining the order of execution of operations when processing parts on machines. The proposed approach considers the following types of flexibility: machine flexibility, tool flexibility, process flexibility, and operation sequence flexibility. To represent the flexibility of the machining process of part processing, the method of representing the manufacturing process through networks was chosen, while for the described mathematical model, the criteria for optimization are minimum production time and minimum costs. Experimental results show that the presented algorithm is more efficient, i.e. to give optimal orders of operations in less time and fewer iterations compared to single GA, SA and hybrid GA-SA algorithm.

U radu je prikazan i izveden novi sekvencijalni algoritam za obučavanje višeslojnog perceptrona u prisustvu autlajera. Autlajeri predstavljaju značajan problem, posebno ukoliko sprovodimo sekvencijalno obučavanje ili obučavanje u realnom vremenu. Linearizovani Kalmanov filtar robustan na autlajere (LKF-RA), je statistički generativni model u kome je matrica kovarijansi šuma merenja modelovana kao stohastički proces, a apriorna informacija usvojena kao inverzna Višartova raspodela. Izvođenje svih jednakosti je bazirano na prvim principima Bajesovske metodologije. Da bi se rešio korak modifikacije primenjen je varijacioni metod, u kome rešenje problema tražimo u familiji raspodela odgovarajuće funkcionalne forme. Eksperimentalni rezultati primene LKF-RA, dobijeni korišćenjem stvarnih vremenskih serija, pokazuju da je LKF-RA bolji od konvencionalnog linearizovanog Kalmanovog filtra u smislu generisanja niže greške na test skupu podataka. Prosečna vrednost poboljšanja određena u eksperimentalnom procesu je 7%., We derive a new sequential learning algorithm for Multilayered Perceptron (MLP) neural network robust to outliers. Presence of outliers in data results in failure of the model especially if data processing is performed on-line or in real time. Extended Kalman filter robust to outliers (EKF-OR) is probabilistic generative model in which measurement noise covariance is modeled as stochastic process over the set of symmetric positive-definite matrices in which prior is given as inverse Wishart distribution. Derivation of expressions comes straight form first principles, within Bayesian framework. Analytical intractability of Bayes' update step is solved using Variational Inference (VI). Experimental results obtained using real world stochastic data show that MLP network trained with proposed algorithm achieves low error and average improvement rate of 7% when compared directly to conventional EKF learning algorithm.

In this paper, we present new Learning from Demonstration-based algorithm that generalizes and extracts relevant features of desired motion trajectories for differential drive mobile robots. The algorithm is tested through series of simulations and real world experiments in which desired task is demonstrated by the human teacher while teleoperating the mobile robot in the working environment. In the first step of the developed method, Gaussian Mixture Model (GMM) is built for incremental motions of the mobile robot between two consecutive poses. After this, the hidden Markov model is used to capture transitions between states (temporal variations of the data between clusters) which are missing from static GMM representation. Generalization of the motion is achieved by using the concept of keyframes, defined as points in which significant changes between GMM/HMM states occur. In the second step, the resulting GMM/HMM representation is used to generate optimal state sequences for each demonstration and to temporally align them, using 1D dynamic time warping, with respect to the one most consistent with the GMM/HMM model. This phase implies extraction of keyframes along all state sequences and projecting them into control space, in which controls are aligned in time as well. Finally, the generalized controls are obtained by averaging over all controls at the keyframes; simple piecewise cubic spline method is used for interpolation between generated control values. The main advantage of the developed algorithm is its ability to learn and generalize from all demonstrated examples which results in high quality reproductions of the motion. The proposed approach is verified both in simulated environment and using real mobile robot.

In this paper, we present new Learning from Demonstration ((LfD) - based algorithm that generalizes and extracts relevant features of desired motion trajectories for differential drive mobile robots. The algorithm is tested through series of simulations and real world experiments in which desired task is demonstrated by the human teacher while teleoperating the mobile robot in the working environment. The human teacher shows the specifics of the new skill or motion behavior that robot needs to learn; the robot extracts information from demonstrations and builds the internal model based on the learning algorithm and in this paper we will refer to it as the Learning from Demonstration.

The hybrid mobile robot control algorithm consists of two independent control loops. By developing two control phases, the transportation task is separated into two parts: movement from the initial position to a position at a great distance from the machine tool (global control) and movement from this position to the machine tool/intermediate point (local control). The original control system based on epipolar geometry as well as on learning of motion trajectories and visual commands was implemented on the Khepera II nonholonomic mobile robot (with additional equipment: KheGrip gripper and Prestigio PWC 2 WEB camera) by using metaheuristic technique in a laboratory model of the technological environment.

It is known that the supervision and learning of robotic executions is not a trivial problem. Nowadays, robots must be able to tolerate and predict internal failures in order to successfully continue performing their tasks. This study presents a novel approach for prediction of robot execution failures based on neural networks. Real data consisting of robot forces and torques recorded immediately after the system failure are used for the neural network training. The multilayer feedforward neural networks are employed in order to find optimal solution for the failure prediction problem. In total, 7 learning algorithms and 24 neural architectures are implemented in two environments - Matlab and specially designed software titled BPnet. The results show that the neural networks can successfully be applied for the problem in hand with prediction rate of 95.4545%, despite having the erroneous or otherwise incomplete sensor measurements invoked in the dataset. Additionally, the real-world experiments are conducted on a mobile robot for obstacle detection and trajectory tracking problems in order to prove the robustness of the proposed prediction approach. In over 96% for the detection problem and 99% for the tracking experiments, neural network successfully predicted the failed information, which evidences the usefulness and the applicability of the developed intelligent method.