The extraordinary capability of swarming ant species in route finding and foraging efficiency through trail trail development has been studied for many years. Scientists have been able to capture the behavior of individual ants in control algorithms and used the resulting artificial swarm of ants to solve difficult problems. These problems include function optimization, sorting, clustering and route finding. So far, the route finding capabilities, where agents solely rely on local available information, have only been demonstrated in virtual discrete domains. The aim of this thesis is the development of an agent based control method that allows a swarm of robots to establish a trail between a source and target, placed in an unknown domain. The method solely relies on information locally available to the agent. No direct communication among the agents is necessary as agents mark the environment to relay information. Simulations are performed in a Python and C-based computer program, developed in tandem with this study. A versatile controller is presented that, in conjunction with suggestions for the swarm size and pheromone characteristics, can be implemented on real robots to explore unknown environments in a multitude of scenarios.

Blind Source Separation (BSS), the separation of latent source components from observed mixtures, is relevant to many fields of expertise such as neuro-imaging, economics and machine learning. Reliable estimates of the sources can be obtained through diagonalization of the cumulant tensor, i.e., a fourth-order symmetric multi-linear array containing the cross-kurtosis values of observed mixtures. The downside of such diagonalization methods is that they scale quartically with the increase of the amount of source components to estimate due to the tensor’s quartic size. Tensor decomposition can simultaneously diagonalize the cumulant tensor and address its size. However, it does not resolve the scalability issue due to the restriction of having to first explicitly compute the tensor. It is studied how decomposing the cumulant tensor in implicit fashion can be used to solve the BSS problem while simultaneously addressing its scalablity issue. A class of implicit cumulant tensor decomposition algorithms is derived which scale more favorably than their explicit counterparts in terms of either computational cost, storage cost or both. Firstly, a novel QR-Tensor algorithm (QRT) is introduced which allows for the simultaneous diagonalization of a tensor’s outer-slices. It is theoretically shown how an implicit version of the QRT algorithm can be used to solve the BSS problem at a linearly scaling computational cost. Secondly, a fixed-point Canonical Polyadic Decomposition (CPD) iteration method is presented. It reduces the computational complexity from a quartic dependence to a linear dependence on the amount of signals to estimate. The source estimation performance of the devised implicit decomposition methods is compared to that of the state-of-the-art FastICA for an artificial linear BSS problem. Results show that both fixed-point CPD and QRT are superior to FastICA when it comes to the computation time needed to reach convergence, while producing estimated sources of similar quality. It is shown that when the amount of sources to estimate is increased both QRT and FastICA struggle to converge. In contrast, the fixed-point CPD method converges within a consistent amount of iterations, suggesting a method more suitable for the estimation of a large amount of sources.

One of the most promising ideas in autonomous vehicle control systems is letting the vehicle drive autonomously outside the normal, linear, operating region and letting it "drift". By doing so, the maneuverability of the vehicle could be enhanced. To enable systems that can control this behaviour, estimation of certain vehicle states is needed with high accuracy and high frequency.In this project, a new solution to this problem is proposed by combining a mixed dynamic-kinematic observer with a single camera that produces velocity measurements based on tracking the ground plane. To improve filtering of the camera velocity measurements, the measurement error covariance matrix is updated online based on a model of the camera measurement error. Evaluation of the new methodology was done on data recorded from a 1:10 scale test vehicle and performance was assessed based on ground truth data obtained using a Motion Capture System.In normal driving conditions with correctly identified vehicle parameters, an observer without camera performs better by 25% in terms of RMSE on lateral velocity and sideslip angle estimation. However, the online adaptation of the covariance matrix results in an estimate that is at least 45% more accurate in terms of RMSE than the same observer without online covariance adaptation. Next to that, experiments show that the proposed observer with camera has better robustness to uncertainty in model parameters by almost a factor five in terms of RMSE than the observer without camera.When the grip of the tires is physically lowered and the vehicle is drifting, the proposed observer can track large sideslip angles (>30°), where the state-of-the-art observer without camera is not able. The state-of-the-art observer has an increase in RMSE of 75% on all estimated quantities in comparison to the proposed methodology. These results show that adding a camera to an existing sideslip angle observer greatly enhances robustness of the observer to uncertainty in model parameters and violation of model assumptions. This comes dat the cost of losing some accuracy in normal driving conditions.

This thesis studies the Canonical Polyadic Decomposition (CPD) constrained kernel machine for large scale learning, i.e. learning with a large number of samples. The kernel machine optimization problem is solved in the primal space, such that the complexity of the problem scales linearly in the number of samples as opposed to scaling cubically in the dual space. Product feature maps are applied to transform the input data. The weights are constrained to be a CPD, so the number of weights scales linearly in the number of features. The CPD introduces a nonlinearity, so nonlinear optimization must be applied. It is studied in which situation it is more advantageous to apply iterative all-at-once opti- mization compared to Alternating Least Squares (ALS) to solve the CPD constrained kernel machine problem. Specifically, all-at-once gradient descent methods are studied. An efficient analytical algorithm for the all-at-once gradient is derived. Furthermore, it is shown that automatic differentiation (AD) can also be applied, but it is found to be slower than the analytical method. The selection of a step size is found to be challenging. It is shown that the magnitude of the gradient of the mean squared error (MSE) term decreases for an increasing number of features. As a result, selecting the step size becomes more difficult for more features. To overcome this, the Line search and the Adam method are studied. A general expression for the exact line search solution is derived. It can be applied to compute the optimal step size for any step direction and any number of features. However, the Adam method performs better in terms of loss after training, convergence and the training run time. The mini-batch Adam method is used to evaluate the performance of all-at-once optimization for large scale learning. It is found that the Adam method no longer performs well for data sets with around 16 features or more, likely due to the decrease in the magnitude of the gradient of the MSE term. On large scale data sets with fewer features, the Adam method outperforms ALS in terms of run time until convergence while achieving similar training and validation losses. The Adam method reached convergence on a data set with 11 million samples within ten minutes. Furthermore, it is shown that the scaling of the run time of the Adam method in terms of the feature map order and the CP-rank is more than an order lower than the scaling of ALS when the methods are run on a GPU. This makes to Adam method more suitable for more complex models.

Due to new legislation, a lot of research is done in the development of a Collision Avoidance Systems (CAS) for surface mines. Almost all CASs studied in literature are decentralized and developed for passenger vehicles or aviation applications. That makes them unsuitable for the mining environment. A centralized system can ensure optimal control, which means scenarios like intersections can be managed much better. This results in a decreased impact on production. It also increases acceptance, which means operators will be less inclined to turn off the CAS. For this reason, a centralized CAS for mining vehicles was developed. Literature shows that c{MILP} can create relatively accurate models that can be solved very fast. Since only linear constraints can be used, a new geometric model was formulated using four squares that are equally spaced on the longitudinal axis of the vehicle. The vehicle dynamics are described using a linear model. When the optimization problem becomes infeasible, no control is issued to the vehicle. The construction of three different operation modes for the vehicle allow the algorithm to deal with the non linear event of a collision which would otherwise result in an infeasible problem. Weight can be assigned to individual vehicles to determine which vehicle should be prioritized or to model traffic situations. Simulations of six scenarios commonly encountered in mines, involving two vehicles, show that the system is able to avoid all collisions. When a third vehicle is added, collisions are still avoided but solution times increase exponentially. As simulation results were promising, the system was implemented on two test vehicles. Quantitative tests show that the optimization runs as fast using noisy real world data as it does using clean, artificially created data.

For large-scale system with tens of thousands of states and outputs the computation in the conventional Kalman filter becomes time-consuming such that Kalman filtering in large-scale real-time application is practically infeasible. A possible mathematical framework to lift the curse of dimensionality is to lift the problem in higher dimensions with the use of tensors and then decompose it. The tensor-train decomposition is chosen due to its computational advantages for systems with low tensor-train rank. Within this thesis two main limitations of the existing tensor Kalman filter are solved. First, a method is developed based on tensor-train rank truncation of the covariances to increase the computational speed for more general systems. Second, a MIMO tensor Kalman filter is developed for a specific class of systems. The power of the developed methods is shown on the example of adaptive optics which fits into the framework. A comparison with state-of-the-art large-scale estimation algorithms shows the computational advantage of the tensor Kalman filter at the cost of approximation errors.

The world population is growing rapidly and the demand for healthy food grows with it. Greenhouse cultivation provides an efficient way to grow crops in a protected and controlled environment. In the past, many greenhouse control algorithms have been developed. How- ever, the majority of these algorithms rely on an explicit parametric model description of the greenhouse. These models are often based on physical laws such as conservation of mass and energy and contain many parameters which should be identified. Due to the complex and highly non-linear dynamics of greenhouses, these models might not be applicable to control greenhouses other than the one for which the model has been designed and identified. Hence, in current horticultural practice these control algorithms are scarcely used. Therefore, the need rises for a control algorithm which does not rely on a parametric system representa- tion but rather on input/output data of the greenhouse system, hereby establishing a way to control the system with unknown or unmodeled dynamics. A recently proposed algo- rithm, Data-Enabled Predictive Control (DeePC), is able to replace system identification, state estimation and future trajectory prediction by one single optimization framework. The algorithm exploits a non-parametric model constructed solely from input/output data of the system. This algorithm is employed in order to control the greenhouse climate. It is shown that in numerical simulation the DeePC algorithm is able to control the greenhouse climate. A comparison is made with the conventional Nonlinear Model Predictive Control (NMPC) algorithm in order to show the differences between a predictive control algorithm that has direct access to the non-linear greenhouse simulation model and a purely data-driven predic- tive control algorithm. Both algorithms are compared based on reference tracking accuracy and computational time. Furthermore, it is shown in numerical simulation that the DeePC algorithm is able to cope with changing dynamics within the greenhouse system throughout the crop cycle.

This research aims at quantifying the uncertainty in the predictions of tensor network constrained kernel machines, focusing on the Canonical Polyadic Decomposition (CPD) constrained kernel machine. Constraining the parameters in the kernel machine optimization problem to be a CPD results in a linear computational complexity in the number of features, whereas the original problem suffers heavily from the curse of dimensionality as the number of parameters scale exponentially. By employing a product feature map with polynomial features, the original data input is transformed to a higher-dimensional space. Three different methods are investigated for quantifying the uncertainty of the predictions of the CPD constrained kernel machine. Firstly, the delta method is proposed which is a frequentist approach that linearizes a nonlinear parametric model around the estimated model. By estimating the covariance of the model parameters, the delta method can estimate the uncertainty in the model predictions based on the estimated parameter uncertainties. The delta method is compared to two other methods that are able to reflect the prediction uncertainty: the Bayesian method and Single Bayesian Core (SBC) method. The Bayesian method treats the parameters in the factor matrices of the CPD as probability distributions rather than single values and the SBC method incorporates both frequentist and Bayesian aspects. A comparison between the three different methods is performed based on an assessment on the correctness and informativeness of the uncertainty measures of prediction intervals and confidence Intervals. It was found by regression and classification experiments that all three methods can provide valuable uncertainty quantification measures in terms of correctness and informativeness for the CPD constrained kernel machine. However, the Bayesian method provides in general more conservative uncertainty intervals compared to the delta and SBC method. A major drawback of the Bayesian method is its lack of scalability as the size of the mean and covariance, constructed by the unscented transform in the Bayesian method, scale exponentially. Furthermore, the delta and SBC method produce high quality uncertainty intervals and the methods provide remarkably similar uncertainty quantification on the prediction error variance.

Compressed sensing is a framework in signal processing that enables the efficient acquisition and reconstruction of sparse signals. A widely-used class of algorithms that are used for this reconstruction, called greedy-algorithms, depend on non-convex optimization. With increasing signal size, these problems become computationally very hard. Block compressed sensing is a framework in compressed sensing that divides the compressed sensing problem into sub-problems, gaining a better storage complexity. However, block compressed sensing has not yet been studied form a computational complexity perspective. This thesis focuses on the application of block compressed sensing to signals of high dimension to gain insight into the relation between reconstruction performance and computational complexity. This is done by, first investigating how theoretical reconstruction guarantees change, when the problem is divided into smaller sub-problems and by doing a complexity analysis of the reconstruction itself. Each sub-problem solves for a portion of a signal, defined as a block. Next, experiments are conducted in order to get insight into the trade-off between computational complexity and quality of the reconstruction. It can be found that, by using this block-wise approach, the computational complexity of the reconstruction problem decreases, but at the same time, quality of the reconstruction deteriorates. Besides, a method to compensate for this performance loss is proposed. The key idea of this method is that, by propagating prior information among the different blocks, the reconstructions of the blocks can be improved. Finally, block compressed sensing and prior-aware block compressed sensing are analysed in a higher order tensor compressed sensing setting. Nevertheless, this setting was found to exhibit a less favourable complexity-performance trade-off than the normal one, as this setting resulted in both a more complex and a less accurate reconstruction than the normal one.

Algorithms which can effectively detect epileptic seizures have the potential to improve current treatment methods for people who suffer from epilepsy. The current state-of-the-art methods use neural networks, which are able to learn directly from the electroencephalogram (EEG) data without feature extraction. However, neural networks have drawbacks—they are time and data intensive to train and require a high number of parameters for efficient classification. This thesis proposes a novel approach to the epileptic seizure detection problem using Support Tensor Machines (STMs). The final models learn directly from EEG data and use far fewer model parameters compared to a state-of-the-art model using a convolutional neural network. Three types of experiments have been conducted using different representations for the EEG data. The STMs used in the experiments are the Support Higher-Order Tensor Machine, Dual Structure-preserving Kernel (DuSK), and Tensor Train Multi-way Multi-level Kernel (TT-MMK). The results show that by using TT-MMK (with a tensorized data representation) for leave-one-seizure-out validation and DuSK (with the original data representation) for leave-one-patient-out validation, the models are able to rival a state-of-the-art epileptic seizure detector.

This thesis studies the application of the Canonical Polyadic Decomposition (CPD) in unsupervised transfer learning methods for cardiac arrhythmia detection. Unsupervised learning methods have become more prevalent in the healthcare sector due to the abundance of unlabeled data. Labeling of medical data is often non-trivial as it is labor-intensive and requires expert knowledge. Transfer learning can utilize the large number of unlabeled data by extracting relevant features, which can in turn be used for a smaller supervised learning part. Furthermore, in the medical field, AI models are often deployed on embedded systems, requiring efficient model architectures while maintaining high diagnostic performance. The unsupervised transfer learning method was designed with an autoencoder for the unsupervised part and a linear network for the supervised part. The first experiment explored four models to select the most optimal architecture to apply the CPD to. These models include an autoencoder adaptation of the ResNet and ConvNeXt models, the U-Net autoencoder and a basic implementation of a convolutional autoencoder. In the second experiment, the autoencoder model is decomposed using the CPD and evaluated at various compression ratios on its reconstruction capabilities, classification accuracy and computational performance. The CPD implementation is also tested on its convergence speed and data efficiency as compared to its uncompressed counterpart. The first experiment found that the basic implementation of a convolutional autoencoder performed best overall. The U-Net model had high reconstruction quality, however lacked the predictive accuracy. The ResNet model was found to have slightly worse reconstruction and prediction capabilities while having a larger parameter count. The ConvNeXt model failed to accurately reconstruct the images. The second experiment showed the CP-decomposed model approached the uncompressed model in terms of predictive capabilities, while having lower reconstruction qualities. This is likely due to the regularization effect of the CPD, suggesting significant redundancy in the uncompressed model. Despite the reduction in forward pass FLOPs for the CP-decomposed model, it was found that both the computational complexity of the backpropagation process was higher than the uncompressed model at the lower compression ratios and that the memory allocation suffered a significant increase. This resulted in longer and less efficient training of the CP-decomposed models. It was furthermore found that the CP-decomposed models converged faster and had higher data efficiency as compared to the uncompressed model.

This thesis studies the application of the Canonical Polyadic Decomposition (CPD) in unsupervised transfer learning methods for cardiac arrhythmia detection. Unsupervised learning methods have become more prevalent in the healthcare sector due to the abundance of unlabeled data. Labeling of medical data is often non-trivial as it is labor-intensive and requires expert knowledge. Transfer learning can utilize the large number of unlabeled data by extracting relevant features, which can in turn be used for a smaller supervised learning part. Furthermore, in the medical field, AI models are often deployed on embedded systems, requiring efficient model architectures while maintaining high diagnostic performance. The unsupervised transfer learning method was designed with an autoencoder for the unsupervised part and a linear network for the supervised part. The first experiment explored four models to select the most optimal architecture to apply the CPD to. These models include an autoencoder adaptation of the ResNet and ConvNeXt models, the U-Net autoencoder and a basic implementation of a convolutional autoencoder. In the second experiment, the autoencoder model is decomposed using the CPD and evaluated at various compression ratios on its reconstruction capabilities, classification accuracy and computational performance. The CPD implementation is also tested on its convergence speed and data efficiency as compared to its uncompressed counterpart. The first experiment found that the basic implementation of a convolutional autoencoder performed best overall. The U-Net model had high reconstruction quality, however lacked the predictive accuracy. The ResNet model was found to have slightly worse reconstruction and prediction capabilities while having a larger parameter count. The ConvNeXt model failed to accurately reconstruct the images. The second experiment showed the CP-decomposed model approached the uncompressed model in terms of predictive capabilities, while having lower reconstruction qualities. This is likely due to the regularization effect of the CPD, suggesting significant redundancy in the uncompressed model. Despite the reduction in forward pass FLOPs for the CP-decomposed model, it was found that both the computational complexity of the backpropagation process was higher than the uncompressed model at the lower compression ratios and that the memory allocation suffered a significant increase. This resulted in longer and less efficient training of the CP-decomposed models. It was furthermore found that the CP-decomposed models converged faster and had higher data efficiency as compared to the uncompressed model.

The use of artificial neural networks is becoming ever more ubiquitous as the computational power available to use grows. The widespread implementation of neural networks as controllers in the field of systems and control is however being hindered by the lack of verifiability of these controllers. One type of controller that does not lack verifiability is the correct-by-design controller. The main drawback of correct-by-design controllers is that they inherently produce large data structures in order to store their control rules. In this work, a novel methodology to synthesize correct-by-design neural network controllers is presented in order to alleviate these issues. This methodology combines reinforcement learning techniques and abstraction based correct-by-design control verification techniques in order to synthesize neural network controllers that are correct-by-design. The procedure does so by alternating between a controller training routine and a controller verification routine in a system abstraction framework. This framework ensures that numerical training and verification results in a controller with formal guarantees applicable to real systems. Using the proposed methodology, neural network controllers are synthesized and verified in order to prove that the methodology works. In addition to this, the resulting correct-by-design neural network controllers are compared to conventional correct-by-design controllers in order to judge their performance and data requirements. This comparison also includes alternative structures used to store these different controllers. Based on this comparison, a conclusion is drawn regarding when to use which type of controller. The proposed methodology is implemented into a correct-by-design neural network synthesis framework called COSYNNC. This framework is intended as a basis for further research into correct-by-design neural network control and is publicly available.

B-splines are basis functions for the spline function space and are extensively used in applications requiring function approximation. The generalization of B-splines to multiple dimensions is done through tensor products of their univariate basis functions. The number of basis functions and weights that define a multivariate B-spline surface, therefore, increase exponentially with the number of dimensions, i.e. B-splines suffer from the curse of dimensionality. Tensor network theory provides a mathematical framework to alleviate the curse of dimensionality of B-splines by representing the high-dimensional weight tensor as a low-rank approximation. This thesis presents the Tensor Network B-spline (TNBS) model, along with an optimization algorithm that allows the estimation of the exponentially large weight tensor directly from data, without ever needing to explicitly construct it. P-spline regularization is incorporated to induce additional smoothness and ensure the B-spline hypersurface generalizes well across the high-volume domain. The developed TNBS framework opens doors for the application of B-spline theory in high-dimensional function approximation. This thesis provides an overview of both B-spline and tensor network theory, then uses it to derive the TNBS model. We validate the effectiveness of the model through an application in black-box nonlinear system identification using a NARX approach. An open-source MATLAB implementation of TNBS is made available on GitHub. The work is concluded with some recommendations for further research on this topic.

Humans make decisions when presented with choices based on influences. The Internet today presents people with abundant choices to choose from. Recommending choices with an emphasis on people's preferences has become increasingly sought. Grundy (1979), the first computer librarian Recommender System (RS), provided users with book recommendations. Growing volumes of user data in the '90s saw increased usage of commercially available RS for e-commerce, music, movies, books, and social networking services. Due to their effectiveness in providing recommendations, Collaborative Filtering (CF) algorithms are predominantly used to build these RS. However, traditional CF algorithms adopting Matrix Factorization (MF) and Nearest Neighbor (NN) methods suffer from handling sparse data or model scalability. With exponentially increasing sparse data, building scalable and accurate RS models is of focus. This thesis uses tensors and graphs to represent available data. Emphasis is given to capturing higher-order interactions present between the data. The use of tensors is motivated as matrices cannot capture data with higher-order relations, such as variation of user ratings to items with time. The transition to using tensors has been promising with the development of efficient tensor decomposition methods and powerful machines. Graphs can capture the correlation between different entities, providing additional information intrinsic to the underlying graph structure. A Graph Regularized CANDECOMP/PARAFAC (GRCP) tensor decomposition model framework is proposed in this thesis. The thesis highlights how to graph Laplacian regularizers (GLRs) benefit CP tensor decomposition methods to build RS. The model framework is evaluated with the MovieLens data set. The model records lesser Normalized Mean Squared Error (NMSE) values than those reported in the literature. The combination of varied data sources notably aids in overcoming the drawbacks of current RS models, offering scalability with computational efficiency in linear time.

Streaming video completion is the practice that aims to fill in missing or corrupted pixels in a video stream by using past uncorrupted data. A method to tackle this problem is recently introduced called a Tensor Networked Kalman Filter (TNKF). It shows promising results in terms of performance compared to state-of-the-art methods for high percentages of missing pixels (≥ 95%). The main drawback of using a TNKF is the computational speed, which needs to be improved to compete with other existing methods and to be carried out in real-time by a regular computer. This work discusses three methods that reduce the computational load of the algorithm, which speeds up computations. The first method is replacing the existing algorithm with a Block Update TNKF. Secondly, the use of randomized rounding instead of deterministic rounding is investigated. The last method is the simplification of the TNKF update. Results that are presented in this report show that significant speedups of up to +132% can be achieved. In most situations, the considered speedup methods compromise the reconstruction’s accuracy. This thesis discusses the effects this has on the quality of the reconstruction.

Traditionally, Event-Triggered Control (ETC) methods are sample-and-hold control schemes that implement a triggering condition in order to reduce the number of control updates. Given a decay rate of the Lyapunov function, they focus on minimizing the (average) Inter-Sample Time (IST). In this thesis, we focused on the scheduling of Periodic Event-Triggered Control (PETC) controllers. By dynamically switching between triggering conditions, we are maximizing the average rate of decay of the Common Lyapunov Function (CLF) given a minimum Average Inter Sample Time (AIST) or burst condition. Given the physical system, we construct a switched system which captures all possible scheduling behaviors. The l-complete abstraction of the switched system is constructed by solving a conjunction of quadratic equations. By setting a minimum AIST or burst condition, a set of states in the abstraction is marked and a safety game is played to construct the Maximal Permissive Controller (MPC). On the safe behaviors inside the MPC, the guaranteed minimum control performance is maximized for the infinite horizon problem, i.e. by maximizing the minimum weighted time average of the primitive cycles in the MPC. First, several energy games are played to estimate the maximum control performance. Thereafter, a mean-payoff game is played to generate the strategy securing this maximum control performance, which is used to construct the infinite horizon controller.

In modern society cars are one of the most important means of transportation. Unfortunately, many people die in car accidents around the world. Research shows that the number of fatal casualties in car accidents has been increasing for the past decade and that the largest cause of these accidents is the human driver. For this reason, research on fully autonomous vehicles has gained a lot of attention. However, currently autonomous driving is only implemented to reduce the errors of human drivers. More research is necessary in order for fully autonomous vehicles to be implemented and to remove the human driver completely. A robust navigation algorithm which is able to run in real time is one of the challenges in development of fully autonomous vehicles. Important topics in navigation of autonomous vehicles include the path planner and the motion controller. The path planner finds a path for the vehicle from its current location to the target location. At the same time the path planner avoids obstacles and fulfills the non-holonomic constraints of the autonomous vehicle. The motion controller tries to follow the path the path planner made as close as possible by controlling the vehicle. These two topics influence each other and are therefore dependent. In literature little research is done on integrated algorithms that combine path planning and motion control. Therefore, this thesis will research navigation of autonomous vehicles by using an integrated algorithm that includes both path planning and motion control. The objective of this thesis is to develop a Lyapunov stable control algorithm that is capable of planning a path for all possible vehicle maneuvers. Besides path planning the proposed algorithm must be capable of controlling the vehicle along this path. Furthermore, the algorithm needs to include obstacles and the non-holonomic dynamics of an autonomous vehicle. The main contribution of this thesis is an integrated path planner and motion controller for navigation of autonomous vehicles. The stability of the proposed algorithm is proven by using the Lyapunov method. Simulation results prove that the algorithm is capable of planning the path and the motion of the autonomous vehicle with non-holonomic constraints and with the presence of obstacles.