The various elements of a modern target tracking framework are covered. Background theory on pre-processing, modelling and estimation is presented as well as some novel ideas on the topic by the author. In addition, a few applications are posed as target tracking problems for which solutions are gradually constructed as relevant theory is covered. Among considered problems are how to constrain targets to a region, use state-independent measurements to improve estimation in jump Markov models and how to incorporate observations sampled at an uncertain time into a state-space model. A framework is developed for tracking dolphins constrained to a basin using an overhead camera that suffers from occlusions. In this scenario, conventional motion models would suffer from infeasible predictions outside the basin. A motion model is developed for the dolphins where collisions with nearby walls are avoided by turning. The basin is modelled as a polygon where each point along the edge influences the turn rate of the dolphin. The proposed model results in predictions inside the basin, increasing robustness against occlusions. An extension to a Gaussian mixture background model providing a degree of confidence for detections is used to improve tracking in the presence of shadows. A probabilistic data association filter is also modified to estimate the dolphin extension as an ellipse. The proposed framework is able to maintain tracks through occlusions and poor light conditions. A framework is developed for estimating takeoff times and directions of birds in circular cages using an overhead camera. A jump Markov model is used to model the stationary and flight behaviours of the birds. A proposed extension also incorporates state-independent measurements, such as blurriness, to improve mode estimation. Takeoff times and directions are estimated from mode transitions and results are compared to manually annotated data. The cameras are inaccessible in both applications, disallowing proper calibrations. As an alternative, a method is proposed to estimate stationary camera models from available data and known features in the scene. A map of the basin and the funnel dimensions are used respectively. The method estimates a homography and distortion parameters in an invertible mapping function. An extension to the linear Gaussian state-space models is proposed, incorporating an additional observation with an uncertain timestamp. The posterior distribution of the states is derived for the model, which is shown to be a mixture of Gaussians, as well as some estimators for the distribution. The effects of incorporating the observation with an uncertain timestamp into the model are analysed for a one-dimensional scenario. The model is also applied to improve the GPS position of an orienteering sprinter by using the control position as an observation with an uncertain timestamp. / LINK-SIC

In model predictive control (MPC), an optimization problem is solved at each time step, in which the system dynamics and constraints can directly be taken into account. The MPC concept can be further extended to the control of hybrid systems, where a part of the state and control variables has a discrete set of values. When applying MPC to linear hybrid systems with performance measures based on the 1-norm or the∞-norm, the resulting optimal control problem can be formulated as a mixed-integer linear program (MILP), while the optimal control problem with a quadratic performance measure can be cast as a mixed-integer quadratic program (MIQP). An efficient method to solve these non-convex MILP and MIQP problems is branch and bound (B&B) which relies on solving convex relaxations of the problem ordered in a binary search tree. For the safe and reliable real-time operation of hybrid MPC, it is desirable to have a priori guarantees on the worst-case complexity such that the computational requirements of the problem do not exceed the time and hardware capabilities. Motivated by this need, this thesis aims to certify the computational complexity of standard B&B methods for solving MILPs and MIQPs in terms of, e.g., the size of the search tree or the number of linear systems of equations (iterations) that are needed to be solved online to compute optimal solution. In particular, this knowledge enables us to compute relevant worst-case complexity bounds for the B&B-based MILP and MIQP solvers, which has significant importance in, e.g., real-time hybrid MPC where hard real-time requirements have to be fulfilled. The applicability of the proposed certification method is further extended to suboptimal B&B methods for solving MILPs, where the computational effort is reduced by relaxing the requirement to find a globally optimal solution to instead finding a suboptimal solution, considering three different suboptimal strategies. Finally, the proposed framework is extended to the cases where the performance of B&B is enhanced by considering three common start heuristic methods that can help to find good feasible solutions early in the B&B search process.

No description available.

Dynamical network systems are complex interconnected systems describing many real world problems. The current trend is to connect more and more systems together, and at the same time requiring continuous availability. To this end, it is crucial to understand the dynamic behaviors of networked systems.This thesis makes three contributions in this area. First, we study the important problem of gathering data that are distributed among the nodes in a network. Two specific tasks are considered: to estimate the size of the network, and to aggregate the distribution of local measurements generated by the nodes. We consider a framework where the nodes require anonymity, and restricted computational resources. We propose probabilistic algorithms with low resource requirements, that quickly generate arbitrarily accurate estimates. For dynamical networks, we improve the accuracy through a regularization term which captures the trade-off between the gathered data and a-priori assumptions on the dynamics. In the second part of this thesis, we consider a dynamical network system where one node is misbehaving due to a failure. We specifically seek robustness conditions that guarantee that the entire network system is still functional. The nodes' dynamics is governed by consensus updates, and we present thresholds on the interaction strengths that determines if the system will reach consensus, or if the system will diverge. Finally, a P2P network is utilized to improve a live-streaming media application. In particular, we study how an overlay network, constructed from simple preference functions, can be used to build efficient topologies that reduce both network latency and interruptions. We present necessary and sufficient convergence conditions, as well as convergence speed estimates, and demonstrate the improvements for a real P2P video streaming application. / <p>QC 20131111</p>

The next generation electric power system, known as Smart Grid, is expected to alleviate the energy shortage problem by exploiting renewable energy resources. The Smart Grid communication network, with its diverse structure, constitutes an indispensable component in the new power system. In terms of power industry standards, the International Electrotechnical Commission (IEC) 61850 framework is of particular note. Originally defined to cover the stringent requirements for automation within the substation, IEC 61850 is proving to be a versatile standard that can also be applied to the medium- and low-voltage networks while facilitating control applications. Long Term Evolution (LTE) appears as a remarkable candidate for supporting remote automation tasks in the electricity grid, offering low latency, high throughput and quality of service differentiation in a single radio access technology. In the context of the thesis, a performance evaluation of the integration of LTE technology with IEC 61850 communication services is carried out. A characterization of the network architecture and the performance requirements for intelligent power system management is performed and an analytical model for the scheduling framework is proposed. Emphasis is given on the development of optimal prioritization schemes and techniques in order to ensure control data scheduling in situations when LTE background traffic coexists in the network. In addition, realistic communication scenarios specifically designed to emulate real network operations are considered and extensive simulations are performed with the use of Ericsson’s radio system simulator platform. The results have demonstrated that LTE networks successfully meet the performance requirements for wide-area automation tasks within a Smart Grid context. Given the size of the LTE ecosystem, such an evolution constitutes an attractive path for future wireless communication.

House heating is responsible for substantial energy consumption in industrialized countries. However, the climate control in detached houses is often sub optimally done. These climate controllers are usually composed by subsystems that have limited information about the state of the building. Improvements usually requires large investments from the house owner, unless new sensor networks technologies are adopted. In this Master thesis, how to improve the eciency of house heating using instant messaging (IM) is investigated. Devices in an inhabited house have been congured to be connected by an IM client. This allows to perform automatic control through a simple software. The dynamical model of the heating and cooling system is studied and the relevant sensor measurements are identied. The possible reduction in heat consumption is quantied by simulations. It is shown that the method is applicable, and that it can result in a reduction of the heat consumption in detached houses, even by using a small subset of potential sensor measurements.

Many companies today, in different fields of operations and sizes, have access to a vast amount of data which was not available only a couple of years ago. This situation gives rise to questions regarding how to organize and use the data in the best way possible. In this thesis a large database of pricing data for products within various market segments is analysed. The pricing data is from both external and internal sources and is therefore confidential. Because of the confidentiality, the labels from the database are in this thesis substituted with generic ones and the company is not referred to by name, but the analysis is carried out on the real data set. The data is from the beginning unstructured and difficult to overlook. Therefore, it is first classified. This is performed by feeding some manual training data into an algorithm which builds a decision tree. The decision tree is used to divide the rest of the products in the database into classes. Then, for each class, a multivariate time series model is built and each product’s future price within the class can be predicted. In order to interact with the classification and price prediction, a front end is also developed. The results show that the classification algorithm both is fast enough to operate in real time and performs well. The time series analysis shows that it is possible to use the information within each class to do predictions, and a simple vector autoregressive model used to perform it shows good predictive results.

This thesis focuses on the collaborative action planning for humanoid robot agents alone with each other or with a human agent. The action performed by these agents consists of exchanging a small deformable monochrome object using one arm. This situation occurs for instance when robots are employed to help elderly or disabled persons at home or retirement houses get an object dif- cult to reach, tidy rooms,etc. The used robots are NAOs made by Aldebaran Robotics. The thesis covers details about behavior trees which are a formal method used to plan the exchange and synchronize the NAOs, the Robot Operating System ROS which gives a global structure for the computer code, and NAOs with their software environment. This work also focuses on the image processing algorithms (HSV-lters, Connected- Components Labeling and Union-Find) that were used to detect an object and get its contour in the NAO's camera image, and the geometric applications that allow to get the 3D-position of a quasi-spherical object, knowing its dimensions and its projection on a camera focal plane.

This thesis focuses on temporal logic-based motion planning for autonomous ve- hicles. Specically planning based on Linear Temporal Logic statements. This type of motion planning allows for automatic generation of correct by construc- tion controllers that implement missions dened in a language that is both quite expressive and easy to understand. The main contribution of this thesis is to provide a framework of reuseable Python modules that implement algorithms for the production of plans based on Linear Temporal Logic and ecient online revising of existing plans when the environment changes. The general case of oine planning is presented as well as several options for online planning based on partially known and dynamic environments. The dif- ferent online planning methods are simulated and their feasibility in dierent scenarios is discussed. Finally, implementation concerns and future directions are discussed.

For a matter of efficiency and robustness, it is often better to use a team of robots instead of a single agent to solve a given problem. A key challenge with multi-robot systems is the collaboration in order to accomplish complex tasks. To coordinate them, we can pre-compute their behavior. However, this method might not be robust to some events such as modification of environment or robots team. To overcome this issue, an adaptive decentralized coordination framework is needed for heterogeneous multiple robot systems. We consider a team of two robots NAOs which can only exchange information when they are close to each other, or via symbols grounded to each embodiment. They are initially in a room a few meters away from each other. The goal is to make them meet and then perform an action such as exchanging an object or some information. In this thesis, we study first robots specifications and adopt tools used for robot control. A tracking method in a simple situation is then described. The robots’ strategy is structured and improved adding obstacles limiting the two agents’ motion. The achieved robust framework allows two humanoid robots to meet, even if one has a problem and can not move.