We develop a Deep Reinforcement Learning (DeepRL)-based, multi-agent algorithm to efficiently control autonomous vehicles that are typically used within the context of Wireless Sensor Networks (WSNs), in order to boost application performance. As an application example, we consider wireless acoustic sensor networks where a group of speakers move inside a room. In a traditional setup, microphones cannot move autonomously and are, e.g., located at fixed positions. We claim that autonomously moving microphones improve the application performance. To control these movements, we compare simple greedy heuristics against a DeepRL solution and show that the latter achieves best application performance. As the range of audio applications is broad and each has its own (subjective) per-formance metric, we replace those application metrics by two immediately observable ones: First, quality of information (QoI), which is used to measure the quality of sensed data (e.g., audio signal strength). Second, quality of service (QoS), which is used to measure the network's performance when forwarding data (e.g., delay). In this context, we propose two multi-agent solutions (where one agent controls one microphone) and show that they perform similarly to a single-agent solution (where one agent controls all microphones and has a global knowledge). Moreover, we show via simulations and theoretical analysis how other parameters such as the number of microphones and their speed impacts performance.

The scope of the thesis is to estimate the parameters of continuous-time models used within control and communication from sampled data with high accuracy and in a computationally efficient way.In the thesis, continuous-time models of systems controlled in a networked environment, errors-in-variables systems, stochastic closed-loop systems, and wireless channels are considered. The parameters of a transfer function based model for the process in a networked control system are estimated by a covariance function based approach relying upon the second order statistical properties of input and output signals. Some other approaches for estimating the parameters of continuous-time models for processes in networked environments are also considered. The multiple input multiple output errors-in-variables problem is solved by means of a covariance matching algorithm. An analysis of a covariance matching method for single input single output errors-in-variables system identification is also presented. The parameters of continuous-time autoregressive exogenous models are estimated from closed-loop filtered data, where the controllers in the closed-loop are of proportional and proportional integral type, and where the closed-loop also contains a time-delay. A stochastic differential equation is derived for Jakes's wireless channel model, describing the dynamics of a scattered electric field with the moving receiver incorporating a Doppler shift.

A continuous-time description of networked control systems is considered and the parameters are estimated. The discrete-time description is time-varying due to the random time-delays in the wireless links and therefore difficult to work with. Off-line as well as on-line situations are considered for parameter estimation. In the off-line situation, a linear regression is formed and then the parameters are estimated by the least squares method. In the on-line situation, the estimates of the parameters are recursively updated for each time instance. A comparative study of two different parameter estimation approaches is presented. In the first approach, the parameters are estimated by a simple linear regression. In the second approach, transformation of the differentiation operator to another casual and stable linear operator is made in linear regression to estimate the parameters. A numerical study of these approaches is also presented for comparison.

The thesis project conducted at Combitech Arboga aimed to develop a microcomputer-based prototype with a graphical user interface for implementation in a landing system. Today's technical development is progressing rapidly, especially in the defense industry. The need to modernize outdated technical systems is increasing due to the current socio-political situation, so the materials used within the industry need to be reliable and user-friendly to support the personnel using it.

The thesis work was carried out during the spring of 2023 over a period of 20 weeks and began with an analysis of the current system to generate concepts for the prototype. New hardware was developed and implemented using the widely available microcontroller Arduino Mega 2560. A graphical display was also utilized in the project. The objective of the thesis work was to emulate the current layout of the troubleshooting unit’s information on a digital display.

The thesis work encompassed various areas, including circuit design, circuit construction, software development, as well as several tests and simulations. The outcome of the work was a functional prototype that fulfills the specified requirements, thereby establishing a solid foundation for further development and eventual production.

A subset of long-range dependent FARIMA processes is considered. A method for estimating the parameter that describes the long-range dependency of such a process is suggested. The method is based on an asymptotic expression for the covariance function of the process and gives a closed form solution by means of a weighted linear least squares estimate. The variance of the estimate given by themethod is analyzed and, at the same time, the optimal choice of the weighting is expressed. A numerical illustration of the method and the material in the paper is provided

Fractional Gaussian noise, given as the increment of fractional Brownian motion, is a stationary Gaussian process characterized by the Hurst parameter. In the paper, moments based estimators of the Hurst parameter are presented and analyzed with respect to asymptotic variance

Zero crossing data contain information of a process in compact form and is therefore of interest in wireless sensor networks where only reduced amounts of data can be transmitted. When analyzing the properties of certain algorithms using zero crossing data, the cross-covariance between the zero crossing rates of two jointly Gaussian and stationary processes is needed. The evaluation of such a cross-covariance is considered in the paper and an exact numerical expression as well as an asymptotic expression are presented.

This thesis discusses and investigates the estimation of carrier offset frequency in

orthogonal frequency division multiplexing (OFDM) mobile systems. The investigation

starts by using Mobile WiMAX wireless communication specifications described

in IEEE 802.16e as the primary system setup. Under this setup orthogonal

frequency division multiple access (OFDMA) is used as a physical layer scheme; it

also involves the use of pilots in the OFDM symbol for channel estimation.

Although OFDM is resistant to multipath fading, it requires a high degree of synchronisation

to maintain sub-carrier orthogonality. Therefore the level of performance

of the system depends first on the accuracy in estimating the carrier frequency

offset and then the estimation of the channel. Maximum likelihood estimator

is used for estimating carrier frequency offset; its performance under different conditions

for example SNR, number of virtual carriers needed for estimation etc. are

simulated and compared with theoretical results. The optimality of IEEE 802.16e

specifications was also examined during the simulations and results analysis.

Neural networks are ubiquitous components of Machine Learning (ML) algorithms. However, training them is challenging due to problems associated with exploding and vanishing loss-gradients. Gradient clipping is shown to effectively combat both the vanishing gradients and the exploding gradients problems. As the name suggests, gradients are clipped in order to prevent large updates. At the same time, very small neural network weights are updated using larger step-sizes. Although widely used in practice, there is very little theory surrounding clipping. In this paper, we analyze two popular gradient clipping techniques-the classic norm-based gradient clipping method and the adaptive gradient clipping technique. We prove that gradient clipping ensures numerical stability with very high probability. Further, clipping based stochastic gradient descent converges to a set of neural network weights that minimizes the average scaled training loss in a local sense. The averaging is with respect to the distribution that generated the training data. The scaling is a consequence of gradient clipping. We use tools from the theory of dynamical systems for the presented analysis. 

Artificiell efterklang används i ljudinspelning och produktion för att placeraljud spatialt. En algoritm som simulerar efterklang med hjälp av ett filternätverk togs fram.Fokus vid resultat lades på objektiv analys och framför allt spektrogram ochvattenfallsdiagram. Den slutgiltiga modellen gav mätvärden lika målsättningen. Viktigaslutsatser drogs om justering av parametrar och filtermodeller.

The decentralized nature of multi-Agent learning often requires continuous information exchange over a (wireless) communication network, in order to accomplish common global objectives. Uncertainty and delay in communication induce large Age of Information (AoI) for data available at the agents, possibly affecting algorithm performance. In order to understand this, one needs communication models that are representative of practical wireless networks. In this paper, we present a representative model based on the Signal-To-Interference-plus-Noise Ratio (SINR) between pairs of agents. Further, we present a novel medium access control (MAC) protocol that is sensitive to local AoI. Our SINR model facilitates the representation of practical dependency effects like shadowing, fading, interference and external noise. The model is driven by an underlying geometrically ergodic Markov chain, which can represent agent mobility. Our MAC protocol enables that the aforementioned dependency effects decay over time. With this dependency decay, we then control the asymptotic growth of the AoI, to facilitate the convergence of distributed algorithms. Finally, we illustrate our ideas by analyzing the distributed stochastic gradient descent scheme that uses delayed communicated data. 

Existing proposals for positioning in NB-IoT networks based on range estimation are characterized by either low accuracy or lack of compliance with 3GPP standards. While range-free approaches taking advantage of Machine Learning (ML) have been recently proposed as a potential way forward, their evaluation has been carried out only in simulated environments, with the exception of Weighted k Nearest Neighbours (WkNN), recently tested on experimental data. This work inves-tigates four ML strategies for range-free positioning in NB-IoT networks, based on WkNN and its combination with preprocessing and classification algorithms as well as on Artificial Neural Networks (ANNs). The strategies are evaluated on experimental data and are compared based on a set of Key Performance Indicators (KPIs) measuring both positioning performance and computational complexity. Results show that range-free positioning using ML is a viable solution in commercial NB-IoT networks, and that WkNN and ANNs are at the two extremes in terms of a performance/complexity trade-off; intermediate trade-offs can be achieved by combining WkNN with preprocessing techniques and classification models.

An autoregressive (AR) first order model is often used for modelling wireless channels. This is done in spite of the fact that a satisfactory physical explanation for this model has been missing. However, in the recent paper [Feng, Field and Haykin. 2007] derive a model in form of a first order stochastic differential equation (SDE) from a stochastic description of the scattered electric field. After discretizing this SDE, a physically motivated first order AR model is obtained.