This thesis presents the development and validation of a new wireless Impact Echo (IE) system for condition assessment of reinforced concrete slabs. The new IE prototype was compared with other commercially available non-destructive testing (NDT) devices used for similar purposes, namely Ground-Penetrating Radar (GPR) and Ultrasonic Pulse Echo (UPE). Monitoring and structural inspections are critical to effective management of civil infrastructure and NDTs can enhance the quality of condition assessments by providing objective visualizations of the interior of a structural element. The IE method, first developed in the 1980s, has seen few advancements in the last 20 years. The method has been standardized and used on site, but the underlying technology has become outdated. The data obtained from the transducer is difficult to interpret and requires a computer to post-process it before being usable, thus limiting the direct feedback of the method when conducting tests on-site. Because of those limitations and the test being relatively more time consuming than other alternatives, the method is lacking in usability. A new prototype IE device was designed and built by the project industry partner, FPrimeC Solutions. The methodology followed the traditional approach, but it was designed to work with today’s technology. The device is operated wirelessly via a Bluetooth connection, uses smaller-sized electronic components, and connects with a user-friendly interface on a small tablet to set-up the tests and compute the results immediately. The first part of the project focused on product development by testing iterations of the prototype and providing user feedback to improve the device and accompanying software. The second part of the project aimed to validate the new technology using a set of three large reinforced concrete slabs containing artificial defects. The studied points of interest were sound concrete, effect of boundaries and steel reinforcements, vertical cracks, presence of a hollow conduit, artificial voids and delamination. The IE results were also compared with those from commercial GPR and UPE devices. GPR was found to be the quickest method by far, although the results gathered seemed to be limited by the presence of steel reinforcement and also failed to locate certain defects. UPE was a bit slower than GPR, but was generally able to locate more accurately the artificial flaws created in the test specimens. The results showed poor definition of the flaws making it difficult sometimes to properly locate them. The UPE results also seemed to be negatively affected by the presence of reinforcement which were causing frequent abnormal values. Lastly, the IE method was used. This method was greatly improved during the first phase, but it is still a time-consuming method. The value of the data, however, has great potential when compared to the other options. It accurately located most of the flaws generated and was practically unaffected by the presence of steel reinforcing bars. Also, with further analysis of the data, it was possible to determine the depth of some of the flaws accurately. Due to the time-consuming testing phase and the longer analysis of the data required to obtain the higher quality of results, this study suggests that IE is not likely to be the best choice for a general inspection of a large area (depending on the nature of the information needed). Rather, it is suggested to first conduct a general review of the structure using a quicker method like GPR to locate the problematic areas. After that, refining the grid at key locations to test with IE should provide the best quality of data in a reasonable amount of time.

The purpose of this research is to measure the social shifts that take place in a home where artificial intelligent (AI) devices like Echo Dot and Google Home are fully integrated into their everyday life. Research is currently limited, being that the widespread use of these devices is roughly seven years old. Three main outcomes of this study were related to how often Alexa is being used in homes to solve everyday problems, the lack of overall privacy and security concerns users had, and the level of integration into the home as a member of the family. Some limitations and challenges are my ability to compare the households before and after installing these devices in the home; pinpointing when and where the device is used (i.e., room placement); collecting data on whether the device is used often or sparingly; and the depth of interactions these families actually have with the device on a whole. The broader implications behind the increased integration of AI devices is centered around health, labor, social inequality and ethics.

Maintenance and repair of deteriorating civil infrastructure are global problems requiring significant attention and resources. Accurate measurements of civil infrastructure enable lower repair and rehabilitation costs if mitigation techniques are deployed at earlier stages of deterioration. This research describes an infrastructure inspection solution to scan concrete bridge decks for internal cracking at high speeds. Internal cracking within bridge decks, known as delamination, is a particularly difficult defect to identify because it is often not detectable through visual inspection. State-of-the practice testing approaches involve the use of slow and subjective manual sounding techniques and costly lane closures. The need for an improved testing approach has led to decades of research investigating the use of acoustic impact-echo testing to detect bridge deck delaminations. The research presented here consists of a study of the acoustic radiation patterns of delamination defects when they are impacted. Acoustic data were collected on an in-service bridge deck and compared to acoustic data collected on defects in decommissioned bridge deck slabs and on simulated delaminations. This study examined cases of ideal and non-ideal delaminations on the in-service bridge deck and identified characteristics of non-ideal delaminations. An apparatus consisting of a high-speed impact-echo platform and recording suite was designed and constructed. Using this towed apparatus, an order-of-magnitude increase in scanning speed was obtained over other reported methods. Significant design effort was employed to achieve synchronization between different sensing devices using networked computer systems. Analysis was also developed to process and automatically classify acoustic responses to determine the presence and location of delaminations. Demonstrated performance against ground truth data obtained on an in-service bridge deck includes an achievement of approximately 90% probability of detection with only a 2% false alarm rate within 0.30 m. Because of the need to classify acoustic data when ground truth may not be obtainable, a new outlier rejection algorithm, which robustly removes outliers for classification on both simulated and field test data, was also developed. These contributions advance state-of-the-art bridge inspection and also lay the groundwork for additional studies of bridge deck deterioration processes. The framework also demonstrates how a tedious, subjective, and manual inspection process can be automated using advanced excitation tools, signal processing, and machine learning.

No description available.

The diploma thesis called "The Woolgathering" is connected with the idea of gather a wave. It is the image of vanishing moment, effort for hold something that escapes us. Practical part is represented by installation of text, image and audio based on experience, reflection and reverie.

<p> A self-synchronized echo ranging system with optimum utilization of signal estimation and detection strategies has been designed and simulated. A binary convolution code has been utilized to modulate the transmitter signal. The random medium is modelled by a vector sum of a fixed and a random component; the medium fading process has a Rician distribution density. A channel estimator has been derived using a maximum a posteriori probability criterion. The estimator is an adaptive processor whereby the variance of the medium fading process is recomputed during each updating cycle. The estimator attempts to provide a coherent input to the correlator. An optimum processor for the signalling described is an ordered serial estimator-correlator combination. It is conjectured that the estimator offers an improvement in signal processing gain of approximately 5 dB over and above the non-optimized system. Accompanying this is an improvement in peak-to -sidelobe ratio and in false alarm probability. A 3 bit (8 level) quantized system is conjectured to be a 'good' trade-off between degradation in system performance and simplification in system implementation.</p> / Thesis / Master of Engineering (MEngr)

No description available.

This dissertation considers a deep reinforcement learning (DRL) setting under the practical challenges of real-world wireless communication systems. The non-stationary and partially observable wireless environments make the learning and the convergence of the DRL agent challenging. One way to facilitate learning in partially observable environments is to combine recurrent neural network (RNN) and DRL to capture temporal information inherent in the system, which is referred to as deep recurrent Q-network (DRQN). However, training DRQN is known to be challenging requiring a large amount of training data to achieve convergence. In many targeted wireless applications in the 5G and future 6G wireless networks, the available training data is very limited. Therefore, it is important to develop DRL strategies that are capable of capturing the temporal correlation of the dynamic environment that only requires limited training overhead. In this dissertation, we design efficient DRL frameworks by utilizing echo state network (ESN), which is a special type of RNNs where only the output weights are trained. To be specific, we first introduce the deep echo state Q-network (DEQN) by adopting ESN as the kernel of deep Q-networks. Next, we introduce federated ESN-based policy gradient (Fed-EPG) approach that enables multiple agents collaboratively learn a shared policy to achieve the system goal. We designed computationally efficient training algorithms by utilizing the special structure of ESNs, which have the advantage of learning a good policy in a short time with few training data. Theoretical analyses are conducted for DEQN and Fed-EPG approaches to show the convergence properties and to provide a guide to hyperparameter tuning. Furthermore, we evaluate the performance under the dynamic spectrum sharing (DSS) scenario, which is a key enabling technology that aims to utilize the precious spectrum resources more efficiently. Compared to a conventional spectrum management policy that usually grants a fixed spectrum band to a single system for exclusive access, DSS allows the secondary system to dynamically share the spectrum with the primary system. Our work sheds light on the real deployments of DRL techniques in next generation wireless systems. / Doctor of Philosophy / Model-free reinforcement learning (RL) algorithms such as Q-learning are widely used because it can learn the policy directly through interactions with the environment without estimating a model of the environment, which is useful when the underlying system model is complex. Q-learning performs poorly for large-scale models because the training has to updates every element in a large Q-table, which makes training difficult or even impossible. Therefore, deep reinforcement learning (DRL) exploits the powerful deep neural network to approximate the Q-table. Furthermore, a deep recurrent Q-network (DRQN) is introduced to facilitate learning in partially observable environments. However, DRQN training requires a large amount of training data and a long training time to achieve convergence, which is impractical in wireless systems with non-stationary environments and limited training data. Therefore, in this dissertation, we introduce two efficient DRL approaches: deep echo state Q-network (DEQN) and federated ESN-based policy gradient (Fed-EPG) approaches. Theoretical analyses of DEQN and Fed-EPG are conducted to provide the convergence properties and the guideline for designing hyperparameters. We evaluate and demonstrate the performance benefits of the DEQN and Fed-EPG under the dynamic spectrum sharing (DSS) scenario, which is a critical technology to efficiently utilize the precious spectrum resources in 5G and future 6G wireless networks.

Echo state network (ESN) is a specific neural network structure composed of high dimensional nonlinear dynamics and learned readout weights. This thesis considers applying ESN for symbol detection in multiple-input, multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) systems. A new ESN structure, namely, windowed echo state networks (WESN) is introduced to further improve the symbol detection performance. Theoretical analysis justifies WESN has an enhanced short-term memory (STM) compared with the standard ESN such that WESN can offer better computing ability. Additionally, the bandwidth spent as the training set is the same as the demodulation reference signals defined in 3GPP LTE/LTE-Advanced systems for the ESN/WESN based symbol detection. Meanwhile, a unified training framework is developed for both comb and scattered pilot patterns. Complexity analysis demonstrates the advantages of ESN/WESN based symbol detector compared to conventional symbol detectors such as linear minimum mean square error (LMMSE) and sphere decoder when the system is employed with a large number of OFDM sub-carriers. Numerical evaluations show that ESN/WESN has an improvement of symbol detection performance as opposed to conventional methods in both low SNR regime and power amplifier (PA) nonlinear regime. Finally, it demonstrates that WESN can generate a better symbol detection result over ESN. / Artificial neural networks (ANN) are widely used in recognition tasks such as recommendation systems, robotics path planning, self-driving, video tracking, image classifications, etc. To further explore the applications of ANN, this thesis considers using a specific ANN, echo state network (ESN) for a wireless communications task: MIMO-OFDM symbol detection. Furthermore, it proposed an enhanced version of the standard ESN, namely, windowed echo state network (WESN). Theoretical analyses on the short term memory (STM) of ESN and WESN show that the later one has a longer STM. Besides, the training set size of this ESN/WESN based method is chosen the same as the pilot symbols used in conventional communications systems. The algorithm complexity analysis demonstrates the ESN/WESN based method performs with lower complexity compared with conventional methods, such as linear mean square error (LMMSE) and sphere decoding. Comprehensive simulations examine how the symbol detection performance can be improved by using ESN and its variant WESN when the transmission link is non-ideal.

<p>The acoustic echo problem arises whenever there is acoustic coupling between a loudspeaker and a microphone, such as in a teleconference system. This problem is traditionally solved by using an acoustic echo canceler (AEC), which models the echo path with adaptive filters. Long adaptive filters are necessary for satisfactory echo cancellation, which makes AEC highly computationally complex. Recently, a low-complexity echo suppression scheme was presented, the perceptual acoustic echo suppressor (PAES). Spectral modification is used to suppress the echoes, and the complexity is reduced by incorporating perceptual theories. However, under ideal conditions AEC performs better than PAES. </p><p>This thesis considers a hybrid system, which combines AEC and PAES. AEC is used to cancel low-frequency echo components, while PAES suppresses high-frequency echo components. The hybrid system is simulated and assessed, both through subjective listening tests and objective evaluations. The hybrid scheme is shown to have virtually the same perceived quality as a full-band AEC, while having a significantly lower complexity and a higher degree of robustness.</p>