This paper presents a comprehensive investigation into novel Recurrent Neural Network (RNN) architectures for enhancing the efficiency and performance of Digital Predistortion (DPD) for Radio Frequency (RF) Power Amplifiers (PAs). The research first introduces Delta Just Another Network (DeltaJANET), a computationally efficient model that leverages a sparse delta update rule. By updating only a small fraction of the hidden state at each time step, DeltaJANET significantly reduces the complexity and computational cost associated with traditional RNNs, establishing a new baseline for efficient DPD solutions. Building upon this foundation of efficiency, the paper then proposes a second, more powerful architecture: the Temporal Convolutional Just Another Network (TC-JANET). This advanced hybrid model synergistically combines a Temporal Convolutional Network (TCN) for long-range feature extraction with a lightweight Just Another Network (JANET) unit. Key innovations, including a direct memory input module and a dynamic phase normalization scheme applied to the recurrent state, enable the TC-JANET to robustly model complex PA behaviors. A systematic multi-seed evaluation demonstrates the exceptional performance and scalability of this architecture, showing that it significantly surpasses existing benchmarks and establishing a new benchmark for high-performance DPD solutions.

Occupancy grid mapping is a method of representing the environment and its obstacles and drivable areas in a discretised grid that is usually constructed with point cloud data from sensors such as radars. These maps facilitate path planning and decision-making in autonomous driving applications, making them a crucial component. This thesis addressed the problem of dynamic occupancy grid mapping, which additionally models moving objects. The aim was to jointly model the spatial structure and temporal dynamics of the environment. Building on an existing sparse Bayesian learning framework exploiting sparsity and spatial structures, we proposed a dynamic extension that integrates radar range-rate measurements into a motion prediction module. This module predicted the future positions of moving objects directly from the point cloud and incorporated them into the Bayesian inference process by altering the Gamma hyperprior, enabling better tracking of moving objects. The proposed method was evaluated on the real-world View-of-Delft dataset containing urban driving scenarios and compared against three other benchmark methods. Experimental results demonstrated superior detection of moving objects and improved shape reconstruction without significantly increasing false positives.

Not only did the fffth generation mobile communication system increase the data rate, capacity and coverage, but also raise a public concern regarding human exposure to electromagnetic ffelds (EMFs). This thesis project assessed EMF exposure in practical indoor scenarios with a beamforming transmitter through electromagnetic and biologic level of simulations and a validation framework. 3D indoor environments were constructed based on measured dielectric constants of materials, enabling realistic ray-tracing simulations in beamforming cenarios. The obtained simulation results were then combined to a multi-layer skin model to evaluate localized absorption mechanisms. This study also highlighted the inffuence of environmental geometry on exposure levels. An algorithm of bridging the gap between indoor wave propagation and exposure assessment regarding biological tissues was proposed for the ffrst time. It is illustrated by the results that beamforming technology introduced time and user dependency to the exposure levels, while these predicted values remained well below the limit in ICNIRP of 20 W/m2.

Synthetic Aperture Sonar (SAS) is an advanced sonar imaging technique that uses multiple pulses or high duty cycle waveforms from a moving surface vehicle to create a synthetic aperture array for imaging. One of the key challenges of SAS systems is the design of continuous waveforms that enable high-quality imaging under practical constraints (e.g., unimodular constraint). This thesis addresses the design of such waveforms and observes that waveform design for a SAS system over a short time window can be approximated as a wideband ambiguity function shaping problem. To tackle this, we formulate the wideband ambiguity function shaping problem as a non-convex optimization problem and propose four methods to solve it. Among the proposed methods, the wideband gradient descent method is proven to be the most efficient and effective in minimizing the average sidelobe energy in the region of interest of the wideband ambiguity function. Simulation and field trial results show that, although waveform design for SAS systems is not strictly equivalent to wideband ambiguity function shaping, the waveform obtained through this approach still yields good SAS imaging performance compared to conventional sonar waveforms, such as the random Binary Phase-Shift Keying (BPSK) waveform. These findings provide a direction for the waveform design of future SAS systems. Beyond SAS, the proposed wideband ambiguity function shaping methods also show potential for other wideband applications, such as waveform design for underwater target tracking.

Terahertz technology has received continuously increasing attention in recent years. A key enabler of this technology is the photo-conductive antenna. A time-domain Norton equivalent circuit representing the antenna and a time-step algorithm provide the tools for analysis and design of such structures. However, the antenna impedance impulse response must be known beforehand. In this thesis, a more comprehensive characterization of the algorithm is provided through a detailed investigation. A numerical error in its output is discovered, the source of which identified, and an error minimization solution proposed. Using the newly acquired knowledge of the algorithm’s properties, an infinite photo-conductive slot is analyzed. First, a time-domain energy balance equation is derived using switched capacitor formalism, and verified. Then, a comparison to constant antenna impedance approximation is performed. A study on the frequency and time-domain far-fields, as well as, on the influence of parameters is also performed. Finally, the analysis is extended to a 1D connected array, in which the mutual coupling and performance are investigated.

Conventional techniques for audio monitoring, including Passive Audio Monitoring (PAM), often call for the use of environmentally unsustainable and challenging equipment. This among other limitations can disrupt the very environment that is meant to be studied and may also limit the ability to observe an ecosystem in its natural state.

This thesis explores the possibility of a passive wireless biodegradable microphone for the purpose of audio monitoring in fragile environments. This project introduces the "The BioThing - A biodegradable microphone for monitoring biodiversity".
The BioThing uses a biodegradable material, to address the current challenges of state-of-art monitoring techniques. The design utilises a resonant cavity to function as a listening device through the implementation of the backscatter wave modulation technique and with a primary working resonant frequency of around 1.7 GHz.

This report presents a detailed analysis of the device's design and function to optimize its geometry for various frequencies, operational ranges, and functionalities by investigating the effect of critical parameters such as the variable capacitances namely the distance between the diaphragm and the tuning post, and the distance between the tuning post and the antenna. Additionally, it also analyses the effect of device geometry on resonance properties by investigating the impact of parameters such as tuning post height, membrane thickness, and cavity height. The report also investigates the possibilities for device enlargement for long-distance operation and device miniaturisation to facilitate faster degradation. This optimization process aims to achieve improved performance, ultimately leading to a more effective listening device. A prototype was fabricated using commercially available Zinc and plant-based resin.

This research works towards a new generation of ecologically sustainable audio monitoring equipment that will allow a deeper insight into the soundscapes of fragile environments without disturbing the ecosystems.

In this thesis, a direction-reconfigurable wide-input-range radio frequency energy harvesting
system is described. In order to improve the receiving power and the performance in low
power range, three blocks of optimization are placed in different part of the system. Subse
quently, a two paths 3-stage body biased cross-coupled differential-drive rectifier with antenna
beam forming control and 3-bit capacitor bank is designed and fabricated in a standard 180nm
CMOS technology. The rectifier is brought at resonance with four high-Q patch antennas by
means of a control loop that compensates for any variation at the antenna-rectifier interface
and a beam forming control that change the largest receiving direction of the antennas to allow
higher power input in multiple directions. A body-bias method with cross-coupled capacitors
and a diode to limit the body-bias voltage makes the rectifier works better in low power range.
The chip is integrated with four patch antennas. Measurements in an anechoic chamber at
2.45 GHz demonstrate a–10dBm sensitivity for 1 V output across a capacitive load. The
end-to-end power conversion efficiency reaches 65% at–6dBm.

New generations of mobile communications are much more demanding in order to provide critical real-time services and extremely fast connections that have become feasible with 5G. A wider spectrum is needed to achieve such high data rates and millimeter-wave frequencies are used for this purpose. Adaptive beamforming is applied at base stations through active phased antenna arrays to shape and steer the beams at each user, so coverage distance is extended and resource sharing among users is more efficient.

Different approaches to the design of active phased arrays for mm-wave base stations to realize the "one beam per user" concept are analyzed and compared. Particular attention to mutual interference between the beams is given. A novel aperiodic, uniformly-fed, active integrated antenna array design is presented as an appealing alternative to minimize user inference based on a multi-disciplinary array synthesis approach. Implementation challenges of such aperiodic array are investigated and the performance is compared to those of a periodic array. The complete design and its simulation results show an interference reduction around 6 dB and similar gain as compared to the regular array taken as reference.