Photovoltaic devices are increasingly relevant not only as energy generators, but also as dynamic electrical components for applications such as sensing and optical communication. This project focused on adapting an impedance and capacitance measurement setup for both commercial-scale and laboratory-scale solar cells. The setup combines a controlled DC bias voltage with a small AC excitation signal and uses lock-in amplifiers to measure the voltage and current response. From these measurements, the impedance and capacitance behaviour of the solar cell can be determined.
In addition to the electrical measurement setup, the illumination system was characterized using spectral irradiance measurements at different LED heights. A height of 20 cm was selected as the preferred reference height when illumination uniformity over the full commercial-scale cell area is most important. A lower height of 10 cm was used for representative electrical validation measurements, because the increased irradiance produced clearer capacitance and impedance results.
To make the setup suitable for repeated use, software was developed as a control and processing layer between the user and the measurement instruments. The software provides a GUI, automates DC and AC measurement routines, processes measured data, calculates power, impedance, admittance and capacitance, and stores the results in CSV files. Reliability features such as retry logic, safety limits, repeated readings, median filtering and rejection of unstable AC measurement points were included to improve consistency.
The adapted setup was successfully used to measure both lab-scale and commercial-scale solar cells over a frequency range of approximately 5 Hz to 10 kHz and within the required DC bias voltage range. The obtained capacitance and impedance curves showed stable and physically meaningful behaviour, comparable to trends reported in the literature. Although limitations remain, the final system provides a functional and user-friendly basis for future solar-cell characterization and research into dynamic photovoltaic behaviour and photovoltatronics applications. ... Read more

This project was carried out to accompany an NLT module by creating an interactive sound sculpture to motivate high school students to pursue STEM fields. The report discusses multiple solutions to this problem and selects one to develop into a prototype. The sound sculpture has been divided into two parts: signal processing and hardware & sensing. The modules are evaluated against an agreed programme of approach between the project team and its client. The prototype shows promise but cannot yet be rolled out as a standalone product, as classroom tests have not yet been performed. Once completed, the sound sculpture will help students understand the principles of sound waves, physics, and electrical engineering by providing an interactive learning experience. ... Read more

This project was made to accompany an NLT-module by creating an interactive sound sculpture to motivate high school students into STEM-fields. The report discusses multiple solutions to this problem and picks one to turn into a prototype. The sound sculpture has been divided in two parts: Signal Processing and hardware & Sensing. The modules are checked against an agreed upon programme of approach between the project team and its client. The prototype shows promise but cannot yet be rolled out as a standalone product as in classroom test have not been performed. Once complete, the sound sculpture will help students understand the principle of soundwaves, physics and electrical engineering by giving an interactive learning approach. ... Read more

Bidirectional brain-computer interfaces (BCIs) are essential for next-generation neuroprosthetics as they enable both neural stimulation and the simultaneous recording of neural activity. However, a significant challenge in these systems is the presence of stimulation artifacts: large voltage transients that saturate sensitive recording electronics and mask critical neural signals.

This thesis investigates the fundamental causes and characterization of stimulation artifacts in the context of visual prostheses through the lens of an electrode-tissue interface (ETI) model. A comprehensive review of existing artifact reduction techniques is provided, assessing their efficacy and trade-offs regarding implementation complexity and signal integrity. The review serves as a framework for the introduction of two novel artifact reduction techniques that target the residual artifact to permit recording of post-stimulation action potentials (APs).

A residual artifact reduction technique is proposed that replicates and cancels the artifact at the input of the recording analog front-end (AFE). Replication of the artifact is performed through the voltage decay of an RC network with a characteristic time constant equal to the ETI model. Although the concept of this technique is completely novel, the system implementation is impractical due to technological limitations.

A rapid charge reset technique for fast settling of artifact transients is also proposed and implemented with an AFE designed in the TSMC 40 nm CMOS technology node. The system features an AC-coupled low-noise boxcar sampler (LNBS) followed by a switched-capacitor low-pass filter (SC-LPF). The maximum artifact considered for the application of this work can be reduced from 725 mVpp to 2.1 mVpp—a reduction of approximately 50.8 dB—within a recovery time of 50 μs while maintaining a low power consumption of approximately 205 nW per channel and an input-referred integrated noise performance of 7.6 μVrms over the system bandwidth of 300 Hz - 5 kHz. Further reduction of the residual artifact requires minimization of the residual charge across the ETI that is present post-stimulation. ... Read more

This thesis explores the design and evaluation of a novel multi-channel hybrid SAR–RAMP analog-to-digital converter (ADC) for high-density neural-recording probes by leveraging the spatial signal correlation between neighbouring channels. Increasing electrode counts in micro- electrode arrays (MEAs) demand compact, low-power signal acquisition. MEArec-generated neural signals were analysed at probe pitches of 10–50 μm, revealing that neighbouring chan- nels in denser arrays often carry overlapping information. Exploiting this property, a hybrid ADC architecture was modelled in Simulink, combining the low-power binary-search algorithm of a successive-approximation register (SAR) with a ramp (RAMP) converter. System-level simulations show that, for 8-bit resolution, the proposed design achieves a 63% reduction in switching activity compared with a conventional RAMP ADC. The 8-bit, 4-channel, 25 kHz sampling-rate (fs) hybrid ADC was implemented and analysed at system level in MATLAB/Simulink, examining the relationship between input signal char- acteristics and ADC performance. Finally, key circuit blocks were implemented in 1.1 V, 40 nm CMOS, including a sample-and-hold stage, a single-ended-to-differential dynamic switching scheme, a strongARM comparator, a binary-weighted CDAC, and the associated digital logic, demonstrating the practical feasibility of the proposed architecture. ... Read more

The hardware demand to accelerate neural network inference in the context of machine learning (ML) and artificial intelligence (AI) has been rapidly growing. Complex models and the widespread use of AI in various domains have led to the problem of energy budget for AI in datacenters becoming a critical problem. The primary contributor is the frequent data movement between memory and processing units, requiring new computing paradigms supporting fast yet energy-efficient neural network inference. Computing-in-Memory (CIM) addresses the “memory wall” of von Neumann architectures by performing parallel computations directly in the memory array. Various memory technologies, including emerging non-volatile memories (NVMs) such as RRAM, FeRAM, PCRAM, STT-MRAM, and optimised structures of conventional SRAM, have been explored for CIM applications.

Designing CIM systems requires multi-level simulation approaches because device or circuit-level design choices can significantly impact system-level efficiency and accuracy. Simulations for CIM architectures span a broad spectrum, from high-level analytical models that provide quick but coarse estimates to detailed circuit-level simulations that provide accuracy at the cost of scalability. However, this leaves a gap in evaluating CIM architectures at realistic workload scales while maintaining sufficient fidelity. This work targets the intermediate cycle- and system-level domain to address this, aiming to bridge the gap between abstract analytical evaluations and low-level hardware description implementations.

This work develops an event-based CIM architecture simulation framework showcased by the simulation of an accelerator system defined with a target multiply and accumulate (MAC) block architecture. Workloads, including basic MLP, CNN, and NLP models, were executed to analyse metrics such as cycle delay, tile count and utilisation as an efficiency indicator. The impact of tunable simulation parameters was also evaluated considering four defined MAC block topologies. Simulation results show for an MLP workload an increase of 7.6% tile utilisation between two topologies, while reporting on both the number of execution cycles and necessary hardware.

Overall, the CIM architecture simulation platform can effectively serve as a tool for mid-stage design exploration and performance evaluation of CIM-based accelerators. The framework’s modular, eventbased structure enables architectural exploration while providing realistic timing behaviour, distinguishing it from analytical and device-level tools. ... Read more

... Read more

This report outlines the design and implementation of a subsystem within the automated e-nose prototype, which is based on a CMOS Pixelated Capacitive Sensor (PCS) array. The objective of the e-nose prototype is to detect the presence and identify the type of volatile organic compounds (VOCs). A potential use case is to accommodate greenhouses with these e-noses to detect infestations in plants.
The subsystem presented in this report is tasked with programming the FRDM-MCXN947 development board to enable and improve communication and control between the PCS array and the microcontroller unit (MCU). In addition, the development board is also explored to investigate the feasibility of replacing currently used external modules with on-board peripherals. To achieve this, the CTIMER and SCTIMER are analyzed and implemented with the MCUXpresso IDE. The results showed that both of these peripherals have difficulties in simultaneously generating clock signals of different frequencies required for the communication protocol between the sensor and the MCU. However, the SCTIMER is capable of generating a differential clock, currently produced by the AD9552 external clock, though with lower signal quality. ... Read more

This thesis presents the design and implementation of a socket, a chamber and an automated control system for an electronic nose (e-nose) based on CMOS Pixelated Capacitive Sensor (PCS) array chip in a Quad Flat No-Leads (QFN) 5 mm x 5 mm chip package. The objective of this project is to enable repeatable and reliable detection of volatile organic compounds (VOCs) by ensuring a gas-tight sensing environment and minimizing human intervention in experimental procedures, while ensuring good electrical connections by the socket.

A socket was developed to ensure good electrical contact using an anisotropic conductive sheet, allowing the QFN-packaged PCS chip to be inserted and removed without the need for soldering. The socket interfaces with a tightly sealed chamber to prevent gas leakage (< 0.1 ml/min) and maintain a stable gas flow across the sensor. Both components were prototyped in Polyethylene Terephthalate Glycol (PETG) and therefore do not fulfill the final requirement for heat and chemical resistance yet.

To automate gas delivery, the system integrates Bronkhorst mass flow controllers and RVM industrial microfluidic rotary valves. A Python-based graphical user interface (GUI) was developed to schedule gas flow profiles, control valve positions and compute VOC concentrations using the Antoine equation and Dalton's Law of Partial Pressures. This enables dynamic delivery of programmable VOC concentration. Although the automated control system meets most mandatory requirements and all trade-off requirements, some limitations remain. These include unreliable serial communication with the valve during automated operation, even though manual operation through the GUI is executed reliably. This inconsistency is likely caused by timing or firmware-related issues. ... Read more

This thesis addresses the development of a supporting electronics platform for a CMOS Pixelated Capacitive Sensor (PCS) array intended for high-sensitivity applications. The work focuses on enabling accurate extraction of capacitance changes at the sensor interface and translating them into signals suitable for digital processing. The design process emphasizes functional integration and signal integrity.

A custom analog front-end was developed, featuring a transimpedance amplifier (TIA) optimized for low input-referred noise and sufficient bandwidth to preserve pixel-level signal integrity. The complete analog signal path supports capacitive measurements with attofarad-level resolution and readout frequencies ranging from 1 to 100 MHz.

The main PCB integrates the readout circuitry with a central microcontroller (MCXN947), DAC-controlled programmable power supplies, and a variety of user interface connectors, all within a compact six-layer mixed-signal stackup. Particular attention was given to minimizing electromagnetic interference (EMI) and power supply noise through careful grounding, power segmentation, and layout strategies. Although the PCB theoretically satisfies most mandatory and trade-off requirements, including spatial and interface constraints, final verification of the noise performance remains pending due to fabrication lead times. ... Read more

This work contains two systems created to raise abstraction for the Haskell-based HDL Clash.

A common tool in hardware design is the waveform viewer. Although Clash could already generate waveform files, these only contained binary representations of the values. Without translating these to Haskell values, they are difficult to interpret. Shockwaves was created to perform this translation. Unlike other typed waveform solutions, Shockwaves performs the translation fully in the Haskell runtime, and stores the results in lookup tables. This gives the programmer full control over the waveform representation of data. There are two methods of generating VCD files from Clash, and Shockwaves was designed to work with both. The system is fully functional for signals traced during direct simulation. The alternative approach of simulating a design after compiling it to a different HDL depends on the Clash compiler adding type annotations. This requires an overhaul of the Clash compiler beyond the scope of the project.

The second system, Tydi-Clash, is a library for the Tydi streaming specification in Clash. Tydi was designed around transferring complex data structures, and allows for multiple related streams carrying typed, multi-dimensional data. The Tydi-Clash library supports Tydi data types, physical streams, and logical stream constructs. To encourage correct usage of the streams, the internal signals are encapsulated in algebraic and abstract data types that prevent defining or accessing undefined values. Additionally, tests are supplied for behavioral restrictions. An example implementation revealed implementations using Tydi-Clash are unfortunately still a bit cumbersome, but this is believed to be solvable by adding a library of utility modules for common situations. ... Read more

Autonomous planetary rovers require obstacle detection capabilities to navigate hazardous terrain without Earth-based intervention, yet currently deployed methods are limited. This research develops and validates a complete deep learning system for autonomous rock detection on a resource-constrained planetary rover.
We present a lightweight MobileNetV2-based U-Net architecture with dual attention mechanisms, optimized for edge deployment with only 0.31 million parameters. A new dataset MarsTanYard was created via a semi-automated dataset creation pipeline, enabling efficient annotation of Mars-analogue terrain imagery. Our deep learning network was trained on this dataset and integrated with a ROS2 navigation stack through a modular architecture that transforms segmentation masks into 2D occupancy grids that can be used for rover path planning. Our network achieves a 77% intersection-over-union accuracy for rock segmentation, and physical validation on a testing rover in a Mars analogue environment demonstrates a 94% detection rate for large rocks at close-range. An inference time of 4.49ms was achieved on the target rover hardware using model optimization techniques. The system maintains reliable operation across varying lighting conditions with less than 15% performance degradation.
Results show theoretical collision probability of 7.8 × 10^-7 per rock encounter, enabling months of autonomous operation for typical planetary missions. This work provides an end-to-end validation of the deep learning obstacle detection system, establishing a foundation for enhanced rover autonomy in future Mars exploration missions. ... Read more

The usage of robots replacing human tasks has become more prevalent. Controlling multiple of these robots can be useful in applications such as disaster response, surveillance and exploration. This form of control is often achieved by using geometric patterns, such as triangles, squares, etc. Drones can employ this concept of formation control to fly and maneuver through environments and obstacles.
In this work a distributed affine formation control algorithm is implemented onto a Crazyflie drones from Bitcraze. Ultra-wideband is used for positioning and communication between the drones. The implementation of the affine formation control algorithm is optimised such that it is only executed when new information is available, to prevent onboard microcontroller from bottlenecking. This resulted in the drones flying in formation successfully with an accuracy of approximately 6.80 cm from its expected position.
Additionally, this algorithm is extended to manage cases where unexpected missing drones could comprise the stability of the formation. The implementation uses the CMSIS library that is optimised for matrix operations. This resulted in the drones flying in formation successfully even in the case of an observation loss with an accuracy of approximately 17.69 cm.
This work not only provides empirical data of experiments with an affine formation control algorithm, but also provides a baseline implementation for future research in the field of affine formation control, which can potentially lead to noise analysis or the application affine formation controls under different circumstances. ... Read more

This study aims to systematically design and implement a voltage regulation control loop (VRCL) for an unspecified power management unit (PMU) to maintain the desired load voltage for electronics worn by group-housed rodents. This is crucial for analyzing the spontaneous behaviour of group-housed rodents, focusing on social interactions and environmental exploration in a simulated natural environment.

The study explores a potential high-level design of a power link, selecting resonant inductive coupling for wireless power transfer (WPT), enabling continuous power supply over larger areas (for instance, 5 by 5 meters). A literature review reveals a research gap in integrating encircling and underneath configurations for sufficient uniform power distribution. A block diagram of the WPT system is provided, outlining the transmitter and headstage receiver components. Subsequently, a hybrid layout is suggested and significant challenges like optimizing driving current and minimizing angular and vertical misalignments are addressed.

Next, a control loop is systematically designed and implemented. Potential loads and a PMU are
selected, followed by the development and verification of an ideal power converter and its derived
and proposed plant model. Control specifications derived from this model suggest the tuning of the
controller parameters using a tailored model-based control system approach. MATLAB simulations
confirm that the control specifications are met. A non-ideal power converter is then integrated with the control loop, including ideal gate drivers, a voltage-controlled oscillator, a bandgap reference, and a PI controller. Simulation results show that the control loop meets the specifications. Despite limitations in robustness, particularly regarding load and input voltage transient response, the study also highlights the need for verification of the power converter model for load capacitors in the pF-nF range, and identifies discrepancies in overshoot behaviour. Future work includes an analysis of the system’s robustness during controller tuning and the incorporation of a transient controller. The combination of the suggested plant model and model-based tuning approach offer an alternative option for the power converter’s control loop design. ... Read more

The device that is designed and described in this thesis aims to non-invasively stimulate neural activity in an earthworm at multiple, specific locations. This is done by inducing an electric field from two inductors with a specific electric field strength in mind. The response of the earthworm is recorded and amplified to enable analysis of this signal. Initially, requirements are set that need to be met by this device. From these requirements a design process was started and simulations were made. After this, a device was created on which the requirements could be evaluated. Due to the physical trials not being completed the requirements could not be fully verified. ... Read more

We conduct a simulation study of an insect-inspired navigation method that combines visual learning in a small area around a home location with path integration to successfully navigate over distances 8 to 10 times larger than the learning radius, while only requiring 6MB of memory. Our simulations show that the method is capable of learning with on-board data only and handling the noise levels and errors it may encounter in real-world conditions. Furthermore, we investigate the integration of online learning with a structured buffer to mitigate catastrophic forgetting, allowing continuous learning during flight. Additional experiments assess the system's generalization capabilities beyond the learned area, its robustness to small variations in pitch and roll, and the relationship between network size and the learnable area size. ... Read more

To support the exploration of the Moon, wireless sensor networks could be deployed. However, using a very large number of tiny sensing nodes to collect data in such a harsh environment has many challenges. Integrated circuits are exposed to a wide temperature range (-253 °C / +120 °C) during the lunar days and nights. To meet the very constrained power budget of an autonomous node, the node time references must consume ultra-low power while maintaining an accurate frequency. To allow accurate operation over the wide target temperature range and be robust to radiations, this thesis proposes to lock the frequency of a frequency locked loop (FLL) to an LC filter. In the FLL, a high-frequency low-power ring oscillator is locked to a separate LC filter to potentially save power with respect of a standard LC oscillator. To further save power, the FLL is heavily duty-cycled and turned on only to calibrate a lower-frequency oscillator to compensate its drift due to noise and environmental changes, thus achieving ultra-low-power dissipation. This thesis explores the design and implementation of the first LC-based FLL by proposing a system architecture, analyzing its limitations and proposing a transistor-level implementation. The resulting time reference, when implemented in TSMC 40-nm process, consumes 773.29𝜇𝑊 power and achieved an estimated 41.89𝑝𝑝𝑚/ 𝑜𝐶 temperature coefficient within an 0.389𝑚𝑚2 chip area. ... Read more

This thesis project has been done with Crownstone, a subsidiary company of Almende. One of their project goals was to develop indoor localization algorithms to determine room-level location of an asset within the Crownstone network for smart-building, home automation and healthcare applications. An asset is a wireless device transmitting Bluetooth messages that are heard by Crownstones(sensors) and measure the strength of the received signal(RSSI). The terms position and location are used to denote two different concepts in this thesis. Position refers to estimating the specific coordinates, whereas a location refers to a much wider space like a room. In this thesis, we are interested in getting a location estimate.
There are two most widely used and researched localization techniques to determine the location of the asset. First, a model based(MB) method (e.g. Trilateration algorithm) which uses a mathematical model based on distance and second is a data-driven(DD) method (e.g. Fingerprinting algorithm) that relies on existing data, like RSSI to directly get the location. Algorithms are tested on real data collected by the Crownstones at the Almende office(test environment) divided into a finite number of locations or rooms. Metrics are defined based on the requirements of Almende to compare the MB algorithms with the DD algorithms. In this thesis, firstly, a centralized multilateration(MB-C) algorithm is implemented taking into account distances from N Crownstones at the office. Since one of the requirements was to perform in-network localization, a simple averaging consensus based distributed(MB-D) algorithm was selected and compared against the MB-C algorithm. Results show that the MB-D algorithm is faster, scalable and robust against single-point of failure than the MB-C but is less accurate and does not converge to the centralized solution for a noise variance greater than 10dB.
The MB algorithms have limitations in terms of selecting a model, learning the
model parameters and an additional step of mapping the position output of the implemented MB algorithms to a location is also required. To deal with these challenges, a Machine-learning(ML) based data-driven algorithm is proposed. In this, training datasets were iteratively improved with different features. Then, an Ensemble based centralized ML algorithm (DD-C) is implemented, giving a classification accuracy of 65%. Algorithm is further improved by distributed data handling leading to a classification accuracy of 77%. There has been very little to no study on finding the room-level location of an asset in an indoor setting using a distributed ML based data-driven algorithm. A consensus based distributed ML algorithm (DD-D) is proposed that performs local predictions within the Crownstone network using the same globally trained model giving a classification accuracy of 73%.
The results show that the proposed DD algorithms perform better than the MB
algorithms in terms of accuracy and are comparable in terms of prediction time. Results also indicate the proposed DD algorithms are more scalable, robust against noise but are computationally expensive. ... Read more

Single-photon detection is a critical step in both quantum computing and quantum information technology. For instance, the measurement of qubit states and the establishment of entanglement in nitrogen vacancy (NV) center quantum computing requires the detection of single photons. Superconducting nanowire single-photon detectors (SNSPDs) are very competitive devices due to their performance in high detection efficiency, high count rates, low dark count rates, and low jitter. However, the readout electronics are usually implemented at room temperature by connecting the dilution refrigerator through coaxial cables, while SNSPD need to work in a cryogenic environment that is close to the qubits. In order to have better integration and reduce the complexity of the wiring when the number of qubits increases, it is essential to position the electronics close to, or even integrated with, the SNSPD and qubits.
In this thesis, a cryogenic CMOS readout for SNSPD is designed and taped-out using TSMC 40 nm technology. The SNSPD is designed for color-center quantum computing, which is anticipated to work in the wavelength range of 619-620 nm and 625-750 nm, and at a temperature of 1.8 K. The readout electronics are expected to operate at 4 K. The system is required to have a detection efficiency of more than 90 % and a dark count rate of less than 1 Hz. With the help of SPICE dynamic model, the SNSPD is reproduced in Cadence Spectre for circuit design. Active quenching is implemented in the readout architecture, allowing for an increased readout resistor, which improves the output slew rate and count rates without any latching while still keeping a high bias current for a higher detection efficiency. Under a -40 degree Celsius simulation, the readout system achieved count rates greater than 20 MHz, an average jitter of 25 ps rms, and a power consumption of 36 uW, while simultaneously expecting to significantly suppress the dark count rates and after pulses. ... Read more

The problem of finding a target and forming a path to it is known and is well-explored in the field of swarm robotics. Using swarm intelligence, even swarms of simple robots are capable of solving this problem. However, often these works assume that two robots can perceive each other can physically reach each other, or that robots can detect any target within communication range. These assumptions do not hold for all swarms and environments. This thesis introduces the CSP (Concurrent Searching and Pathfinding) algorithm to make path finding possible in simple swarms without relying on these assumptions. By strategically constructing and reconstructing an ephemeral network of stationary robots and relocating the rest of the robots throughout this network, swarms performing CSP can efficiently find targets with minimal risk of fracturing the network. The effect of packet loss, and swarm size, as well as the the execution time vis-à-vis creation of new anchors are shown. ... Read more