This thesis presents a comprehensive study on enhancing the frame rate of CMOS image sensors through both circuit-level and system-level innovations. Motivated by the growing demands inmachine vision and artificial intelligence applications such as ADAS and industrial automation, this work addresses the limitations of conventional high-speed image sensor designs, which typically require increased ADC sampling rates or a higher number of ADCs—often leading to trade-offs in area, power consumption, and image quality.
The first contribution of this research lies in the analog front-end design, particularly in optimizing the ADC architecture. A column-parallel SAR ADC is developed, with both conventional and advanced two-step structures implemented to reduce noise, improve linearity, and minimize layout area. These ADCs are designed to operate efficiently at 250kS/s while maintaining low temporal noise. Techniques such as analog correlated double sampling (CDS), comparator offset calibration, and FPN suppression using optical black (OB) pixels are applied to ensure image quality despite the high-speed operation. The overall noise model is theoretically analyzed and validated, showing that with proper bandwidth and CDS timing, temporal noise remains within acceptable limits without requiring high-frequency operation.
Beyond the circuit-level improvements, a major innovation is the implementation of an adaptive reconfigurable system that switches between full-frame and region-ofinterest (ROI) modes. Recognizing that in many practical scenarios—such as tracking a moving object in a fixed background—only a small part of the image needs to be updated, the proposed system dynamically identifies and captures a small ROI (e.g., 10×10 pixels) around the object. By capturing multiple ROI images within a single full-frame cycle, the effective frame rate is dramatically improved without increasing ADC speed or interface bandwidth. The intelligent switch network, pixel array design, and address decoder architecture are all tailored to support seamlessmode switching, enabling up to 750,000fps in effective frame ratewhile operating under fixed ADC sampling constraints.
The image data path, from pixel readout to ADC conversion and digital transmission, is carefully coordinated. In zoom-in mode, 300 ADCs sample and convert ROI data across three separate sets per horizontal line, managed by FPGA-controlled timing. Image data is reassembled and analyzed in real-time within the FPGA to update the ROI position based on intensity variation. The image analyzer performs high-speed comparison using a sliding 10×10 mask and updates ROI coordinates frame-by-frame. This feedback loop ensures accurate object tracking while minimizing redundant data processing.
Measurement results confirm that the proposed design successfully achieves highspeed performance and high image quality with improved energy efficiency. The background offset calibration achieves VFPN levels as low as 0.37LSB, and the system avoids the power and timing penalties typically associated with high-speed ADC operation and high-throughput interfaces. The architecture demonstrates a scalable, application-specific approach to image sensor design, suitable for future intelligent sensing systems.
In summary, this thesis delivers a novel image sensor readout architecture that enhances frame rate without compromising image quality or power efficiency. It introduces a practical framework for high-speed operation through a dual approach: a compact, high-performance SAR ADC and a flexible reconfigurable system architecture. Together, these solutions forma robust platformfor next-generation machine vision applications that demand both speed and intelligence. ... Read more

In the rapidly developed market of smartphones, drones, robots and automobiles, 3D sensing based on NIR light detection becomes a standard feature. For example, functions like car parking assistance and pre-crash detection in the automobile’s domain; the facial recognition and field of depth detection in smartphone cameras. As a result, the combination for visible and NIR light detection becomes a wise choice for the future applications. Instead of merging two different types of image sensor pixels together, a new image sensor pixel design is developed in this thesis.
In the thesis, a novel CMOS image sensor (CIS) pixel architecture capable of operating across both the visible and near-infrared (NIR) spectral ranges is presented. The proposed design integrates an n-type four-transistor (4T) pinned photodiode (PPD) stacked back-to-back with a p-type 4T PPD structure. Considering the wavelength-dependent photon absorption properties of silicon, the shallower n-type PD is optimized for collecting electrons generated by visible light, whereas the deeper p-type PD is configured to collect holes induced by NIR illumination. The pixel structure was designed and simulated using Sentaurus TCAD and fabricated using a standard 0.18 μm CIS process on an n-type silicon substrate. The resulting pixel forms a multi-junction photodetector (MJPD) featuring a vertically stacked p+-nwell-pwell-nsub configuration—representing a significant difference comparing to the standard 4T PPD architectures. To understand the electrical behavior of the MJPD structure, a three-series-capacitor model is proposed and analyzed in detail. The theoretical analysis demonstrates good agreement with both TCAD simulation and measurement results.
... Read more

Light detection and ranging (LIDAR) is rapidly emerging as a key technology for depth sensing applications. LIDAR acts as an auxiliary ranging technique, such as radar, thermal image ranging, image recognition, and time-of-flight (TOF) sensor. As a special type of direct time-of-flight (D-TOF) system, this work presents the system design and the hardware implementation of a one dimension time-of-flight (1D-TOF) sensor that uses LIDAR with single-photon avalanche diodes (SPADs). The system is designed for consumer electronics, serving for short-distance target ranging in proximity and face detection.
A typical 1D-TOF LIDAR system is composed of a pulsed laser which emits photons, and a sensor which measures the photons that are reflected back from the target. In 1D-TOF LIDAR systems, the following elements are of great importance: SPAD, time-to-digital converter (TDC), TOF histogram, and depth estimation algorithm.
SPAD is the key component of the system, which is designed to detect a single photon. The target ranging can be estimated by measuring the travel time of the photons. With the help of SPADs, the detection of the photons is converted into a response of the SPAD circuity, in which the optical signal is converted to the electrical signal. And, we can measure the moment in which the electrical signal is active to estimate the travel time of the photons.
TDC is a tool for time measurement. In a 1D-TOF LIDAR system, it measures the exact moment, when a SPAD is triggered by a single photon. The measured time is called the timestamp of the photon.
A TOF histogram accumulates the timestamps of the photons at every time interval. In a TOF histogram, we can observe the time distribution of the photons. In essence, the photons come from two sources, noise and signal. With the help of algorithms, we can derive signal information and reject noise information from a TOF histogram. Also, the algorithm enables us to retrieve the TOF information of the histogram, in which the distance between the target and the sensor lies within.
In this work, the basic concepts of the 1D-TOF LIDAR system is described in chapter 1, and a prototype architecture is discussed in chapter 2. Next, the design is decomposed into stages, including system modeling in chapter 3, system trade-off analysis in chapter 4, algorithm design in chapter 5 and hardware implementation in chapter 6. Moreover, we also propose novel approaches both in system level modeling and algorithm design based on artificial intelligence methods. ... Read more

This thesis describes the integration of temperature sensors into a CMOS image sensor (CIS). The temperature sensors provide the in-situ temperature of the pixels as well as the thermal distribution of the pixel array. The temperature and the thermal distribution are intended to be used to compensate for dark current affecting the CIS. Two different types of in-pixel temperature sensors have been explored. The first type of temperature sensor is based on a substrate parasitic bipolar junction transistor (BJT). The second type of temperature sensor that has been explored is based on the nMOS source follower (SF) transistor of the same pixel. The readout system that is used for the temperature sensors and for the image pixels is based on low noise column amplifiers. Both types of in-pixel temperature sensors (IPTS) have been designed implementing different techniques to improve their accuracy. The use of the IPTSs has been proved by measuring three prototypes chips. Also, a novel technique to compensate for the dark current of a CIS by using the IPTS has been proposed. ... Read more

Diagnostic imaging is a fundamental requirement for the effective treatment of about 25 % of patients globally. Ultrasound imaging is considered the safest, least expensive and most convenient diagnostic imaging modality. Its widespread adoption has been in the area of 2D imaging. However, the accuracy of diagnosis with 2D ultrasound imaging is highly dependent on the experience and expertise of the sonographer because the sonographer has to mentally create a 3D impression from multiple 2D images of the region of interest and this could lead to erroneous diagnosis. 3D ultrasound imaging addresses this concern, however, it drives up the cost of ultrasound imaging. Also, disadvantages such as difficulty in taking quality images, prolonged time in learning how to effectively take these images and human distress during the acquisition of these images have been reported. By leveraging the high level of precision, accuracy and maneuverability provided by robotic systems in conjunction with low-end ultrasound imaging platforms, 3D ultrasound images can be captured, the problem of human distress alleviated and subsequently reducing the cost of 3D ultrasound imaging. This project sought to explore the feasibility of designing a low-cost 3D ultrasound robotic system. The design followed a distributed system approach using ROS (Robotic Operating System) where the data acquisition decoupled from the processing and visualization unit. The performance of the design was tested by constructing 3D ultrasound images of custom made phantoms. These results were compared with 3D ultrasound images of the phantoms acquired using the Philips EPIQ 7 - a high-tier ultrasound machine. The results obtained with this design had a low resolution in comparison with those obtained with the Philips EPIQ 7 however, a 3D point cloud representation of the object embedded in the phantom can be seen. Also, a full 3D image of the phantom could not be acquired due the transducer movement limitations of the design. This design successfully demonstrates the feasibility of integrating low-end electronics with robotic systems to acquire 3D ultrasound images. ... Read more

High speed imagers find applications in many fields such as scientific and medical imaging, automotive applications, machine vision and much more. In this thesis, the design of a high speed, high dynamic range (HDR) CMOS sensor with electronic global shutter (GS) and flexible exposure control is presented. The sensor is designed in the 0.11μm CIS process, features 1k(H) x 1k(V) pixels and achieves frame rates greater that 10.000 fps.A review of the architecture of the sensor is given, along with functional illustrations for each comprising block. The quadrant-based approach is described, along with the selectable region-of-interest capability. The pixel design is a eleven-transistor (11T) pinned photodiode global shutter pixel, implementing HDR by means of two in-pixel capacitors. The design of the pipelined Sample & Hold, column gain and column-level Correlated Double Sampling (CDS) circuits are shown. ... Read more

Imaging sensors are remarkable devices which are able to capture moments and present them in a form available to us for years to come. With the use of material properties and photon particles, charges can be produced which, in combination with electric circuitry, transform into information understandable to the human eye and eventually the brain. When a photon particle hits a silicon photodiode, an electron-hole pair forms. By using readout techniques optimized for noise, area and power, images of different quality can be read out. Imaging concepts such as Signal-to-Noise Ratio, Dynamic Range, Resolution, Contrast and more can be used to describe the quality of the sensor. These concepts of quality are of interest when designing an imaging sensor. Throughout this thesis, design methodologies are applied to both the field of Optical Tomography and Neural Networks.

Optical Tomography is an imaging technique capable of detecting the internal structure of a subject with the use of image sensors. Various Optical Tomography techniques exist such as Diffusion Optical Tomography, Optical Projection Tomography and Optical Coherence Tomography (OCT). Optical Coherence Tomography is of special interest due to published papers which show the potential of the use with CMOS image sensors. The technique is based on the Michelson interferometer which is used in order to retrieve data of micrometer resolution. A feasibility study was conducted on OCT which shows potential for future research due to its recent development using CMOS image sensors. A review of state-of-the-art solutions is presented. With the most recent publication on a CMOS image sensor in OCT, with a dynamic range of 66 dB and a frame rate of 730 frames per second, showed the possibility of retrieving OCT images of higher quality compared to conventional sensors.

Artificial Neural Networks are inspired by the complex structure of our brain. By using its unique way of parallel computation, algorithms are capable of teaching our electronic devices human capabilities such as speech and face recognition. Artificial Neural Networks are identified by neurons and synapses in analogy to the nervous system. In this thesis, pixel design concepts are applied to neural networks where the pixel structure is modelled as a neuron. As a part of a larger project, low power design cells are collected into a combined library and presented. The library, in the end of development, will be used for finalizing and optimizing the complete neural network. One of the project’s challenges was to design a novel absolute value filter. The filter’s importance is presented as a part of a learning algorithm where it takes an absolute value of the difference between two signals. The resulting filter is operational and has the low power dissipation of 87 nW. The area was minimized and includes 10 transistors and a current mirror. The signal gain is around -13.2 dB which shows attenuation of the signal. The amount of gain required for this structure, as part of the learning algorithm, is not yet determined. Future work includes improving the circuit’s gain with a gain boost technique and minimizing noise levels with the use of larger transistors and new technologies.
... Read more

The depletion of domestic reserves and the growing use of sustainable resources forces a transition from the locally produced natural gas with a well-known composition toward the ‘new’ gas with a more flexible composition in the Netherlands. For safe combustion and proper billing, the natural gas must be monitored at both distribution points and households. This calls for a robust and low-cost gas sensor that does not require sample preparation or a lab technician to analyze the results.

Optical absorption spectroscopy is a widely used method for material identification due to its nondestructive and self-referencing properties. However, the gas cell, where the sample to be measured is stored, limits the miniaturization of spectrometers to maintain the sensitivity. The gas-filled linear variable optical filter (LVOF) functionally integrates a wideband optical filter and a gas cell by using the resonator cavity of the LVOF as a chamber. The highly reflective mirrors and high operating order of the LVOF enhances the physical dimensions of the cavity into an effective optical path comparable to an external gas cell. Therefore, when combined with a wideband emitter and a detector array, the gas-filled LVOF enables wafer-level spectroscopy with high sensitivity and selectivity.

In this dissertation, the design, fabrication and characterization of a gas-filled LVOF for the composition measurement of natural gas is described. The functionality of the sensor is demonstrated by actual gas measurements using the main constituents of natural gas; methane, ethane and propane.
... Read more