Preignition is an ignition event that happens before it is expected to happen and, many times, where it is not expected to happen. Understanding this phenomenon is of great importance as it influences the design and operation of modern downsized boosted internal combustion engines. To gain a fundamental understanding of preignition, homogeneous reactors like shock tubes and rapid compression machines may be used to decipher the influence of fuel chemical structure, temperature, pressure, equivalence ratio and bath gas on preignition. In this thesis, a comprehensive study of the preignition tendency of various chemical systems is presented. Firstly, renewable fuels like ethanol, methanol and a surrogate of conventional fuels, n-hexane, are characterized by traditional shock tube techniques, such as the measurements of ignition delay times and pressure-time histories, to identify thermodynamic conditions which promote non-ideal ignition behavior. Preignition pressure rise and the expedition of measured ignition delay times are identified as the indicators of non-homogeneous combustion. It is shown that preignition effects are more likely to be observed in mixtures containing higher fuel concentration and that preignition energy release is more pronounced at lower temperatures. High-speed imaging was implemented to visualize the combustion process taking place inside the shock tube. End-wall imaging showed that low-temperature ignition may be initiated from an individual hot spot that grows gradually, while high-temperatures ignition starts from many spots simultaneously which consume the reactive mixture almost homogeneously. Simultaneous lateral and endwall imaging was implemented in both low- and high-pressure shock tube facilities. All tested fuels exhibited localized ignition at low temperatures, and methanol showed a higher propensity than ethanol to ignite far from the endwall. Imaging experiments were also performed in a rapid compression machine to understand preignition at lower temperatures. Herein, ethanol showed non-homogeneous ignition while iso-octane and diethyl ether exhibited homogeneous ignition at the low-temperature conditions. Various criteria for the onset of preignition were tested against experimental observations to propose an adequate predictor of non-ideal ignition phenomena in practical applications. A non-dimensional number, relating the ignition delay sensitivity and laminar flame speed of the mixtures, was found to be the best criterion to elucidate ignition regimes.

UWB has the huge potential to impact the present communication systems due to its enormous available bandwidth, range/data rate trade-off, and potential for very low cost operation. According to FCC, Ultra Wideband (UWB) radio signal defined as a signal that occupies a bandwidth of 500 MHz or fractional bandwidth larger than 20% with strict limits on its power spectral density to -41.3dBm/MHz in the range 3.1GHz to 10.6GHz. Decades of research in the area of wide-band systems have lead us to new possibilities in the design of low power, low complexity radios, comparing with existing narrowband radio systems. In particular, impulse radio based ultra wideband (IR-UWB) is a promising solution for short-range radio communications such as low power radio-frequency identification (RFID), wireless sensor network's and wireless personal area network (WPAN) etc. Since a simple circuit, architecture adopted in the IR-UWB system, the non-idealities of receiver front end may lead to degrade the overall performance. Therefore, it is important to study these effects in order to create robust and efficient UWB system. However, majorities of recent studies are formed on the channel analysis, rather than the receiver system. The main objectives of this thesis work are, (a) System level modeling of non-coherent IR-UWB receiver, (b) Performance analysis of IR-UWB receiver with the help of bit error rate (BER) estimation, (c) A study on the impact of receiver front end non-idealities over BER, (d) Analysis of charge leakage in integrator and its effect on overall performance of UWB receiver. In this work, IR-UWB non-coherent energy detector receiver operating in the frequency band of 3GHz-5GHz based on the on-off keying (OOK) modulation was simulated in Matlab/Simulink. The effect of receiver front end non idealities and integrator charge leakages were discussed in detail with respect to overall performance of the receiver. The results show that non idealities and leakage degrade the performance as expected. In order to achieve a specific BER of 10-2 with the integrator leakage of 25%, the SNR should be increased by 2.1 dB compared to the SNR with no leakage at a data rate of 200Mbps. Finally, integrator design and its specifications were discussed.

<p dir="ltr">Deep learning has achieved remarkable success across various domains, largely relyingon assumptions of ideal data conditions—such as balanced distributions, accurate labeling,and sufficient computational resources—that rarely hold in real-world applications. Thisthesis addresses the significant challenges posed by data non-idealities, including privacyconcerns, label noise, non-IID (Independent and Identically Distributed) data, and adversarial threats, which can compromise model performance and security. Additionally, weexplore the computational limitations inherent in traditional architectures by introducingin-memory computing techniques to mitigate the memory bottleneck in deep neural networkimplementations.We propose five novel contributions to tackle these challenges and enhance the efficiencyand robustness of deep learning models. First, we introduce a gradient-free machine unlearning algorithm to ensure data privacy by effectively forgetting specific classes withoutretraining. Second, we propose a corrective machine unlearning technique, SAP, that improves robustness against label noise using Scaled Activation Projections. Third, we presentthe Neighborhood Gradient Mean (NGM) method, a decentralized learning approach thatoptimizes performance on non-IID data with minimal computational overhead. Fourth, wedevelop TREND, an ensemble design strategy that leverages transferability metrics to enhance adversarial robustness. Finally, we explore an in-memory computing solution, IMAC,that enables energy-efficient and low-latency multiplication and accumulation operationsdirectly within 6T SRAM arrays.These contributions collectively advance the state-of-the-art in handling data non-idealitiesand computational efficiency in deep learning, providing robust, scalable, and privacypreserving solutions suitable for real-world deployment across diverse environments.</p>