Hamam Abd

• The rapid evolution of micro and nano-scale technologies has propelled advancements in sensory systems, driving their adoption across IoT, Industry 4.0, autonomous vehicles, healthcare, and robotics. However, conventional amplitude-based sensory systems struggle with adaptability, energy efficiency, and reliable performance under fluctuating environmental and operational conditions. This thesis introduces a Neuromorphic Adaptive Spiking Analog-Front-End with Self-X Capabilities for Sensors (SAFEX), offering a novel and comprehensive design inspired by neuromorphic principles to overcome these limitations. The proposed system combines two key components: the Adaptive Sensor Signal-to- Spike Converter (ASSC) and the Self-Adaptive Spike-to-Digital Converter (SA-SDC), working in tandem to deliver a highly flexible and efficient sensory interface. The ASSC transforms analog signals into spike timings, employing adjustable synaptic weights to ensure alignment of the sensor’s input span with subsequent processing stages. This design eliminates the need for external level-shifting and amplification circuits, streamlining the signal-conditioning process. By leveraging the time domain for signal processing, the ASSC achieves a compact and adaptive implementation suitable for a variety of operating conditions. The transfer function of the ASSC was obtained using measurement processes for various gains. Adjustments in gain extended the duration of the ASSC, increasing the number of bits resolved by the SA-SDC to 11.98 bits. An offset of 77.6 LSB was observed in the transfer function due to specific synapse weights, which was rectified through synaptic weight adjustments. Benchmarking against SoA sensory systems, the Neuromorphic Spiking Sensory System demonstrates superior adaptability and resource efficiency. Initial tests showed that the automatic adaptation mechanism reduced Differential Non-Linearity (DNL) from 0.6 to 0.34 LSB (a 43% improvement) and Integral Non-Linearity (INL) from 2.07 to 0.53 LSB (a 74% improvement) at nominal conditions. Under reduced power supply (from 3.3 V to 2.4 V), the system corrected eight missing codes, improving DNL from 1.53 to 0.4 LSB (a 74% reduction) and INL from 4.4 to 0.46 LSB (a 90% reduction). These performance enhancements surpass those reported in recent SoA designs, which often struggle to mitigate errors without extensive manual recalibration or complex error correction algorithms. Furthermore, the system’s intrinsic self-adaptation at the hardware level directly contributes to yield improvement by automatically compensating for process variations, ensuring that each instance of the SAFEX system continuously selfoptimizes without any supervision or external handling. This eliminates the need for manual recalibration and post-fabrication calibration, effectively lowering production costs. Real-world validation using Tunnel Magnetoresistance (TMR) sensors for angle measurements further emphasized the system’s robustness. When the TMR sensor’s output decreased from 0.225 Vp-p to 0.104 Vp-p due to repositioning from 4 mm to 7 mm, the ASSC adapted by adjusting synapse weights, reducing the angle error from 24.95° to 12.72° a 49% improvement. This level of autonomous correction sets the system apart from traditional sensory interfaces, which lack adaptability and require manual recalibration. In summary, this thesis introduces a transformative Neuromorphic Spiking Sensory System, bridging analog signal processing and time-domain techniques with self-X capabilities self-calibration, self-compensation, self-optimization, and self-repair. The design not only addresses the core technical and industrial challenges faced by existing sensory systems but also establishes a scalable, energy-efficient framework tailored to modern applications. This work takes a step towards dynamic adaptability, robust performance, and yield improvement through hardware-level self-adaptation, building on existing developments in neuromorphic engineering and sensory technology. Future research will focus on extending system resolution, broadening application domains, and integrating higher-level self-X functionalities. With these advancements, the Neuromorphic Spiking Sensory System is well-positioned to meet the evolving demands of edge computing and intelligent, resource-constrained environments.