Name: Qosay Wattad Seminar Title: Integrated Phased-Array RFIC Transmitter with Simultaneous Data and Coupling Indication Transmission for Shape Estimation of Flexible Arrays Name: chen chefer Research: super resolution and image reconstruction in event cameras

Full Name: Qosay Wattad Degree: Master’s Degree with Thesis Advisor: Dr. Matan Gal Katziri Seminar Title: Integrated Phased-Array RFIC Transmitter with Simultaneous Data and Coupling Indication Transmission for Shape Estimation of Flexible Arrays Abstract: This research presents the design and implementation of an integrated phased-array RFIC transmitter aimed at enabling both high-speed data transmission and coupling indication for the estimation of flexible phased-array surface curvature. The chip is implemented in TSMC 65 nm CMOS technology and combines two transmitters and one receiver, designed to comply with 5G 3GPP requirements. The architecture allows real-time monitoring of array deformation by transmitting auxiliary coupling information alongside communication signals, which enables curvature reconstruction without external sensors. The expected outcomes include the fabrication and measurement of a working prototype, along with the evaluation of its performance in terms of efficiency, stability, and shape-sensing capabilities. This work contributes toward the development of adaptive, conformal phased-array systems for next-generation communication platforms. Name: chen chefer Electrical Engineering Supervised by prof. Adrian stern Research: super resolution and image reconstruction in event cameras Abstract —Event cameras provide asynchronous, event-driven measurements with unique advantages: ultra-high temporal resolution, wide dynamic range, and low power consumption. However, their relatively low native resolution together with the unique data structure makes high-quality image reconstruction a significant challenge. In this work, we study the possibility of achieving image super-resolution by exploiting the high temporal resolution of the camera. Motivated by this, we depart from common algorithms that primarily process the space domain and add the temporal information, and instead we propose a time-based event alignment method that reconstructs images by accumulating events temporally along the motion trajectory instead of spatially accumulate them. This physically motivated approach reduces sensitivity to pixel threshold variability and unknown initial conditions while enabling sub-pixel detail recovery. Simulations and experiments demonstrate that our method can reconstruct spatial frequencies up to 1.25 cycles per pixel (CPP) in simulation and 0.91CPP in real-world data, far surpassing the Nyquist limit of the native sensor array (0.5CPP) .