Abstract
Retinitis pigmentosa (RP) and age-related macular degeneration (AMD) are diseases that notably affect photoreceptors of the retina and cause progressive vision loss. This thesis proposed two pulse modulation CMOS imager with integrated sense-and-stimulus (SAS) techniques, which are used to replace the sensing capability of damaged photoreceptor. The first work exhibits a 0.8 V CMOS SAS imager with 4096 pixels for retinal prosthesis. The pixel consists of a pulse-frequency modulation (PFM) photon sensor (for sensing) and a balanced current-mode stimulator (for stimulating) to achieve a highly integrated and low-power solution for high-resolution vision recovery. With the PFM CMOS image sensor, an ultra-low-power operation is achieved. Three operation modes (test mode, programming (PG) mode, and implanted (IP) mode) have been implemented for various purposes. In test mode, the internal signals are multiplexed out serially for chip verification. In PG mode, the output pattern of current stimulator array is programmable by external addresses for patterned electrical stimulus experiments of retina. In IP mode, the chip is fully functional with a minimized number of I/O as 4 for in vivo operation. The second work exhibits a dual-supply high dynamic range (HDR) SAS CMOS imager with adaptive gain control for artificial retina applications. When compared with the first work, the 0.5 V operated pulse-width modulation (PWM) based HDR image sensor is adopted to reduce power consumption with dynamic range extension. In this PWM sensor, the proposed threshold-variation cancelling (TVC) scheme is adopted to efficiently eliminate the FPN and achieve a low-noise image quality in the front-end. The 1.8 V operated in-pixel pulse-to-current stimulator provides a biphasic current pulse with sufficient intensity to activate neuron cells for artificial vision recovery applications. The time-to-voltage (T-V) conversion technique with a programmable gain is employed to achieve a reduced fixed-pattern-noise (FPN) and an adaptive sensitivity. The cellular mechanism in the retina varies its sensitivity at different intensity of light. An adaptive gain control system is also proposed to modulate a suitable response for mimicking the mechanism of retinal cells.