In recent years, there is a growing demand of wireless sensory devices, especially in biomedical applications. The circuits and systems deployed in such applications require a low-power and low-complexity design. One of the most difficult challenges in this field is that the large data amount generated by the sensors is hard to be transmitted due to the limited speed of the low power wireless device. In order to solve this problem, in this dissertation we explored the asynchronous analog and RF signal processing techniques, and demonstrated the feasibility through circuits and systems implementation.;In the beginning of this thesis, wireless bio-potential recording system will be introduced to present the background and motivation of this study. In order to minimize the power and complexity of the quantization circuits and reduce the total output data size, the fixed window level crossing sampling method is applied in the sensor. Since the input spike signals are with low activity ratio, by taking the advantages from this signal property, the idea of saving the system power is to shut down the circuits when the input signal is silent. To achieve this, we used asynchronous delta modulator as the analog to digital convertor in the sensor. In order to further reduce the power and circuit complexity, we proposed and designed a fixed window level crossing sampling method. This method minimized the power and complexity of the quantization circuits. When the resolution is less than 7 bits, the data compression rate is more than 50%. Asynchronous pulse sequences will be generated to represent the spike signal. The integrated 4-channel sensor front end was fabricated with AMI 0.5um CMOS process, with total power consumption at 0.5mW when the sampling rate is 60kS/s with 5-bit resolution.;In the wireless part, we use a non-coherent UWB impulse radio system as the wireless data link in order to minimize the power consumption for the asynchronous wireless sensor and communication system. We demonstrate a UWB impulse radio transmission system with a integrated front-end transmitter and a receiver using off-the-shelf-component. The system can support 14Mbps data rate. The circuits and systems are applied to a wireless neural recording system and a wireless temporal difference image sensor with Manchester Encoding / Decoding to provide data and clock recovery. Experiment results from the two systems are reported.;Another challenge in the low power wireless system is to perform power efficient clock and data recovery. Current solutions use digital baseband circuits, phase locked loops circuits and a synchronization header in the preamble of data package. Those methods fall short of meeting the requirements of the low power low complexity transmission, because the wireless circuits are running even when the sensor input signal is inactive. That means the system wastes power if the input signal is with low active rate. Also, the training time of the phase locked loop by synchronization preamble will reduce effective data rate in the wireless link.;To solve this problem, we proposed a non-coherent FSK-OOK modulation and demodulation for UWB impulse radio. The radio uses different carrier frequencies to represent the symbol 1 and 0 while using on-off keying modulation to isolate two adjacent symbols in each frequency band. By doing so, the receiver can perform data and clock recovery simultaneously without PLL training or wake-up time. The pulse duration is controllable to guarantee the low duty cycling of the pulse in order to meet the UWB mask from Federal Communication Commission. The wireless link can be silent when there is no event or no data to transmit. Also the digital baseband and phase locked loop are not necessary in the FSK-OOK receiver. This could save power and silicon area of the receiver. We have fabricated the integrated FSK-OOK UWB transmitter on 0.25um Silicon on Sapphire process. The system can send and receive data with a maximum 30Mbps data rate. This modulation method is an ideal candidate for wireless bio-potential recording systems, temporal different image sensors or other sensor network systems.;At the end of this thesis, we summarized the contribution of this work. Based on this thesis, we also proposed some potential projects for future such as wireless address event representation systems using the asynchronous wireless sensory technologies.
展开▼
