Font Size: a A A

Architectural Design Of Ultra-Low Power Image Sensor Based On Information-Time Transformation

Posted on:2021-02-16Degree:MasterType:Thesis
Country:ChinaCandidate:Z P LiFull Text:PDF
GTID:2518306107493154Subject:Engineering
Abstract/Summary:PDF Full Text Request
Image sensor,a device for converting optical signal into electric signal,have been widely used in digital cameras,electronic optics,medical imaging,and other fields.Up to now,image sensors are mainly divided into two types:charge-coupled device(CCD)and complementary metal oxide semiconductor active pixel sensor(CMOS Active pixel sensor).Compared to CCD,CMOS image sensors manufactured using standard CMOS processes can integrate timing and control electronics,sensor arrays,signal processing electronics,and analog-to-digital converters on the same chip.It works on standard logic supply voltage and consumes very little power,which has gradually replaced CCD in most fields in recent years.The brain-computer interface system,a bridge between the brain and electromechanical devices,is an important field of contemporary scientific research.It is widely used in the field of medical nerve repair,such as cochlear implants,artificial retinas,artificial prostheses,etc.,in order to realize the control of the neural prosthesis by the brain.First,it is necessary to collect the nerve signals of the brain through a digital imaging system,and then turn the nerve signals of the brain into electrical signals and then analyze and process them through subsequent circuits to control these neural prostheses.The method of collecting brain signals is traditionally realized by electroencephalogram(EEG),while this paper uses an intrusive image sensor.To avoid damaging brain cells with too much heat from the chip,the image sensor chip is required to consume as little power as possible.Therefore,low power consumption has become an important research direction.In this thesis,an ultra-low power image sensor based on information-time transformation(Pulse-Width Modulation,PWM)is proposed.Firstly,the light intensity signal is converted into a current signal by photodiode.Secondly,the current signal is converted into a voltage signal and sent to the information-time transformation circuit.In the information-time transformation circuit,the voltage signal is converted into a pulse width signal.Finally,the pulse width of each waveform corresponds to the light intensity signal.This structure is simpler than the image sensor using ADC,which makes the number of transistors greatly reduced,so that the power consumption is significantly reduced.The proposed image sensor is designed with a 0.18-um CMOS process.The total consumption is about 0.57mW from a 3.3-V power supply.The pixel array is 400×400,where the area of each pixel is 6×6?m2.The rolling shutter mode is adopted,and the dynamic range and the peak signal-to-noise ratio are 65.5dB and 52.75dB,respectively.The output mode is PWM.
Keywords/Search Tags:Brain-computer interface, Image sensor, Low power, Biomedicine
PDF Full Text Request
Related items