Hardware Efficient Analog Front-End Acquisition Circuits Based On Frequency Division Multiplexing Technique

Lung disease and cardiovascular disease are the top two leading causes of deaths worldwide according to the statistics from the World Health Organization(WHO).Electrical impedance tomography(EIT)and electrocardiography(ECG)are reported to be the only viable portable real-time monitoring method for lung and heart.For lung disease,if real-time lung monitoring is applied,many lung ventilated patients can be saved.For cardiovascular disease,ECG is the most effective method to evaluate the functional status of the heart.Real-time ECG monitoring is essential especially for preventing severe consequences caused by heart attack and arrhythmia.Multi-channel signal acquisition is needed to fulfill the precise medical diagnosis requirement.The area and power penalties in wearable application pose great challenges for biomedical acquisition circuit design.With the advancement of CMOS technology,it is possible to integrate EIT or ECG signal acquisition circuits into an integrated circuit(IC),which supports lung or heart condition monitoring.Integrating multi-channel biomedical signal acquisition circuits into a single chip enables lower cable distortion,smaller device size,lower cost,and most importantly,the possibility for portable/wearable monitoring application.However,many of the existing works have insufficient integration level due to the chip area and power penalties.Presently,most of the designs are based on Time Division Multiplexing(TDM)or Active Electrode(AE)architecture.Circuit architecture based on the TDM scheme reduces chip area,but it is limited by the settling time of channel switching.Furthermore,with a large number of inputs to multiplexer(MUX),the power penalty arises.AE system eliminates the cable transmission loss,however,a more substantial area penalty is incurred due to the redundant design such as inactive current stimulator and communication block.How to reduce the area and power of the chip,while increase the multi-channel integration level and monitoring period is the critical issue.In this dissertation,we have proposed a new design architecture based on Frequency Division Multiplexing(FDM)technique.The FDM scheme enjoys the merits of lower power consumption per channel,smaller area per channel,and on-chip data combination,while the main limitation is the bandwidth resource.Due to the narrow bandwidth nature of biomedical signals,multi-channel narrow-band signals can be combined to form a wide-band FDM signal.Therefore,only one analog to digital converter(ADC)or voltage sensing channel is required to quantize or acquire the wide-band FDM signal.In this dissertation,we have developed two different biomedical signal acquisition ICs to verify the feasibility and advantages of FDM technique.The first design is for lung tomographic imaging based on a 13-channel EIT IC and the second design is for 12-Lead ECG recording based on an 8-channel ECG IC.With the proposed FDM based architecture,the EIT IC achieves the lowest power consumption per channel and the smallest area occupation per channel compared to the current state-of-the-art,while the performances of the ECG IC are comparable with the current state-of-the-art.The 13-channel 1.53-mW 11.28-mm~2 EIT IC based on FDM technique is proposed for lung tomographic imaging.By using early in-phase/quadrature-phase(I/Q)demodulation for13 input signals,DC coupled analog front-end can be used with ultra-low bandwidth thus achieving lower power consumption and 26 signal components are produced.Only two 16-bit Delta-Sigma Modulators(DSMs)are used to sample the two assembled FDM signals,with each containing 13 signal components.As DSM enjoys high over-sampling ratio,this eliminates the need of high-order analog anti-aliasing filter.The proposed EIT IC reduces the power per channel by 10 times to 118μW and reduces the area per channel by 58%to 0.87mm~2 compared to other reported state-of-the-art.Spectrum estimation based on window corrected fast Fourier transform is proposed to batch extract all the 26 signal components.As a result,5 frames per second lung EIT images can be generated by the IC.The IC is reported at ISSCC 2019,the world’s highest-level conference in the field of integrated circuits recognized by academia and industry circles.The 8-channel ECG IC based on FDM technique is presented for 12-Lead ECG recording application.Different from EIT signals,the ECG input signals across channels do not vary much.Therefore,the FDM technique can be employed at the instrumental amplifier stage,and only one programmable gain amplifier,one buffer,and one ADC are used in the8-channel acquisition system.These functional blocks are shared by 8-channel ECG signals to achieve smaller area with the FDM scheme.The whole 8-channel ECG recording system is implemented in 130 nm CMOS process with a total area of only 4.2 mm~2,and 12-Lead ECG measurement for medical diagnosis is realized.At last,we summarize the research work related to the biomedical signal acquisition circuits design based on FDM technique.The directions for further development and new prospects of biomedical signal acquisition circuits are discussed.