Study on the Low Energy Management Strategy of SoC in Elder Monitor System

We all know that the ratio of the elderly population is increasing continuously in these decades. This tendency will cause many serious social problems, such as, poor, lonely and unhealthy. The typical illnesses of the elder are including elderly deafness, coronary heart diseases and senile dementia.
Due to too busy for the young people, they have no much time to stay with their parents in normal time. So, elder monitor system becomes a hot research topic. Moreover, smart home is a good solution for those unhealthy people. We can also utilize sensors to detect the action of the elder. So, all kind of sensors are needed to be set in the smart home.
Actually, portable medical and wearable electronic are the other effective methods to help the elder people. The supervised people wear several portable sensors and a center controller. The sensor nodes are mainly designed for information acquisition, pre-processing, storage and possibly transmission using raw data and also making interventions to human body. And the center controller used as a CPU with RF receiver and transfer. Actually there is a SoC in the center controller. That means all these jobs of sensor nodes need a SoC to deal with.
You can see SoCs are embedded in the elder monitor systems, the bioelectronics appliance, and so on. Sensor information is converted, stored, processed, and controlled by SoC. SoC plays very important roles in the elder monitor system and bioelectronics.
Usually, there is a battery in the center controller. You know the development of the battery is very slow. But the function of a SoC is increasing rapidly. More function means more energy will be required. So we should make good use of power.
There is a DC-DC converter block or PMU in SoC to regulate the variable battery power to a constant voltage power to supply SoC. In a general SoC for portable medical and bioelectronics applications, there are several blocks, such as sensor, interface, DSP, PMU and so on. Usually,
different block needs different power supply.
Smaller bulk and lighter weighty is the eternal target of our research. PMU is better than discrete DC-DC converters. Smaller feature size fabricated process can be utilized to shrink the area of SoC. On the other hand, with the feature size scaling, both the power supply and the
threshold of IC are decreasing and the leakage power is increasing. It is more difficult for design the PMU or SoC. So low power low dissipation is a key point for SoC design.
For a monitor system, or a SoC, there are many different operation states. When the system or SoC is in power off, shut off, or idle state, some sub-circuits can be shut down. While, in on or wake up state, all most all the sub-circuits are working. Therefore, we can save some energy.
Secondly, DVS, the dynamic voltage scaling, is the most effective power management technique recently. In DVS, the frequency is changed based on workload, and the supply voltageis varied according to operation frequency. As a result, optimal voltage and frequency are supplied to each module of SoC. This ensures much longer battery life while keeping primary user experience.
A real-time solution is AVS, adaptive voltage scaling. AVS is a closed DVS. And can change the output voltage of PMU in time with the changing of its load.
PMU plays a very important role in SoC for saving energy.
Recently, we design a PMU circuit, which includes 4 DC-DC converters and 2 LDOs to send different voltages to different blocks in SoC, such as the CPU core, the interface, the memory, and so on.
We use a PLC, a Phase Lead Compensator, to compensate the DC-DC closed circuit. The zero is much lower than the pole of PLC, so that high frequency gain and large phase margin are obtained.
We also use 100% duty cycle at up-tracking period in order to obtain high charge speed. And 0 duty cycle is adopted to reduce the energy dissipation.
Synchronous Rectifier & Adaptive Dead-Time Control are also adopted in our PMU to obtain high efficiency. Another low energy strategy is utilizing segregated power driving method.
Our PMU is fabricated in 0.13μm 1P8M CMOS process. And its area is 2*2mm2.
We also study on MEPT, the Minimum Energy Point Tracking method. It is a novel low energy management strategy. It is very suitable for wearable electronics and bioelectronics saving energy.
Next generation, maybe we will harvest energy from solar, human body temperature, human movement, or wireless transmission to supply the wearable electronics and bioelectronics. Low energy management strategies are especially important for long time operation for them. I hope we can do something in this field with you.

-> Download presentation