Technical Talks
February 2022
Constant ΔV Charge Redistribution Method for Energy-Starved Power Management Circuits
- Speaker: Eulogio A. Antig III (UP Diliman)
- Date: 28 February 2022
Abstract
Power management integrated circuits (PMIC) help produce regulated power for wireless sensor networks (WSNs) by gathering ambient energy from the environment. For the power management to successfully deliver power to the node they need to harvest energy from various sources, store it in a capacitor, transfer these charges the to supply capacitors of the PMIC itself to power the PMIC as well as transfer more charges down the supply capacitor for the regulator to provide a regulated voltage to the node. This requires transferring charges across different capacitors and charge redistribution can be a method for the transfer. But charge redistribution from a fixed voltage has a drawback of being energy inefficient and having longer capacitor charging times, leading to increased losses due to leakage. This study aims to use a constant voltage difference charge redistribution approach wherein we control the voltage difference between two capacitors at which charge redistribution would begin. This study would investigate the relationship between variation of the constant voltage difference between two capacitors versus the charge redistribution losses. This study would also like to examine the capacitor charging times which are greatly affected by the input power and the voltage level needed to reach the constant voltage difference along with its variations. These charging times greatly influence the leakage loss which affects the overall efficiency of the proposed approach. Overall, the study aims to create a guide on finding optimum constant voltage difference for a charge redistribution approach which factored in the effects of variations of the constant voltage difference, the effects of input power and leakage that would yield the best efficiency.
December 2021
Design Techniques to Mitigate Radiation Effects on Image Sensors
- Speaker: Rico Jossel M. Maestro (KU Leuven, UP Diliman)
- Date: 15 December 2021
Abstract
Image sensors are being used more than ever in different applications: from capturing our daily lives with our smartphones, to capturing how other planets look like. Some of these applications expose the sensor in harsh environments such as the presence of radiation. As such, aside from its typical specifications, the image sensor system should also be designed with reliability and longevity in mind. This talk will provide an overview on image sensor systems and on how radiation affects its performance. Then, design techniques will be discussed to ensure that the system will still be reliable even in the presence of radiation.
An Energy-Efficient Unsupervised Online Learning Classifier for Seizure Detection
- Speaker: Adelson Chua (UC Berkeley, UP Diliman)
- Date: 9 December 2021
Abstract
Epilepsy is a serious neurological disorder affecting around 50 million people worldwide and is usually characterized by recurrent seizures. Implantable devices that record neural activity and detect seizures have been adopted to issue warnings or trigger neurostimulation to suppress epileptic seizures. Typical seizure detection systems rely on high-accuracy offline-trained machine learning classifiers that require manual retraining when seizure patterns change over long periods of time. For an implantable seizure detection system, a low-power, at-the-edge, online learning algorithm can be employed to dynamically adapt to the neural signal drifts, thereby maintaining high accuracy without external intervention. In this talk, I will be presenting an unsupervised online learning classifier using stochastic gradient descent and logistic regression. After an initial offline training phase, continuous online unsupervised classifier updates are applied in situ, which improves sensitivity in patients with drifting seizure features. The classifier was tested on two human electroencephalography (EEG) datasets, where it was able to achieve an average sensitivity of 97.5% and 97.9% respectively, and maintaining <1.2 false alarms per day. The classifier was fabricated in TSMC’s 28 nm process occupying 0.1 and achieves 1.5 nJ/classification energy efficiency, which is at least 24x more efficient than state-of-the-art.
