Publications

Alternating between Surrogate Model Construction and Search for Configurations of an Autonomous Delivery System

Published in 2024 IEEE International Conference on Software Analysis, Evolution and Reengineering, Rovaniemi, Finland, 2024, pp. 383-394 (SANER 2024, Industry Track) , 2024

Autonomous robots are emerging as a solution to various challenges of last mile goods delivery, like reducing traffic congestion, pollution, and costs. The configuration of an autonomous delivery robots system requires balancing aspects like delivery rate, cost of robots’ operation, and required monitoring efforts. Our industry partner Panasonic is employing a search-based approach to find the configurations of the system that optimise these three aspects for a given set of customers’ orders. The approach uses a simulator to assess the different configurations in the fitness functions’ computation. Due to the high cost of the simulation, the whole search-based approach is computationally expensive. A classic approach to speed up such approaches is to use surrogate models trained on example simulation data that allow to approximate the results of a simulated configuration with negligible computational cost. A risk when using such approaches is to underestimate the cost of building the surrogate model itself, that can exceed the computational gain obtained during the search, thus making the adoption of surrogate models detrimental. In this work, we propose an approach in which the surrogate model is not trained before the search; instead, the approach alternates between training the model on subsets of data of increasing size, and searching using these cheaper models until the search stagnates. Experiments over 144,000 settings of the search show that the proposed approach can significantly reduce the cost of searching for configurations, while having an acceptable impact on the quality of the configurations it finds.

Recommended citation: Chin-Hsuan Sun, Thomas Laurent, Paolo Arcaini, Fuyuki Ishikawa. (2024). Alternating Between Surrogate Model Construction and Search for Configurations of an Autonomous Delivery System. 2024 IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER), Rovaniemi, Finland, 2024, pp. 383-394. https://doi.org/10.1109/SANER60148.2024.00045

Quantifying changes in oxygen saturation of the internal jugular vein in vivo using deep neural networks and subject-specific three-dimensional Monte Carlo models

Published in Optics Letters Vol. 49, Issue 10, 2024

Central venous oxygen saturation (ScvO2) is an important parameter for assessing global oxygen usage and guiding clinical interventions. However, measuring ScvO2 requires invasive catheterization. As an alternative, we aim to noninvasively and continuously measure changes in oxygen saturation of the internal jugular vein (SijvO2) by a multi-channel near-infrared spectroscopy system. The relation between the measured reflectance and changes in SijvO2 is modeled by Monte Carlo simulations and used to build a prediction model using deep neural networks (DNNs). The prediction model is tested with simulated data to show robustness to individual variations in tissue optical properties. The proposed technique is promising to provide a noninvasive tool for monitoring the stability of brain oxygenation in broad patient populations.

Recommended citation: Chin-Hsuan Sun, Hao-Wei Lee, Ya-Hua Tsai, Jia-Rong Luo & Kung-Bin Sung (2024). Quantifying changes in oxygen saturation of the internal jugular vein in vivo using deep neural networks and subject-specific three-dimensional Monte Carlo models. Optics Letters. 49(10). https://doi.org/10.1364/OL.517960

Non-invasive quantification of changes in blood oxygen saturation of the internal jugular vein: theoretical evaluation and in-vivo demonstration

Published in Society of Photographic Instrumentation Engineers Photonic West Conference at San Francisco (SPIE Photonic West), 2023

A prediction model based on artificial neural networks was built to quantify changes in blood oxygen saturation of the internal jugular vein (dSijvO2) from diffuse reflectance measured at five wavelengths. The model was trained by Monte Carlo simulations with various tissue optical coefficients and subject-specific tissue structure determined by ultrasound imaging. Errors in dSijvO2 estimated from simulated data are below 2.2% and independent of the initial oxygen saturation. The model was further validated by excellent agreements between modeled and measured in-vivo reflectance spectra from a healthy volunteer undergoing hyperventilation, and the quantified trend of dSijvO2 followed expectations during and after hyperventilation. The proposed method is promising to provide non-invasive quantification of dSijvO2.

Recommended citation: Hsin-Yuan Hsieh, Chin-Hsuan Sun, Yi-Siang Syu, Yin-Fu Chen, Hao-Wei Lee, Kuang Yang, Kung-Bin Sung. (2023). Non-invasive quantification of changes in blood oxygen saturation of the internal jugular vein: theoretical evaluation and in-vivo demonstration. SPIE BiOS, 2023. https://doi.org/10.1117/12.2651106

Perturbation Monte Carlo Applicability for Human Neck Model

Published in The 5th Global Conference on Biomedical Engineering & Annual Meeting of Taiwanese Society of Biomedical Engineering (TSBME) & SPIE Biomedical Imaging and Sensing Conference (BISC), 2022