SmartComp2024

Tutorials

We have 4 tutorials this year; details are shown as follows.

Tutorial 1: 803 (8F), June 29th, 14:45-16:45

The Internet of Bio-Nano Things -Smart Computing in the Human Body

Stefan Fischer, the University of Lübeck, Germany

Abstract
The Internet of Bio-Nano Things (IoBNT) is an innovative field of research located at the intersection of nanotechnology, biotechnology and information and communication technologies. It aims to enable the seamless integration of biological and nanoscale systems into the Internet in order to develop advanced biomedical applications, environmental monitoring sensors and energy-efficient networks. At the core of IoBNT are biocompatible nanodevices that can function in living organisms to monitor or modify specific biological processes in real time. These devices communicate with each other and with the Internet to collect, process and transmit data, opening up entirely new possibilities for health monitoring, disease control, environmental protection and many other areas. By merging biology and nanotechnology, IoBNT promises to push the boundaries of what is technically possible while improving the efficiency and sustainability of technological solutions.

This tutorial first introduces the basics of IoBNT. It explains the basic requirements and existing technologies and looks at a number of example applications, primarily medical applications. In the main part, DNA-based nanonetworks are presented as a promising implementation technology. Before providing an outlook, further concepts are addressed, which focus in particular on gateways between communication inside and outside the body.

Contactless Physiological Health Sensing: Challenges, Solutions & Opportunities

Nirmalya Roy, University of Maryland Baltimore County, USA
Zahid Hasan, University of Maryland Baltimore County, USA
 (*) We are very sorry, but due to unavoidable circumstances, Zahid will not be able to come to Japan. For this reason,  Nirmaly takes over and handles all parts.

Abstract
Contactless health monitoring techniques, such as Remote Photoplethysmography (rPPG) and video-based Respiratory Rate (RR) estimation have emerged as promising methods utilizing regular camera sensors for capturing vital signs like Heart Rate (HR) and Respiratory Rate (RR). These approaches offer cost-effective, widely applicable, and safe solutions for regular and long-term health monitoring. However, the primary challenge lies in extracting accurate vital signals from the captured videos due to their low signal-to-noise ratios.

In this tutorial, we aim to provide a comprehensive background on rPPG and video-based heart rate estimation, covering signal acquisition and physics-inspired estimation principles. We will discuss signal-processing approaches and their limitations. We will delve into DL-based estimation systems, addressing the challenge of inherent aleatoric uncertainty (irreducible uncertainty from various stochastic factors like sensor variations and inter-subject differences) in the ground truth data annotation that hinders the development of generalized deep learning-based rPPG estimation system. To reinforce a robust DL model addressing these inherent uncertainties in rPPG data streams, we will discuss three novel deep-learning approaches – a multi-task learning method, a self-supervised learning method, and a generative adversarial network-based architecture for rPPG estimation. Furthermore, we will demonstrate a real-time heart rate estimation system, RhythmEdge, using a low-cost camera sensor and describe the rPPG-specific pruning techniques to reduce rPPG model size for efficient edge implementation. We will conclude the tutorial with existing research gaps and potential directions in contactless rPPG and respiratory rate estimation research.

---

Tutorial 3: 1101&1102 (11F), July 1st, 13:15-15:15

Advancing Smart Computing: A Comprehensive Tutorial to 3D Point Clouds

Tatsuya Amano, Osaka University, JAPAN
Hamada Rizk, Osaka University, JAPAN and Tanta University, Egypt

Abstract
3D point clouds from LiDAR sensors have emerged as a powerful representation for understanding and interacting with the surrounding environment in smart computing systems. This tutorial aims to provide an overview of fundamental techniques, considerations, and applications of 3D point cloud recognition, with a focus on both deep learning and non-deep learning approaches. We will begin by introducing the unique characteristics and challenges of point cloud processing, highlighting the differences from traditional image-based approaches. Through the example of PointNet, a pioneering deep learning model for point cloud recognition, we will discuss the specific considerations and best practices for handling point cloud data. Recognizing the limitations of deep learning in resource-constrained and mobile environments, we will also explore alternative statistical and probabilistic techniques, such as Fisher Vector-based approaches, which enable efficient and lightweight point cloud processing and context recognition. These techniques are particularly relevant for smart computing applications that require real-time processing on edge devices and in Internet of Things scenarios.

The tutorial will also showcase various applications of point cloud recognition in smart computing contexts. We will showcase a case study of Osaka University smart campus, where we have deployed a large-scale pedestrian tracking system using a network of approximately 70 LiDAR sensors. Additionally, we will discuss the use of point clouds for remote space sharing in extended reality (XR) applications, demonstrating how point cloud recognition enables immersive and interactive experiences.We will also present a case study of our LiDAR mobile device (hitonavi-μ).

---

Tutorial 4: Conf. Hall (12F), July 1st, 13:15-15:15

Science of Cyber Physical Security in Smart Living Applications

Sajal K. Das, Missouri University of Science and Technology, USA
Shameek Bhattacharjee, Western Michigan University, USA

Abstract
The vision behind community-scale smart living applications is to use sensor-actuator devices, the so-called Internet of Things (IoT), to generate sensing data that provide situational awareness of the   physical world to improve the quality of human life at the city scale. Examples applications include smart transportation, customer and distribution layers of the smart grid metering, smart water networks, etc. The effects of threats such cyber-attacks, device/network faults and malfunctions, unsafe events, typically manifest themselves as anomalies that need to be promptly detected. However, there are unique challenges in anomaly detection for smart living : (1) behavioral randomness of humans creates dynamic spatiotemporal variations in data patterns making it difficult to learn the profile of benign behavior leading to unusable false alarm frequencies; (2) High non-linearity, non-IID data, random evolving patterns, cause traditional anomaly detection/learning methods to lose detection sensitivity; (3) smart living sensing data often have privacy and individual device profiling concerns; (4) unlabeled threats present while learning benign profile.
To address such challenges, we present a unified framework to detect various threats in a light-weight, timely, and privacy preserving manner. The tutorial covers the following topics: (1) decentralized graph representations of the wide area CPS; (2) anomaly detection metric learning; (3) resilient machine learning for finding detection thresholds; (4) bio-inspired classification methods to identify individual devices affected by the threats.  We show key results achieved by the overarching framework in smart grid, smart transportation, smart water networks, etc.