CNN-Based QR Code Reading of Package for Unmanned Aerial Vehicle.

This paper plans to establish a warehouse management system based on an unmanned aerial vehicle (UAV) to scan the QR codes printed on packages. This UAV consists of a positive cross quadcopter drone and a variety of sensors and components, such as flight controllers, single-board computers, optical flow sensors, ultrasonic sensors and cameras, etc. The UAV stabilizes itself by proportional-integral-derivative (PID) control and takes pictures of the package as it reaches ahead of the shelf. Through convolutional neural networks (CNNs), the placement angle of the package can be accurately identified. Some optimization functions are applied to compare system performance. When the angle is 90°, that is, the package is placed normally and correctly, the QR code will be read directly. Otherwise, image processing techniques that include Sobel edge computing, minimum circumscribed rectangle, perspective transformation, and image enhancement is required to assist in reading the QR code. The experimental results showed that the proposed algorithm provided good performance of a recognition rate of 94% for the stochastic gradient descent (SGD) and 95% for Adadelta optimization functions. After that, successful QR code reading was presented.

An Artificial Intelligence of Things-Based Picking Algorithm for Online Shop in the Society 5.0's Context.

In this study, an Artificial Intelligence of Things (AIoT)-based automated picking system was proposed for the development of an online shop and the services for automated shipping systems. Speed and convenience are two key points in Industry 4.0 and Society 5.0. In the context of online shopping, speed and convenience can be provided by integrating e-commerce platforms with AIoT systems and robots that are following consumers' needs. Therefore, this proposed system diverts consumers who are moved by AIoT, while robotic manipulators replace human tasks to pick. To prove this idea, we implemented a modified YOLO (You Only Look Once) algorithm as a detection and localization tool for items purchased by consumers. At the same time, the modified YOLOv2 with data-driven mode was used for the process of taking goods from unstructured shop shelves. Our system performance is proven by experiments to meet the expectations in evaluating efficiency, speed, and convenience of the system in Society 5.0's context.

Raspberry Pi-Based Sleep Posture Recognition System Using AIoT Technique.

The relationship between sleep posture and sleep quality has been studied comprehensively. Over 70% of chronic diseases are highly correlated with sleep problems. However, sleep posture monitoring requires professional devices and trained nursing staff in a medical center. This paper proposes a contactless sleep-monitoring Internet of Things (IoT) system that is equipped with a Raspberry Pi 4 Model B; radio-frequency identification (RFID) tags are embedded in bed sheets as part of a low-cost and low-power microsystem. Random forest classification (RFC) is used to recognize sleep postures, which are then uploaded to the server database via Wi-Fi and displayed on a terminal. The experimental results obtained using RFC were compared to those obtained via the support vector machine (SVM) method and the multilayer perceptron (MLP) algorithm to validate the performance of the proposed system. The developed system can be also applied for sleep self-management at home and wireless sleep monitoring in medical wards.

Full-order observer for direct torque control of induction motor based on constant V/F control technique.

This paper presents another control method for the three-phase induction motor that is direct torque control based on constant voltage per frequency control technique. This method uses the magnitude of stator flux and torque errors to generate the stator voltage and phase angle references for controlling the induction motor by using constant voltage per frequency control method. Instead of hysteresis comparators and optimum switching table, the PI controllers and space vector modulation technique are used to reduce torque and stator-flux ripples and achieve constant switching frequency. Moreover, the coordinate transformations are not required. To implement this control method, a full-order observer is used to estimate stator flux and overcome the problems from drift and saturation in using pure integrator. The feedback gains are designed by simple manner to improve the convergence of stator flux estimation, especially in low speed range. Furthermore, the necessary conditions to maintain the stability for feedback gain design are introduced. The simulation and experimental results show accurate and stable operation of the introduced estimator and good dynamic response of the proposed control method.