Position Recognition System Research Articles

DeepLPos: A Comprehensive Hybrid Deep Learning Model for Lying Position Recognition Using a Tactile Sensor Array System

Abstract Unpredictable limb movements or turning motions can significantly disrupt the extraction of respiratory and heart rate signals. Lying position detection plays a vital role in clinical settings, enabling continuous monitoring of patients' sleep conditions and respiratory rates. In this paper, a smart sleeping position recognition system is proposed, which employs a tactile pressure sensor array based on the unique structure of "the double-layer capacitors (EDLC)". The sensor array, comprising 64 rows and 32 columns (2048 nodes), captures four types of healthy lying positions using an 8-bit AD module. Despite challenges arising from limited experimental samples for accurate training, we propose DeepLPos, a hybrid deep learning approach combining Generative Adversarial Networks (GAN) and the YOLO network. To tackle the differentiation challenge between supine and prone positions, we introduce an SPD Conv attention module to enhance the resolution of detailed descriptions in pressure images. The designed model is then pruned to simplify its structure and parameters, ensuring fast and accurate lying position detection. Furthermore, we evaluate the performance of our proposed method on the SLP dataset and demonstrate an impressive accuracy rate of 97.5%. Real-time processing speed reaches 0.069s/frame, demonstrating the efficiency and effectiveness of the developed approach in practical applications.

Real-time face mask position recognition system based on MobileNet model.

COVID-19 is a highly contagious disease that was first identified in 2019, and has since taken more than six million lives world wide till date, while also causing considerable economic, social, cultural and political turmoil. As a way to limit its spread, the World Health Organization and medical experts have advised properly wearing face masks, social distancing and hand sanitization, besides vaccination. However, people wear masks sometimes uncovering their mouths and/or noses consciously or unconsciously, thereby lessening the effectiveness of the protection they provide. A system capable of automatic recognition of face mask position could alert and ensure that an individual is wearing a mask properly before entering a crowded public area and putting themselves and others at risk. We first develop and publicly release a dataset of face mask images, which are collected from 391 individuals of different age groups and gender. Then, we study six different architectures of pre-trained deep learning models, and finally propose a model developed by fine tuning the pre-trained state of the art MobileNet model. We evaluate the performance (accuracy, F1-score, and Cohen's Kappa) of this model on the proposed dataset and MaskedFace-Net, a publicly available synthetic dataset created by image editing. Its performance is also compared to other existing methods. The proposed MobileNet is found as the best model providing an accuracy, F1-score, and Cohen's Kappa of 99.23%, 99.22%, and 99.19%, respectively for face mask position recognition. It outperforms the accuracy of the best existing model by about 2%. Finally, an automatic face mask position recognition system has been developed, which can recognize if an individual is wearing a mask correctly or incorrectly. The proposed model performs very well with no drop in recognition accuracy from real images captured by a camera.

Real-time face mask position recognition system using YOLO models for preventing COVID-19 disease spread in public places

Real-time face mask position recognition system using YOLO models for preventing COVID-19 disease spread in public places

Real-time face mask position recognition system using YOLO models for preventing COVID-19 disease spread in public places

Real-time face mask position recognition system using YOLO models for preventing COVID-19 disease spread in public places

Research on Position Recognition System of Gantry Hoisting Based on Machine Vision

During the process of steel casting, the task of worker is to identify the relative position of gantry hook and lifting lug of hot metal ladle so as to judge the appropriate time when gantry crane is to be lifted. Worker needs to stand near hot metal ladle, which does great damage to human’s body if exposed to such a bad environment. For this reason, the author has designed a machine vision system to displace artificial recognition. The machine vision system consists of the vision recognition algorithm based on AlexNet Convolution Neural Network (CNN) and the improved Canny edge detection algorithm. AlexNet CNN completes the preliminary recognition of the image and provides the samples that meet the recognition conditions to the improved Canny algorithm. The improved Canny algorithm completes the extraction of parameters such as the relative distance between the gantry hook and lifting lug and the tilt Angle of gantry hook. The sample images are extracted from two environments: model simulation and field simulation. Experimental data shows that the recognition algorithm based on AlexNet CNN can achieve 100% image recognition rate, which is higher than other traditional machine learning algorithms; the improved Canny edge detection algorithm can improve edge quality of extracting images in harsh environments.

A Situation-Aware Indoor Localization (SAIL) System Using a LF and RF Hybrid Approach.

Recently, studies focusing on identifying user’s current location for use in a wide variety of differentiated location-based and localization services have steadily increased. In particular, location awareness using wireless communication is gaining attention in indoor environments composed of many obstacles, where GPS signals cannot reach. Previously, studies have focused mostly on location precision, which resulted in using many beacon nodes, not considering the initial installation and maintenance costs, communication robustness, or power consumption. This makes it difficult to apply existing methods to various fields, especially in mobile nodes (i.e., wearable devices, mobile tags, etc.) with limited battery capacity. In this paper, we propose a hybrid situation-aware indoor localization (SAIL) system for real-time indoor localization using a combination of low frequency (LF) and Bluetooth Low Energy (BLE) 4.0. This approach allows us to work with limited battery capacity mobile devices, and identify tagged mobile nodes and their current location in relevance to the anchor node. In our experiment, we attached one anchor node at the entrance to indoor areas such as office or factory settings. Using our hybrid SAIL system, we were able to detect the passing of a mobile node through the entrance and recognize whether the node is entering or exiting the room by calculating the direction of movement as well as the distance from the entrance. This allowed us to distinguish the precise position in an indoor environment with the margin of error being 0.5 m. The signal attenuation due to obstacles is overcome by using LF communication in the 125-kHz band. This approach enables us to reduce the number of initially installed anchor nodes as well as the power consumption of the mobile node. We propose an indoor position recognition system, namely, the hybrid SAIL system, that can be applied to mobile nodes with limited battery capacity by reducing the system complexity and power consumption.

Underwater 3-D Spatial Attenuation Characteristics of Electromagnetic Waves With Omnidirectional Antenna

In this paper, an underwater 3-D spatial attenuation model of electromagnetic waves is proposed to establish a position recognition system in underwater 3-D space using distance estimation. The distance estimation is based on signal attenuation characteristics of electromagnetic waves radiated through several omnidirectional antennas. Positions in underwater environments have been predominantly estimated by an ultrasonic position recognition system using travel time and phase difference. However, such a system provides inaccurate estimation on account of the multipath effect in a structured environment or complex environment with many obstacles. A position estimation method that uses signal attenuation of electromagnetic waves has been proposed to overcome this limitation. That method can precisely estimate position on a structured 2-D plane. In this paper, an existing underwater position estimation system based on electromagnetic waves is expanded into 3-D space using the Friis formula and plane wave equation, thereby classifying attenuation characteristics of electromagnetic waves into the effects of medium, radiation, and antenna. These effects are summarized to deduce the signal attenuation characteristics of electromagnetic waves in 3-D space. In addition, the relative position ( $R$ , ${ \Theta}$ , ${ \Phi}$ ) and attitude ( ${ \phi}$ , ${ \theta}$ , $\psi$ ) of transceiver antennas are used to define a coordinate system for 3-D estimation; moreover, an attenuation model is defined for individual factors in the coordinate system. A generalized attenuation model of an omnidirectional antenna in 3-D space is presented, and the validity of the proposed model is demonstrated through experiments. Based on the results, the proposed model demonstrates the potential application to an omnidirectional antenna and expansion into an attenuation model based on the 3-D position between random antennas.

굴삭기의 버킷 끝단 위치인식에 관한 연구

The accurate calculation of bucket tip position has a large influence on showing the motion of an excavator on the display device of the excavator and controlling the excavator automatically. It is generally known that Inertial Measurement Unit (IMU) sensors are more accurate than accelerometer-based sensors while the boom, arm or bucket moves because additional forces beyond gravity add additional acceleration to the sensors. To prove the accuracy difference between the two types of sensors, a position recognition system using an accelerometer-based sensor and an IMU sensor is implemented on the excavator. The experimental results show that the system using the IMU sensor significantly reduces the position recognition error while bucket moves and additional force beyond gravity exists.

Position Recognition System for Autonomous Vehicle Using the Symmetric Magnetic Field

The autonomous driving method using magnetic sensors recognizes the position by measuring magnetic fields in autonomous robots or vehicles after installing magnetic markers in a moving path. The Position estimate method using magnetic sensors has an advantage of being affected less by variation of driving environment such as oil, water and dust due to the use of magnetic field. It also has the advantages that we can use the magnet as an indicator and there is no consideration for power and communication environment. In this paper, we propose an efficient sensor system for an autonomous driving vehicle supplemented for existing disadvantage. In order to efficiently eliminate geomagnetism, we analyze the components of the horizontal and vertical magnetic field. We propose an algorithm for position estimation and geomagnetic elimination to ease analysis, and also propose an initialization method for sensor applied in the vehicle. We measured and analyzed the developed system in various environments, and we verify the advantages of proposed methods.

指点字の打点指と打点位置の認識システムに関する研究

指点字の打点指と打点位置の認識システムに関する研究

실감 게임 인터페이스를 위한 지능형 위치 보정 기법 개발

Recently, a wide range of next-generation's game consoles has been developed by many game makers. Particularly, active interaction between users and games is required more than ever before for giving gamers the fullest pleasure. In this work, an infrared image sensor based position recognition system which can be used for virtual game interface is proposed. Furthermore, two kinds of compensation algorithms and Extended Kalman Filter are utilized to enhance the performance of the proposed system. The proposed system can effectively generate the position of the gamer in the face of the coordinate distortion and noise. To verity the feasibilities of the proposed system, various experiments are carried out.

Development of a Construction Robot for Marking on Ceiling Boards. 3rd Report. Prototype of the Laser Pointer System.

We are developing positioning systems for mobile robots working at construction sites. In earlier work, we investigated the use of pillars as reference points at construction sites for a position recognition system, and reported on the performance of a prototype system. To further increase the flexibility of the robot system, we have developed a new self-position measurement system for the mobile marking robot. This new system is aimed at expansion of the work area of the marking robot. The system is composed of a prototype of a laserpointer robot and a marking robot with laser-spot detectors. The marking robot uses a point of laser light projected onto it by the laser-pointer robot, placed at a reference point, to determine its position. This report outlines the method for the self-position measurements and describes the performance of the system in a field test.

Packaging machine for “Mozartkugeln” controlled by a position recognition system

This paper explores some of the critical issues in the development of a picture position recognition system. The system uses an on-line correlation procedure. Its output is used to control a positioning and packaging machine, the design of which is described in some detail.

A new imaging method for intraoperative therapy control in skull-base surgery.

CAS-computer assisted surgery-is a new imaging method supporting skull base surgeons. Support is granted not only for preoperative planning of therapy but also for pathfinding during surgery itself and in the postoperative phase as therapy control. The CAS-system consists of high technology items such as -a digital image generation system (CT, MR) -a real-time image processing system -a 3 D position recognition system. Robotics are not incorporated in this system but a hand-guided manipulator houses the surgeon's instrument. Accuracy of the method has been experimentally determined to be within 1 mm. Follow up systems are under development to permit microsurgery support as well.