Real-time Processing Power Research Articles

Exploring the Application of Machine Learning in the Field of Fruit Image Recognition

Image recognition techniques play an important role in business and science fields. The application of these techniques, especially the YOLO algorithm's real-time processing capabilities and high accuracy make it particularly noteworthy in advancing fruit image recognition. This study focuses on the application of machine learning in the field of image recognition, and in particular explores the use of the YOLO (You Only Look Once) algorithm in fruit image recognition. The study experimentally explores the application of the YOLO algorithm to a complex fruit recognition task, with a focus on evaluating the algorithm's real-time processing power and accuracy. Using a wide range of fruit image datasets including COCO, OID and KITTI, the YOLO model was trained and tested with the aim of gaining insights into the algorithm's performance in different real-world scenarios and its limitations. This research points out the importance of continuous model improvement, especially in terms of image quality, model adaptation, and with other techniques such as NLP. Directions for future research include the development of more transparent and interpretable deep learning models.

Heuristic search for multi‐distance measuring equipment station selection

To support alternative positioning, navigation and timing (A-PNT), multi-DME with better geometry than distance measuring equipment (DME)/DME can be used to improve the performance. Different from the existing station selection methods that have been proposed for DME/DME, a station selection method for multi-DME should be considered in order to prevent unnecessary emission of navigation signals because DME receivers have limited number of channels and constrained real-time processing power, and also to limit DME to GNSS L5/E5a interference. A heuristic algorithm is constructed to determine the quasi-optimal selection of DME stations based on geometrical and computational cost restrictions, which is different from a brute-force method checking all possible subsets and selecting the best one. A station selection method for multi-DME is proposed based on the heuristic search. Meanwhile, the degradation of GNSS L5/E5a signal processing that is caused by DME signals is discussed. An advanced station selection method where the DME to GNSS signal interference is taken into account is also proposed. The experimental results show that the methods based on heuristic search can effectively select the quasi-optimal subset of DME stations with less computational expense.

Integrating Scalable Genome Sequencing Into Microbiology Laboratories for Routine Antimicrobial Resistance Surveillance.

Antimicrobial resistance (AMR) is considered a global threat, and novel drug discovery needs to be complemented with systematic and standardized epidemiological surveillance. Surveillance data are currently generated using phenotypic characterization. However, due to poor scalability, this approach does little for true epidemiological investigations. There is a strong case for whole-genome sequencing (WGS) to enhance the phenotypic data. To establish global AMR surveillance using WGS, we developed a laboratory implementation approach that we applied within the NIHR Global Health Research Unit (GHRU) on Genomic Surveillance of Antimicrobial Resistance. In this paper, we outline the laboratory implementation at 4 units: Colombia, India, Nigeria, and the Philippines. The journey to embedding WGS capacity was split into 4 phases: Assessment, Assembly, Optimization, and Reassessment. We show that on-boarding WGS capabilities can greatly enhance the real-time processing power within regional and national AMR surveillance initiatives, despite the high initial investment in laboratory infrastructure and maintenance. Countries looking to introduce WGS as a surveillance tool could begin by sequencing select Global Antimicrobial Resistance Surveillance System (GLASS) priority pathogens that can demonstrate the standardization and impact genome sequencing has in tackling AMR.

Smartphone-Based Point-of-Care System Using Continuous-Wave Portable Doppler

Point-of-care Ultrasound (PoCUS) is a safe, repeatable, and inexpensive bedside diagnostic tool. Over the years, PoCUS services are adopted in resource-limited settings for faster and useful outcomes. For a cost-effective and power-efficient solution, a smartphone-based portable continuous-wave Doppler ultrasound (US) system has been developed for diagnosing peripheral arterial diseases based on the hemodynamic feature values. The proposed system includes the analog front end (AFE), signal processing and display unit (SPDU), and smartphone application. The AFE acquires blood flow signal from the brachial artery using an 8-MHz pencil probe, extracts the Doppler shift frequency, and transfers to the SPDU through 12-bit analog-to-digital converter. To provide an area and power-efficient solution, SPDU is embedded in a field-programmable gate array (FPGA)-based single chip. A COordinate Rotation DIgital Computer (CORDIC)-based custom-designed 512-point fast Fourier transform is implemented in that FPGA for displaying the blood flow spectrogram in real time. For back-end processing, the smartphone application receives a spectrogram through Bluetooth, removes noise, extracts hemodynamic features, and diagnoses using a machine learning framework. The device has been examined on 18 volunteers (normal: 17 and abnormal: 1), while the accuracy is found to be 94% in the pretrained support vector machine classifier. For validation, the spectrogram of the normal and abnormal subjects and parameter values are compared with the commercial device. Overall, the handheld device is minimally trained operator-dependent and consumes < 4 W of power for real-time processing. Such smartphone-based feature extraction and automated diagnosis can facilitate the point-of-care system and provide a baseline for early assessment.

SDN-Based Multi-Tier Computing and Communication Architecture for Pervasive Healthcare

A large spectrum of healthcare applications, ranging from continuous blood sugar level monitoring to sleep apnea detection are nowadays facilitated by modern mobile gadgets. Wearable and ambient sensors generate enormous amounts of physiological data that demand high computation power for real-time processing and large storage area for recording the personal data. In order to conserve the energy on the battery-limited mainstream mobile devices of the end-users, the execution of the healthcare applications may be offloaded to a remote server. While cloud computing provides unlimited pool of resources for latency-tolerant services such as training a machine learning model, the personalization of healthcare services and the delay sensitivity of continuous health assessment necessitate a computation infrastructure in the vicinity of the end-users. As a remedy to address various demands of a wide range of pervasive healthcare applications, we propose a multi-tier computing and communication architecture composed of end-user devices, edge servers, and legacy cloud data-centers. The dynamic management of this architecture, policies to be applied within the network and the orchestration of the healthcare services are carried out by the concept of programmable networks, in the form of software-defined networking (SDN). As a concrete demonstration of our ideas, a fall risk assessment service is implemented and an experimental study is conducted to evaluate its accuracy and the performance of the multi-tier architecture. The results indicate that the proposed architecture is feasible to enable real-time healthcare services and has significant performance advantages over traditional cloud-based approaches.

Aerial MANETs: Developing a Resilient and Efficient Platform for Search and Rescue Applications

Abstract — The ability of first-responders to react to the aftermath of natural disasters depends heavily on receiving accurate, real-time data about the structures that may have been affected. Because transportation infrastructure may be unusable, aerial assessments are the gold standard by which such assessments are performed. The advent of mobile ad-hoc networks (MANETs) and autonomous aircraft represents a unique opportunity to allow for rapid response, while minimizing the cost of deployment and increasing reliability and operator safety. This paper describes the key challenges to implement fault-tolerant and efficient deployments of collaborative autonomous aircraft to increase operational reliability and performance when performing aerial sensing and assessment. Some challenges are affected by mobility, such as wireless communication, group navigation, and data collection. Security also represents a challenge during the operation of the MANET. We consider the effects of limited resources (e.g., real-time processing power, battery packs) available on the aircraft. By understanding both the application context and the resource availability, networked aircraft can reorganize to ensure resiliency for the mission if a resource failure occurs within the network.

A ROBUST PARALLEL FRAMEWORK FOR MASSIVE SPATIAL DATA PROCESSING ON HIGH PERFORMANCE CLUSTERS

Abstract. Massive spatial data requires considerable computing power for real-time processing. With the help of the development of multicore technology and computer component cost reduction in recent years, high performance clusters become the only economically viable solution for this requirement. Massive spatial data processing demands heavy I/O operations however, and should be characterized as a data-intensive application. Data-intensive application parallelization strategies are imcompatible with currently available procssing frameworks, which are basically designed for traditional compute-intensive applications. In this paper we introduce a Split-and-Merge paradigm for spatial data processing and also propose a robust parallel framework in a cluster environment to support this paradigm. The Split-and-Merge paradigm efficiently exploits data parallelism for massive data processing. The proposed framework is based on the open-source TORQUE project and hosted on a multicore-enabled Linux cluster. One common LiDAR point cloud algorithm, Delaunay triangulation, was implemented on the proposed framework to evaluate its efficiency and scalability. Experimental results demonstrate that the system provides efficient performance speedup.

An Emulation Method of Sensor Clutter Plots Elimination Using Statistical Characteristics Analysis

Aiming at problem about real-time processing power is poorer as the cause of too many clutter plots in multisource information fusion conditions when traditional information algorithm is used, method of sensor clutter elimination based on statistical characteristics analysis is put forward. Firstly, statistical characteristics of clutter plots are extracted, calculated and analyzed. Secondly, in combination with the advantages of multi-sensor detection, the plots of real-target have been filtered according to certain threshold values and safeguarded. Finally, a specific simulation experiment is given. Experiment shows this method can achieve the real-time processing power effectively.

Real-time polarization-sensitive optical coherence tomography data processing with parallel computing

With the increase of the A-line speed of optical coherence tomography (OCT) systems, real-time processing of acquired data has become a bottleneck. The shared-memory parallel computing technique is used to process OCT data in real time. The real-time processing power of a quad-core personal computer (PC) is analyzed. It is shown that the quad-core PC could provide real-time OCT data processing ability of more than 80 K A-lines per second. A real-time, fiber-based, swept source polarization-sensitive OCT system with 20 K A-line speed is demonstrated with this technique. The real-time 2D and 3D polarization-sensitive imaging of chicken muscle and pig tendon is also demonstrated.

A Fast Satellite Selection Algorithm: Beyond Four Satellites

Satellite selection is an important executive logic to prevent unnecessary navigation signals in the first place for Global Positioning System (GPS) receivers with limited number of channels and real-time processing power, such as those used in mobile phones, cars, and space crafts. In this paper, we propose a fast satellite selection algorithm to select more than four satellites based on the optimal geometries, which can obtain the smallest geometric dilution of precision (GDOP) values. The main idea of this fast algorithm is to select a subset of all satellites in view whose geometry is the most similar to the optimal geometry. Computer simulation shows that the consumed time of this algorithm is very close to that of the quasi-optimal satellite selection algorithm and obviously lower than that of the traditional optimal satellite selection algorithm to minimize GDOP factor, but the increased GDOP values relative to the minimal GDOP values are much smaller than those of the quasi-optimal satellite selection algorithm.