Anton Kos

New benchmarking methodology and programming model for big data processing

Big data processing is becoming a reality in numerous real-world applications. With the emergence of new data intensive technologies and increasing amounts of data, new computing concepts are needed. The integration of big data producing technologies, such as wireless sensor networks, Internet of Things, and cloud computing, into cyber-physical systems is reducing the available time to find the appropriate solutions. This paper presents one possible solution for the coming exascale big data processing: a data flow computing concept. The performance of data flow systems that are processing big data should not be measured with the measures defined for the prevailing control flow systems. A new benchmarking methodology is proposed, which integrates the performance issues of speed, area, and power needed to execute the task. The computer ranking would look different if the new benchmarking methodologies were used; data flow systems would outperform control flow systems. This statement is backed by the recent results gained from implementations of specialized algorithms and applications in data flow systems. They show considerable factors of speedup, space savings, and power reductions regarding the implementations of the same in control flow computers. In our view, the next step of data flow computing development should be a move from specialized to more general algorithms and applications.

Suitability of Smartphone Inertial Sensors for Real-Time Biofeedback Applications

This article studies the suitability of smartphones with built-in inertial sensors for biofeedback applications. Biofeedback systems use various sensors to measure body functions and parameters. These sensor data are analyzed, and the results are communicated back to the user, who then tries to act on the feedback signals. Smartphone inertial sensors can be used to capture body movements in biomechanical biofeedback systems. These sensors exhibit various inaccuracies that induce significant angular and positional errors. We studied deterministic and random errors of smartphone accelerometers and gyroscopes, primarily focusing on their biases. Based on extensive measurements, we determined accelerometer and gyroscope noise models and bias variation ranges. Then, we compiled a table of predicted positional and angular errors under various biofeedback system operation conditions. We suggest several bias compensation options that are suitable for various examples of use in real-time biofeedback applications. Measurements within the developed experimental biofeedback application show that under certain conditions, even uncompensated sensors can be used for real-time biofeedback. For general use, especially for more demanding biofeedback applications, sensor biases should be compensated. We are convinced that real-time biofeedback systems based on smartphone inertial sensors are applicable to many similar examples in sports, healthcare, and other areas.

Evaluation of Smartphone Inertial Sensor Performance for Cross-Platform Mobile Applications.

Smartphone sensors are being increasingly used in mobile applications. The performance of sensors varies considerably among different smartphone models and the development of a cross-platform mobile application might be a very complex and demanding task. A publicly accessible resource containing real-life-situation smartphone sensor parameters could be of great help for cross-platform developers. To address this issue we have designed and implemented a pilot participatory sensing application for measuring, gathering, and analyzing smartphone sensor parameters. We start with smartphone accelerometer and gyroscope bias and noise parameters. The application database presently includes sensor parameters of more than 60 different smartphone models of different platforms. It is a modest, but important start, offering information on several statistical parameters of the measured smartphone sensors and insights into their performance. The next step, a large-scale cloud-based version of the application, is already planned. The large database of smartphone sensor parameters may prove particularly useful for cross-platform developers. It may also be interesting for individual participants who would be able to check-up and compare their smartphone sensors against a large number of similar or identical models.

Wearable training system with real-time biofeedback and gesture user interface

Wearable computing and sensors are becoming increasingly prevalent in our daily lives. This paper presents a wearable training system designed to facilitate the learning process of proper movement patterns in sports training. The system implements a gesture user interface and real-time biofeedback. The feedback loop consists of one or more body-attached motion sensors, a processing device and a biofeedback device that are interconnected through low-latency communication channels. Due to the diverse number of possible applications, a flexible system architecture, which includes several different system versions, is proposed. Operation of the system is driven by user gestures. To demonstrate the concept of the proposed real-time biofeedback training system, an application for golf swing training is developed. The application implements the system using smartphone motion sensors and audio biofeedback and aids golfers in correcting unwanted head movements during a golf swing. The application is driven by a gesture user interface. During the golf swing, the application provides users with real-time audio feedback that signals head movement errors. The field test results show that the developed application can be used as an efficient tool in golf swing training.

Sorting Networks on Maxeler Dataflow Supercomputing Systems

Abstract The primary contribution of this study is the implementation and evaluation of network sorting algorithms on a Maxeler dataflow computer. Sorting is extensively used in numerous applications. We discuss sequential, parallel, and network sorting algorithms. The major part of this study is dedicated to the properties, construction, and testing of sorting networks. We introduce and compare principal network sorting algorithms with predominant sequential and parallel sorting algorithms. We implement network sorting algorithms in an entry model of the Maxeler dataflow supercomputing system. The goal of our study is to compare the sorting times of network sorting algorithms using a Maxeler dataflow computer with the sorting times of optimal sequential and parallel sorting algorithms using a control flow computer. In different testing scenarios, we demonstrate that high sorting speedups can be achieved with network sorting using a Maxeler dataflow computer. We sorted arrays of 128 values. Using different testing parameters, we achieved speedups that ranged from approximately 10 to more than 200. Sorting networks that execute parallel sorting using the dataflow computational paradigm offer a possible solution for expanding volumes of data. By converting to more advanced Maxeler systems and researching new ideas and solutions, we aim to sort large arrays and achieve large speedups.

Clustering by fast search and merge of local density peaks for gene expression microarray data

Clustering is an unsupervised approach to classify elements based on their similarity, and it is used to find the intrinsic patterns of data. There are enormous applications of clustering in bioinformatics, pattern recognition, and astronomy. This paper presents a clustering approach based on the idea that density wise single or multiple connected regions make a cluster, in which density maxima point represents the center of the corresponding density region. More precisely, our approach firstly finds the local density regions and subsequently merges the density connected regions to form the meaningful clusters. This idea empowers the clustering procedure, in which outliers are automatically detected, higher dense regions are intuitively determined and merged to form clusters of arbitrary shape, and clusters are identified regardless the dimensionality of space in which they are embedded. Extensive experiments are performed on several complex data sets to analyze and compare our approach with the state-of-the-art clustering methods. In addition, we benchmarked the algorithm on gene expression microarray data sets for cancer subtyping; to distinguish normal tissues from tumor; and to classify multiple tissue data sets.

Paradigm Shift in Big Data SuperComputing: DataFlow vs. ControlFlow

The paper discusses the shift in the computing paradigm and the programming model for Big Data problems and applications. We compare DataFlow and ControlFlow programming models through their quantity and quality aspects. Big Data problems and applications that are suitable for implementation on DataFlow computers should not be measured using the same measures as ControlFlow computers. We propose a new methodology for benchmarking, which takes into account not only the execution time, but also the power and space, needed to complete the task. Recent research shows that if the TOP500 ranking was based on the new performance measures, DataFlow machines would outperform ControlFlow machines. To support the above claims, we present eight recent implementations of various algorithms using the DataFlow paradigm, which show considerable speed-ups, power reductions and space savings over their implementation using the ControlFlow paradigm.

Validation of smartphone gyroscopes for mobile biofeedback applications

Smartphones are currently the most pervasive wearable devices. One particular use of smartphone inertial sensors is motion tracking in various mobile systems and applications. The objective of this study is to validate smartphone gyroscopes for angular tracking in mobile biofeedback applications. The validation method includes measurements of angular motion performed concurrently by a smartphone gyroscope and a professional optical tracking system serving as the reference. The comparison of the measurement results shows that the inaccuracies of a calibrated smartphone gyroscope for various movements are between 0.42° and 1.15°. Based on the measurement results and the general requirements of biofeedback applications, smartphone gyroscopes are sufficiently accurate for angular motion tracking in mobile biofeedback applications.

The Role of High Performance Computing and Communication for Real-Time Biofeedback in Sport

This paper studies the main technological challenges of real-time biofeedback in sport. We identified communication and processing as two main possible obstacles for high performance real-time biofeedback systems. We give special attention to the role of high performance computing with some details on possible usage of DataFlow computing paradigm. Motion tracking systems, in connection with the biomechanical biofeedback, help in accelerating motor learning. Requirements about various parameters important in real-time biofeedback applications are discussed. Inertial sensor tracking system accuracy is tested in comparison with a high performance optical tracking system. Special focus is given on feedback loop delays. Real-time sensor signal acquisitions and real-time processing challenges, in connection with biomechanical biofeedback, are presented. Despite the fact that local processing requires less energy consumption than remote processing, many other limitations, most often the insufficient local processing power, can lead to distributed system as the only possible option. A multiuser signal processing in football match is recognised as an example for high performance application that needs high-speed communication and high performance remote computing. DataFlow computing is found as a good choice for real-time biofeedback systems with large data streams.

Multi-sensor Golf Swing Classification Using Deep CNN

Abstract In recent years smart sport equipments have achieved great success in professional and amateur sports, as well as body sensory systems; now discovering interesting knowledge in the surge of data from those embedded sensors used in sports is necessary and the focus of our research. In this paper, we investigate golf swing data classification method based on deep convolutional neural network (deep CNN) fed with multi-sensor golf swing signals. Our smart golf club integrates two orthogonally affixed strain gage sensors, 3-axis accelerometer and 3-axis gyroscope, and collects real-world golf swing data from professional and amateur golf players. Furthermore we explore the performance of our well-trained CNN-based classifier and evaluate it on the real-world test set in terms of common indicators including accuracy, precision-recall, and F1-score. Experiments and corresponding results show that our CNN-based model can satisfy the requirement of accuracy of golf swing classification, and outperforms support vector machine (SVM) method.