Integrating machine learning in embedded sensor systems for Internet-of-Things applications

Published on Dec 1, 2016 in ISSPIT (International Symposium on Signal Processing and Information Technology)
· DOI :10.1109/ISSPIT.2016.7886051
Jongmin Lee4
Estimated H-index: 4
(ASU: Arizona State University),
Michael E. Stanley3
Estimated H-index: 3
(NXP Semiconductors)
+ 1 AuthorsCihan Tepedelenlioglu26
Estimated H-index: 26
(ASU: Arizona State University)
Interpreting sensor data in Internet-of-Things applications is a challenging problem particularly in embedded systems. We consider sensor data analytics where machine learning algorithms can be fully implemented on an embedded processor/sensor board. We develop an efficient real-time realization of a Gaussian mixture model (GMM) for execution on the NXP FRDM-K64F embedded sensor board. We demonstrate the design of a customized program and data structure that generates real-time sensor features, and we show details and training/classification results for select IoT applications. The integrated hardware/software system enables real-time data analytics and continuous training and re-training of the machine learning (ML) algorithm. The real-time ML platform can accommodate several applications with lower sensor data traffic.
  • References (49)
  • Citations (17)
📖 Papers frequently viewed together
2 Authors (Y. Weng, Alex Doboli)
11 Citations
23 Citations
7 Citations
78% of Scinapse members use related papers. After signing in, all features are FREE.
Nov 1, 2016 in ASILOMAR (Asilomar Conference on Signals, Systems and Computers)
#1Henry BraunH-Index: 6
#2Pavan TuragaH-Index: 26
Last. Cihan TepedelenliogluH-Index: 26
view all 4 authors...
We examine a potential technique of performing a classification task based on compressively sensed (CS) data, skipping a computationally expensive reconstruction step. A deep Boltzmann machine is trained on a compressive representation of MNIST handwritten digit data, using a random orthoprojector sensing matrix. The network is first pre-trained on uncompressed data in order to learn the structure of the dataset. The outer network layers are then optimized using backpropagation. We find this app...
5 CitationsSource
#1Ace DimitrievskiH-Index: 2
#2Eftim ZdravevskiH-Index: 9
Last. Vladimir TrajkovikH-Index: 9
view all 4 authors...
As the research in Ambient Assisted Living (AAL) matures, we expect that data generated from AAL IoT devices will benefit from analysis by well established machine learning techniques. There is also potential that new research in ML and Artificial Intelligence (AI) can be used on data generated from the sensors used in AAL. In this paper we present a survey of the research in the related topics, identify its shortcomings and propose future work that will integrate these fields by collecting ambi...
9 CitationsSource
#1Stefan Bosse (University of Bremen)H-Index: 9
The Internet-of-Things (IoT) gets real in today's life and is becoming part of pervasive and ubiquitous computing networks offering distributed and transparent services. A unified and common data processing and communication methodology is required to merge the IoT, sensor networks, and Cloud-based environments seamless, which can be fulfilled by the mobile agent-based computing paradigm, discussed in this work. Currently, portability, resource constraints, security, and scalability of Agent Pro...
9 CitationsSource
#1Henry Braun (ASU: Arizona State University)H-Index: 6
#2S. T. Buddha (ASU: Arizona State University)H-Index: 3
Last. Devarajan Srinivasan (ASU: Arizona State University)H-Index: 7
view all 7 authors...
As more utility scale photovoltaic (PV) power plants are installed, there is a need to improve monitoring and management of PV arrays. A procedure is presented here for optimizing the electrical configuration of a PV array under a variety of operating conditions. Computer simulations and analysis with synthetic and real data are presented in this paper. The performance of the optimization system is evaluated for a variety of partial shading conditions using a SPICE circuit simulator. In general,...
28 CitationsSource
Apr 18, 2016 in MELECON (Mediterranean Electrotechnical Conference)
#1Sunil Rao (ASU: Arizona State University)H-Index: 5
#2David Ramirez Dominguez (ASU: Arizona State University)H-Index: 1
Last. Andreas Spanias (ASU: Arizona State University)H-Index: 29
view all 11 authors...
Monitoring utility-scale solar arrays was shown to minimize cost of maintenance and help optimize the performance of the array under various conditions. In this paper, we describe the design of an 18 kW experimental facility that consists of 104 panels fitted with smart monitoring devices. Each of these devices embeds sensors, wireless transceivers, and relays that enable continuous monitoring, fault detection, and real-time connection topology changes. The facility enables networked data exchan...
15 CitationsSource
Mar 20, 2016 in ICASSP (International Conference on Acoustics, Speech, and Signal Processing)
#1Huan Songg (ASU: Arizona State University)H-Index: 1
#2Jayaraman J. Thiagarajan (LLNL: Lawrence Livermore National Laboratory)H-Index: 16
Last. Pavan Turaga (ASU: Arizona State University)H-Index: 26
view all 5 authors...
The pervasive use of wearable sensors in activity and health monitoring presents a huge potential for building novel data analysis and prediction frameworks. In particular, approaches that can harness data from a diverse set of low-cost sensors for recognition are needed. Many of the existing approaches rely heavily on elaborate feature engineering to build robust recognition systems, and their performance is often limited by the inaccuracies in the data. In this paper, we develop a novel two-st...
6 CitationsSource
#1Mihai DatcuH-Index: 1
2 Citations
#1Visar Berisha (ASU: Arizona State University)H-Index: 13
#2Alan Wisler (ASU: Arizona State University)H-Index: 5
Last. Andreas Spanias (ASU: Arizona State University)H-Index: 29
view all 4 authors...
Information divergence functions play a critical role in statistics and information theory. In this paper we show that a nonparametric f-divergence measure can be used to provide improved bounds on the minimum binary classification probability of error for the case when the training and test data are drawn from the same distribution and for the case where there exists some mismatch between training and test distributions. We confirm these theoretical results by designing feature selection alg...
41 CitationsSource
Jan 1, 2016 in IOT (The Internet of Things)
#1Satyavrat Wagle (Vishwakarma Institute of Technology)H-Index: 1
The emergence of the paradigm of Internet of Things (IoT) has necessitated the development of machine-to-machine (M2M) protocols geared towards wireless sensor network interfacing to the Internet and implementing machine learning algorithms over the cloud. This paper discusses the viability of the MQ Telemetry Transport (MQTT) protocol for such applications. This paper introduces MQTT along with its merits and demerits and suitability towards IoT applications. Then it outlines an implementation ...
9 CitationsSource
#1Mahesh K. Banavar (Clarkson University)H-Index: 15
#2Jun Jason Zhang (DU: University of Denver)H-Index: 15
Last. Antonia Papandreou-Suppappola (ASU: Arizona State University)H-Index: 20
view all 10 authors...
We provide an overview of recent work on distributed and agile sensing algorithms and their implementation. Modern sensor systems with embedded processing can allow for distributed sensing to continuously infer intelligent information as well as for agile sensing to configure systems in order to maintain a desirable performance level. We examine distributed inference techniques for detection and estimation at the fusion center and wireless networks for the sensor systems for real time scenarios....
15 CitationsSource
Cited By17
#1Hamidreza Keshavarz (TMU: Tarbiat Modares University)H-Index: 3
#2Mohammad Saniee Abadeh (TMU: Tarbiat Modares University)H-Index: 12
Last. Reza Rawassizadeh (BU: Boston University)H-Index: 12
view all 3 authors...
One of the fundamental challenges for running machine learning algorithms on battery-powered devices is the time and energy needed for computation, as these devices have constraints on resources. There are energy-efficient classifier algorithms, but their accuracy is often sacrificed for resource efficiency. Here, we propose an ultra-low power binary classifier, SEFR, with linear time complexity, both in the training and the testing phases. The SEFR method runs by creating a hyperplane to separa...
With the rise of Internet of Things (IoT), low-cost resource-constrained devices have to be more capable than traditional embedded systems, which operate on stringent power budgets. In order to add new capabilities such as learning, the power consumption planning has to be revised. Approximate computing is a promising paradigm for reducing power consumption at the expense of inaccuracy introduced to the computations. In this paper, we set forth approximate computing features of a processor that ...
#1Babatunde Seun Awoyemi (University of Pretoria)H-Index: 4
#2Attahiru Sule Alfa (UM: University of Manitoba)H-Index: 27
Last. Bodhaswar Tikanath Jugpershad Maharaj (University of Pretoria)H-Index: 14
view all 3 authors...
Fifth generation (5G), the currently evolving communication standard, promises better performance in terms of capability, capacity, speed, latency, etc. than recent technologies such as WiMax, LTE and LTE-Advanced. Similarly, the internet-of-things (IoT), the newly developing internet computing paradigm, has the potential for providing seamless, efficient human-device and device-device communication and connectivity. Both 5G and IoT technologies are definite key players in achieving a smart, int...
#1Emmanuel A. Castillo (CSUSM: California State University San Marcos)H-Index: 1
#2Ali Ahmadinia (CSUSM: California State University San Marcos)H-Index: 19
Distributed cameras have been used widely for real-time image recognition. There are two main approaches in distributed camera systems: 1. The cameras are equipped with a powerful high-end processor for local image processing, 2. Low-cost cameras with resource-constrained processors are used for capturing the images and transferring them to a cloud server for classification purposes. The first approach is costly and not scalable. The second approach is too slow for real-time object detection due...
#1Sauptik DharH-Index: 4
#2Junyao GuoH-Index: 3
Last. Mohak ShahH-Index: 9
view all 6 authors...
The current paradigm for using machine learning models on a device is to train a model in the cloud and perform inference using the trained model on the device. However, with the increasing number of smart devices and improved hardware, there is interest in performing model training on the device. Given this surge in interest, a comprehensive survey of the field from a device-agnostic perspective sets the stage for both understanding the state-of-the-art and for identifying open challenges and f...
4 Citations
#1Ikram Ud Din (University of Haripur)H-Index: 9
#2Mohsen Guizani (Qatar University)H-Index: 54
Last. Valery V. KorotaevH-Index: 10
view all 5 authors...
Abstract Machine learning is one of the emerging technologies that has grabbed the attention of academicians and industrialists, and is expected to evolve in the near future. Machine learning techniques are anticipated to provide pervasive connections for wireless nodes. In fact, machine learning paves the way for the Internet of Things (IoT)—a network that supports communications among various devices without human interactions. Machine learning techniques are being utilized in several fields s...
16 CitationsSource
#1Kristen Jaskie (ASU: Arizona State University)H-Index: 1
#2Andreas Spanias (ASU: Arizona State University)H-Index: 29
This paper will address the Positive and Unlabeled learning problem (PU learning) and its importance in the growing field of semi-supervised learning. In most real-world classification applications, well labeled data is expensive or impossible to obtain. We can often label a small subset of data as belonging to the class of interest. It is frequently impractical to manually label all data we are not interested in. We are left with a small set of positive labeled items of interest and a large set...
4 CitationsSource
#1Muhammad Usama (Information Technology University)H-Index: 4
#2Junaid Qadir (Information Technology University)H-Index: 24
Last. Ala Al-Fuqaha (Khalifa University)H-Index: 19
view all 8 authors...
While machine learning and artificial intelligence have long been applied in networking research, the bulk of such works has focused on supervised learning. Recently, there has been a rising trend of employing unsupervised machine learning using unstructured raw network data to improve network performance and provide services, such as traffic engineering, anomaly detection, Internet traffic classification, and quality of service optimization. The growing interest in applying unsupervised learnin...
20 CitationsSource
#1Namrata Majumdar (Amity University)H-Index: 1
#2Shipra Shukla (Amity University)H-Index: 2
Last. Anisha Bhatnagar (Amity University)H-Index: 1
view all 3 authors...
Over the past few years, the use of computer has spread through a lot of fields. By the connection of devices over the internet tasks can be carried out with much ease. The advancement of Machine Learning (ML) on the Internet of Things (IoT) has allowed researchers to find easier solutions to complicated day to day problems. ML techniques and algorithms such as Support Vector Machine (SVM), Convolutional Neural Network (CNN), K means Clustering, Logistic regression can be used to generate device...
1 CitationsSource
#1Farib Khondoker (ASU: Arizona State University)H-Index: 1
#2Trevor Thornton (ASU: Arizona State University)H-Index: 23
Last. Uday Shankar Shanthamallu (ASU: Arizona State University)H-Index: 3
view all 4 authors...
Internet of Things (IoT) has enabled several applications related to data analytics. In this paper, an intuitive method for optimizing activity detection data is presented. Further applications include exploring detection accuracies of physical activities such as walking intensity and movement on stairs. This method utilizes different Microcontroller Units (MCUs) with embedded sensors which are used for activity detection. Additionally, this method also incorporates supervised learning - more sp...