SysMART Indoor Services: A System of Smart and Connected Supermarkets
©©2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for allother uses, in any current or future media, including reprinting/republishing this material for advertisingor promotional purposes, creating new collective works, for resale or redistribution to servers or lists, orreuse of any copyrighted component of this work in other works. DOI: 10.1109/CCECE.2018.8447626.O. Raad, M. Makdessi, Y. Mohamad, and I. Damaj, SysMART Indoor Services: A System of Connectedand Smart Supermarkets, The st Canadian Conference on Electrical and Computer Engineering, IEEE,Quebec City, Quebec, Canada, May 13–16, 2018. P. 1–6.https://doi.org/10.1109/CCECE.2018.8447626
SysMART Indoor Services: A System of Smart andConnected Supermarkets
Omar Raad, Majd Makdessi, Yazan Mohamad, and Issam Damaj
Electrical and Computer Engineering DepartmentAmerican University of KuwaitSalmiya, KuwaitEmail: { S00026922,S00027923,S00026193,idamaj } @auk.edu.kw Abstract —Smart gadgets are being embedded almost in ev-ery aspect of our lives. From smart cities to smart watches,modern industries are increasingly supporting the Internet-of-Things (IoT). SysMART aims at making supermarkets smart,productive, and with a touch of modern lifestyle. While similarimplementations to improve the shopping experience exists,they tend mainly to replace the shopping activity at the storewith online shopping. Although online shopping reduces timeand effort, it deprives customers from enjoying the experience.SysMART relies on cutting-edge devices and technology tosimplify and reduce the time required during grocery shoppinginside the supermarket. In addition, the system monitors andmaintains perishable products in good condition suitable forhuman consumption. SysMART is built using state-of-the-arttechnologies that support rapid prototyping and precision dataacquisition. The selected development environment is LabVIEWwith its world-class interfacing libraries. The paper comprisesa detailed system description, development strategy, interfacedesign, software engineering, and a thorough analysis and eval-uation.
Index Terms —Internet of Things, Supermarket, Indoor navi-gation, Health, Safety, Networks
I. I
NTRODUCTION
Shopping is one of the most frequent activities in today’sbusy schedule. Finding the required items inside a supermarketcan prove to be a hurdle, especially, in hyper- and super-markets. Although going from one section to another canprovide a healthy amount of walk for the day, arriving tothe item’s location and discovering that it ran out of stock issometimes annoying. In addition, verifying that temperature-controlled perishable food are in good condition during trans-portation or refrigeration can be challenging to track. A variety of smart shopping systems are presented in theliterature. Wang and Yang [1] worked on developing a cart-based system for smart shopping. The system detects theusers’ presence using sensors connected to the cart’s handleand collect their position, then provides promotional info forproducts related to their position. The system uses a mesh ofwireless routers to detect position and transmit data, and asystem onboard the cart to display information. The systemhas four types of operations including behavioral analysis,where the cart senses the actions done by the customerand provides information based on the actions. In addition,operations include Query and Answer, where the customeruses the devices LCD screen to get information. The thirdtype is asking for help where the customer can get assistancefrom the supermarket’s employee by requesting help fromthe system. The last type is equipment examination wherethe system probes the cart to check its status whether itsmalfunctioning or not being used.Another cart-based smart shopping is developed by Alkha-lawi et al. [2]. They implemented Smart Cart, an automatedpersonal guidance shopping system. They used RFID tagsplaced over the supermarket’s aisles for positioning with anRFID reader attached to the cart. The authors implementedtwo main features: ”Find your way” and ”Track your trolley”.The first is to locate items in the supermarket, the latter fortracking companion’s cart.Detecting the position is also achieved by Philips in a super-market. The author in [3] discusses Philips’ smart light systemthat uses the Li-Fi technology or Visible Light Communicationto provide consumers with useful information based on theirlocation inside the store by using the existing lighting fixtures. a r X i v : . [ ee ss . SP ] M a y he system uses the lighting of the supermarket to transmitdata to smartphones’ camera at high frequency undetectedby the eye thus making it easier to implement. Several othercompanies are developing similar systems.Other options for positioning includes GPS, Wi-Fi, cellularsignals, Q-Track, Ultra-wideband (UWB) and RFID. Posi-tioning systems can rely on triangulation to increase theiraccuracy [4]. In [5], the author compares the current existingpositioning technologies for indoor usage. The GPS offers alow accuracy of 10 meters but fails indoor. Wi-Fi and cellularhave an accuracy of 2 meters, while Q-Track has an accuracyof 15 cm outdoor and 40 cm indoor. The UWB uses lowpower consumption and has an accuracy of 10 cm, but can beeasily jammed indoor as well as signal bouncing problems.In addition, the authors in [6] describe existing problemswith current RFID systems and survey potential solutions forproximity detection to use for positioning.RFID are also used to track objects in real time. In [7], theauthor discusses the advantages of using RFID tags to trackfood from the farm to the consumer. Instead of destroying allthe products due to contamination, it can be tracked to thesource of contamination and decrease the amount of wastedresources. RFID tags can be designed to track the location andtime of loading and unloading products during the distributionprocess from the processing plant to the grocery store. Theproblem with RFIDs is their cost and the need of compatibleimplementations to read the data stored by different manufac-turers and users. The RFIDs can become more attractive whencombined with battery-operated circuitry to log temperatureand other data, and estimating the expiry date of products.Moreover, fewer people will get food poisoning as a result oftheir awareness of product’s expiry date.In [8], the authors present the implementation of VirtualInstrumentation (VI) based system used for remote monitoringof selected environmental parameters: humidity, temperature,light intensity and methane. Distance operation of the appli-cation is available via iOS apps. The authors used DynamicNear Field Communication (DNFC) to transmit data whichtracks environmental parameters of desired objects and displayit in real-time. The authors in [9] state that with the expansionof Internet-of-Things (IoT), smart systems are trending in themarket. Remote access and control presents a challenge forsmart home. To solve the challenge, the authors propose asmart home monitoring system which supports data transmis-sion between local ZigBee network and remote Internet net-work. Another attempt to control devices was done by Damajet al. [10] by implementing a hardware interface connectedto a computer that allows controlling external devices suchas doors or televisions via a web interface accessible usingmobile phones.In this paper, we present an indoor system for connectedsmart supermarkets—SysMART. SysMART Indoor servicesaim at making shopping process simple and safe using IoT. Inaddition, SysMART provides indoor navigation by automat-ically monitoring the cart location, searching for alternativebranches, checkout lane improvement based on number of items to be bought, and food safety using a DNFC tag tomonitor the status of perishable goods. This paper exploresthe system implementation and the components used, thenevaluates and analyzes it is effectiveness.This paper is organized so that Section II presents the sys-tem design, organization and architecture. Section III presentsthe system implementation. A thorough analysis and evalua-tion with a deployment example are presented in Section IV.Section V concludes the paper and sets the ground for futurework.II. S YS MART O
RGANIZATION AND A RCHITECTURE
SysMART consists of three subsystems (see Fig. 1), namely,Indoor, Food Tracker and Outdoor [11]. SysMART includesLocal and Main servers. The Indoor subsystem is responsiblefor providing services inside the supermarket such as navi-gation, finding items and checking-out. The Food Tracker isresponsible of providing status information about perishablegoods. The Outdoor subsystem takes care of customers beforereaching the store.The Local server is used to store data related to a super-market’s branch and push the data to the Main server whichis responsible for communicating with the customer over amobile device. The Main server’s database (Fig. 2) holdsall the functional data for SysMART. The database consistsof tables for inventory, products, store and other informationwhere it gets update periodically from branches’ Local server.The store table contains store location, traffic and parkingstatus. In addition, the inventory table provides a count andavailability of products. Finally, the location table stores cartsposition.The Indoor subsystem provides the customer with thenecessary tools to improve their shopping experience insidethe supermarket. The Indoor subsystem features navigationusing a pair of RF tags and RF reader mounted on the cartand sends the cart position to the Local server over Wi-Fi,fastest-check lane, alternative branches for out of stock items,requesting help on the move and reporting damaged cart usingand Android application.The Food Tracker subsystem allows customers to have asummary of perishable products status, such as, productiondate, expiry date, temperature and humidity info, timestampfor each distributing plant. The Food Tracker allows thecustomer to verify the condition in which the product wasstored in. In addition, it can provide a log of all recorded datato the supermarket management or health department.The Outdoor subsystem is capable of checking the parkinglots status using a sensor and updates the Local server datausing a mesh wireless network. A similar system is used fordetecting traffic status. It checks for products’ availability in aspecific branch or query all branches by communicating withthe main server.III. S YS MART I
MPLEMENTATION
The hardware used is categorized under two subsystems:Indoor and Food Tracker. Starting with the Indoor subsystem nternetSupermarket
ServerSupermarket
Server
Main Server
Outdoor AP Database
Ultrasound sensor Xbee node
Controller Local Database
Local
Database Indoor AP temperatureDNFC tag
MCU
Basket Cart
RFID Tag (positioning)
RFID
ReaderMyRIO
Controller
Arduino FioUltrasound sensor Xbee node
Arduino Fio
Raspberry Pi Arduino
Arduino
Fig. 1. SysMART System Architecture: To the left the Outdoor Services, in the middle the backend and to the right the Indoor Services which requires the customer to have a cart and a smartphoneto get its services. For indoor navigation, a Low Frequency(LF) RFID tags rated at 125 kHz are used to identify thelocation. An LF RFID reader is attached to the bottom of thecart to read the tags. Although the standard LF RFID has arange of few centimeters, non-standard implementation canreach up to 1 meter under controlled conditions. Due to thecart metallic frame, the 1-meter range is reduced to 20 cm.The RFID reader is connected to a National Instruments (NI)myRIO embedded hardware device using Wiegand interfacewhich uses 2 data lines D0 for 0-bit and D1 for 1-bit. The NimyRIO communicates with a controlling server over Wi-Fi.The Food Tracker subsystem makes use of Near FieldCommunication (NFC) available on most smartphones. TheFood Tracker is a DNFC tag, unlike regular NFC tag, the datastored in the DNFC can be updated using an attached micro-controller. The updating feature makes the DNFC suitable forlogging applications. The microcontroller selected is a TexasInstruments (TI) MSP430FR5969. In addition, two types ofsensors are used to get the temperature and humidity readings.The temperature sensor has a high accuracy of 0.5 °C and lowcurrent usage rated at 7 µA active mode, and 1 µA sleep mode.Similarly, the humidity sensor has an accuracy of 3% and lowcurrent usage ranging from 200 nA (sleep mode) and 820 nA(active mode). Both sensors use I2C for communication.Similar to the hardware subsystems, the software has severalcomponents: Backend, Indoor and Food Tracker. The backendof the project is Microsoft Azure, on which a database anda webservice are deployed to store the store data, trafficand parking status, indoor location, and other information.The information is retrieved from the database through thewebservice as well. Fig. 2 presents the database design. Storetable provides store id primary and unique key for eachsupermarket branch.The store id key is used to link entries inlocation, inventory, cart location and mapping tables with eachstore. Each product has a product id key in product table usedin the inventory table to list the supermarket’s inventory. Thelocation id key in location table stores location id specific foreach store and used to map location of inventory product using inventory store_idFK product_location availability_in_store location location_idPK tag mapping store_idFK map traffic traffic_idPK traffic_name store store_idPK store_namestore_long cart_location store_idFK cart_locationFK product product_idPK product_name store_lat store_parking_countavg_trafficstore_parking_availablelocation_id_1FK location_id_2FK store_idFK cart_idPK product_idFK price
Fig. 2. Backend Database Schema the product location field in inventory table and customerlocation for indoor navigation using the cart location field incart location table.The second component is the indoor software subsystem,which is made by developing a myRIO program and anAndroid application. The myRIO program is developed usingNI LabVIEW a visual programming language. The deployedsoftware reads the location tag ID from the RFID readerusing Wiegand 26 protocol. The application server can onlyreceive data from one myRIO at a time, to avoid havingtwo carts sending data at the same time, a random delay isintroduced before transmitting to minimize data collision. Thedata transmitted consists of 16-bit integer store ID, 16-bitinteger cart ID, and a 6 hexadecimal characters for tag ID,totaling to 80 bits with overhead may reach up to 128 bits.The used Wi-Fi transmits at 54 Mbps, therefore, each data sentwill take 2.26 µs or 442,368 data/s, making the random delayup to 1 second adequate to avoid data collision.The application server is developed using LabVIEW and canrun on a personal computer or a tablet. The app server’s mainfunctionality is to update the database using the data it receivesfrom the myRIOs deployed on the carts. A second functional-ity is to display the request of assistance and malfunctioning ig. 3. SysMART Indoor App Activities and Transactions carts. The user interaction with the cart is accomplished byusing an Android app developed using Android Studio. Thecustomer enters the store ID and cart ID to get the relateddata of the cart. The app provides the current location andthe location of the item desired (see Fig. 3). Using the app,customers can get the fastest lane for checkout. The appalso provides the ability to request assistance and report cartmalfunctioning.The last component is the Food Tracker. The Food Trackertag is developed using TI Code Composer Studio (CCS) whichallows to build and debug the code. As the data stored in thetag are sensitive by nature, such as, production data and expirydate. The data must be tamper proof, therefore the code mustnot allow modifying any pre-populated field without resettingthe tag. Moreover, resetting the tag introduces multiple usagesof the tag. An additional security measure is a 20-characterpassword provided to the tag at initialization time after a reset.The reset can be performed after confirming the password andit clears the settings. To avoid tampering with time logging,all timestamps are provided by the microcontroller’s real-time-clock at the time of setting a field (when a product arrives ata distributing plant or departs from one). For temperature andhumidity logging, due to limited space in the microcontroller’smemory, the data logged after the difference between a newreading and the last recorded reading exceeds a specificthreshold. Also to minimize battery consumption and writingspace, the sensor readings are set at specific intervals. Thecommunication with the tag is done over NFC. The DNFCreads or writes data, and after the reader/writer stops the signal, (a) Read Tag (b) Graph DataFig. 4. Food Tracker Master App the DNFC notifies the microcontroller of an action was done.By default, the DNFC holds the latest summary data in itsmemory. The microcontroller is Byte addressable; to avoiddata corruption each log item should be stored separately. Thelog item contains the timestamp, temperature and humidityvalue in raw binary format, this will reduce the space needed tostore the data. The timestamp for the log item is an increment echnology
LabVIEW + NI Devices LabVIEW + Arduino PHP + ArduinoAdvantages:
1. Graphical programming
2. NI devices fully supported
Advantages:1. Simple to prototype
2. Online support
Advantages:
1. Speed and stability
2. Easy to use for simple websitesDisadvantages:
1. Steep learning curve2. Hard to debug for complex applications Disadvantages:
1. Not all Arduino devices are supported
2. Very limited interaction between them Disadvantages:
1. Both Arduino and PHP are limited
2. They cannot directly interact with each others
Fig. 5. Technologies of the first log item’s value, so the first item will have atimestamp of 0. The temperature sensor value occupies 12bits and the humidity sensor occupies 14 bits for the highestpossible accuracy, the total sensors bit count is 26 bits. Thenext 8-bit multiples are 32, 40 and 48. Choosing the 48-bitformat provides 22 bits for timestamp, with increment of 1minute. Thus the timestamp allows for a maximum of 7.98years of log tracking. To communicate with the tag, a masterapplication is developed for Android phones. The applicationprovides the capabilities of setting the tabs fields, reading thesummary info and plot a graph from the detailed log as shownin Fig. 4. IV. A
NALYSIS AND E VALUATION
SysMART is built and realized successfully and proves tobe effective in application. First, the goal of SysMART is tohelp the society and economy. By connecting the stores, thecustomer can have access to information from other stores.Customers should have an account with the supermarket tofacilitate the interaction with the store, to locate friends andrelatives and other features. By having an account, a paymentthrough the account may be possible at in the future. Theaforementioned features raise a concern regarding privacy,security, and safety. Customers information are very sensitiveand must be protected including their location and they musthave an option to disable storing their location. Since shoppingexperience becomes improved as a result of the project, cus-tomers may become more inclined to spend more, therefore,affecting the purchasing power.Another feature, which raises concerns, is the food tracking.The tracking tags must be reusable, to reduce additionalwaste and extra resources. The addition of a tracker mayalso increase the price tag of the product from the trackercost and extra processing overhead that comes with it. Thecomponents used were chosen based on availability, perfor-mance, scalability and cost. From the tree (Fig. 5) the mostpromising technology for our purposes is the LabVIEW andNI devices. Although the decision is to use NI devices, butother microcontrollers for specific purposes.The RFID devices are grouped by frequency. There are threemain categories: Low Frequency (LF), High Frequency (HF)and Ultra High Frequency (UHF). The HF or NFC, operates
Position
RFIDWiFi UWB Bluetooth VLCAdvantages:1. Widely used for communication
Disadvantages:
1. Low accuracy
Advantages:1. High accuracy of few centimetersDisadvantages:
1. Expensive
Advantages:1. Low powerDisadvantages:1. Low accuracy
Advantages:1. Widely used for communicationDisadvantages:1. Limited info2. Best with LED
Advantages:1. Cheap
2. High accuracyDisadvantages:1. Tag count and placement affect accuracy
Fig. 6. Positioning Technology
Processor
NI myRIO Arduino
Advantages:
1. Onboard WiFi
2. Small form factor
3. Compatible with Arduino shields
4. FPGA + MCU
5. Powerful application
Advantages:1. Easy to use and prototype2. Wide range of extension shields3. C/C++ language
4. Online support
Disadvantages:
1. Expensive
2. FPGA design
3. Requires knowledge of LabVIEW programming language
Disadvantages:
1. Simple application
2. Difficult to debug
Fig. 7. Cart Processor at 13.56 MHz, is used in different fields, from identification,smart wallet and transmitting data and has a maximum rangeof 1 m. For the Food Tracker, the NFC is the most suitablesolution as it allows data communication although UHF RFIDprovides similar functionality, the reader and antenna arevery expensive, also detecting the tags from a distance maycause some confusion as to which product is being tracked.Furthermore, most Android smartphones includes and NFCreaders which makes the NFC more favorable over UHF RFID.From Fig. 6 both VLC and RFID are valid options; but asVLC is still a new technology, RFID tags are more favorableto implement positioning. The 10 cm NFC’s range is shortand requires the device to be close to the ground, and theUHF RFID reader is more expensive than its alternatives.While the LF RFID has a short range of 10 cm, non-standardimplementation can support up 80 cm using a 30 cm by 30cm reader.Next, deciding on processing unit to be used from Fig. 5, theNI devices for critical applications is chosen as the processingunit for the cart or basket. NI products are compared to otherMCUs, such as, Arduino as shown in Fig. 7.Food tracker is a device that allows the consumer, vendorand distributor to check the safety of the product. As thelifespan of a product can be affected by several factors, mainlytemperature, so the food tracker must log the temperatureregularly and inform the user of the maximum temperaturereached, average temperature and estimated expiry date basedon the logged temperatures. Consumers can use their smart- ag Data Storage
NFC tag DNFC Tag
Advantages:1. No power source needed2. Fast, easy and wireless Advantages:1. Fast, easy and wireless2. Data updated using an MCUDisadvantages:
1. Need external device to update data
Disadvantages:1. Most references are aimed at experts2. Power source needed
Fig. 8. Food Tracker Tag Type
MCU
TI MSP430EXP5969 Arduino Raspberry PiAdvantages:
1. Onboard debugging2. Code available for development
3. Ultra-low power consumption4. Sleep mode Advantages:
1. Different form factor and single chip usage
2. Code available for development
3. Low power consumption4. Sleep mode
Advantages:
1. Linux OS
2. Python programmingDisadvantages:1. Steep learning curve Disadvantages:1. Limited capabilities Disadvantages:
1. High power consumption
2. Size cannot be reduced
Fig. 9. Food Tracker Processor phones to get a summary of the food status and the vendorscan get a detailed log of the storage conditions and producthand-off along the distribution chain.A useful technology used to access data in a simple andfast way is NFC tags. NFC tags can be accessed wirelesslyto retrieve data and are passive devices so they require nopower source to be active as they become active by the powerprovided wirelessly by the reader device. The NFC tags canbe used to store the condition of the product. But they needto be accompanied by a writer device at a regular intervalto update the value, which makes them inappropriate for theapplication because the tag must be updated as needed and thesystem should be independent from external components. Analternative to the NFC tag is DNFC tag (Fig. 8). The DNFChas the same features as a regular NFC tag, but with the abilityto interface with a microcontroller to update the data. Sincethe DNFC tag needs a microcontroller, the TI microcontrollerwas chosen. The TI microcontroller offers on-board real timeclock and an FRAM to store data. Two other alternatives wereArduino and Raspberry Pi (Fig. 9).To evaluate SysMART from the user’s perspective, a surveyis deployed on a population of 58 people between the ages of18 to 64. The survey collects the average time of shoppingand the feature ratings from customers’ point-of-view. Fig. 10and Fig. 11 provide a bar-graph of the survey’s results.To estimate the cost and feasibility of SysMART, a simu-lated case study was applied to a local supermarket chain. Thesupermarket chain owns several branches and the simulationwas applied to one of the busiest mostly frequented branch.The supermarket has 150 carts. A total of 230 RFID tagsare carefully distributed among different sections. An overallbudget for the simulated case study is $88,730 and $15 per
18 to 24 25 to 34 35 to 44 45 to 54 55 to 64 P e o p l e Age
How long do you spend on grocery shopping and checkout?
15 - 20 min
20 - 30 min
30 - 60 min > 60 min
Fig. 10. Survey: Grocery Time
18 to 24 25 to 34 35 to 44 45 to 54 55 to 64 R a t i n g ( i s h i g h , i s l o w ) Age
Feature rating
Food storagetemperature
Checkout linesuggestion
Navigation to findthe products in thestore
Fig. 11. Survey: Feature Rating food tag. V. C
ONCLUSION
SysMART is a modern IoT system that offers fast andsafe shopping experiences. SysMART supports a bouquet offeatures that include indoor navigation, fast checkouts, andfood tracking. SysMART interacts with the customers’ smart-phone to provide real-time information. The cost associatedwith offering a premium service to customers is expected tohave a high return on investment–with more customers visitingthe supermarket for efficient grocery shopping and checkout.Future works include motorizing and tracking the cart toallow smoother shopping for elderly and kids, and facilitatetransportation of heavy items. Moreoever, future work includesaccelerating security aspects and database queries using high-performance computing [12]–[15].R
EFERENCES[1] Y.-C. Wang and C.-C. Yang, “3s-cart: a lightweight, interactive sensor-based cart for smart shopping in supermarkets,”
IEEE Sensors Journal ,vol. 16, no. 17, pp. 6774–6781, 2016.[2] R. Alkhalawi, R. Alluqman, N. Alalawi, and I. Damaj, “Smart cart:A ubiquitous shopping guidance system,” in , 2014.[3] M. LaMonica. (2014, Feb.) Philips creates shop-ping assistant with leds and smart phone. [On-line]. Available: https://spectrum.ieee.org/tech-talk/computing/networks/philips-creates-store-shopping-assistant-with-leds-and-smart-phone[4] H. Liu, H. Darabi, P. Banerjee, and J. Liu, “Survey of wireless indoorpositioning techniques and systems,”
IEEE Trans. Syst., Man, Cybern.C, Appl. Rev. , vol. 37, no. 6, pp. 1067–1080, 2007.[5] D. Schneider, “You are here,”
IEEE Spectrum , vol. 50, no. 12, pp. 34–39,2013.6] J. Q. Anderson and H. Rainie,
The Internet of things will thrive by 2025 .Pew Research Center [Internet & American Life Project], 2014.[7] E. Strickland, “Tracking Every Tomato,”
IEEE Spectrum , vol. 50, no. 6,pp. 72–77, 2013.[8] F. Zhou and Q. Li, “Parking guidance system based on zigbee andgeomagnetic sensor technology,” in
Distributed Computing and Appli-cations to Business, Engineering and Science (DCABES), 2014 13thInternational Symposium on . IEEE, 2014, pp. 268–271.[9] W. Yiqi, H. Lili, H. Chengquan, G. Yan, and Z. Zhangwei, “A zigbee-based smart home monitoring system,” in
Intelligent Systems Design andEngineering Applications (ISDEA), 2014 Fifth International Conferenceon . IEEE, 2014, pp. 114–117.[10] I. Damaj, R. Abaji, and Z. Osman, “Computer interfacing electronicboards for building management systems,” in
Proceedings of the SixthIASTED International Conference on Advances in Computer Scienceand Technology , vol. 689, no. 022, 2010, p. 129.[11] Y. Mohamad, M. Makdessi, O. Raad, and I. Damaj, “SysMART OutdoorServices: A System of Connected and Smart Supermarkets,” in ,2017, pp. 98–102.[12] I. W. Damaj, “Parallel algorithms development for programmable de-vices with application from cryptography,”
International Journal ofParallel Programming , vol. 35, no. 6, pp. 529–572, 2007.[13] I. Damaj, “Performance analysis of linear algebraic functions usingreconfigurable computing,” Ph.D. dissertation, American University ofBeirut. Department of Electrical and Computer Engineering, 2001.[14] I. Damaj and H. Diab, “Performance analysis of extended vector-scalaroperations using reconfigurable computing,” in
Computer Systems andApplications, ACS/IEEE International Conference on. 2001 . IEEE,2001, pp. 227–232.[15] D. Singh, G. Tripathi, and A. J. Jara, “A survey of Internet-of-Things:Future vision, architecture, challenges and services,” in