Networked Wireless Sensors, Active RFID, andHandheld Devices for Modern Car Park

Management

Djamel Djenouri 1,∗, ElMouatezbillah Karbab 1,Sahar Boulkaboul 1, Antoine Bagula 2

1: CERIST Research Center, Algiers, Algeria2: University of the Western Cape, Cape town, South Africa

Emails: ddjenouri@acm.org, m.karbab, bsahar@mail.cerist.dz, bbagula@uwc.ac.za

Abstract

Networked wireless sensors, actuators, RFID, and mobile computing technologies are explored in this paperon the quest for modern car park management systems with sophisticated services over the emerging internetof things (IoT), where things such as ubiquitous handheld computers, smart ubiquitous sensors, RFID readersand tags are expected to be interconnected to virtually form networks that enable a variety of services. Afteran overview of the literature, we propose a scalable and low-cost car parking framework (CPF) based on theintegration of aforementioned technologies. A preliminary prototype implementation has been performed, as well asexperimentation of some modules of the proposed CPF. The results demonstrate proof of concept, and particularlyreveal that the proposed approach for WSN deployment considerably reduce the cost and energy consumptioncompared to existing solutions.

I. INTRODUCTION

Automatic car park management using modern technologies becomes an inevitable option to rationalizetraffic management in modern cities. With the constant increase of vehicles in agglomerations worldwide,finding a car park with available spots becomes a worry for the citizens in modern societies. This increaseis the main cause of high fuel consumption, increased traffic congestion, and the dramatic impacts on theenvironment, drivers’ health and well-being. New technologies such as mobile computing and emerginghandheld devices [1] (smart phones, smart watches, etc.), sensor/actuator [2], and Radio FrequencyIdentification (RFID) [3], [4] may play a pivotal role to modernize and improve car park management,and to provide modernized services via a new first mile of the Internet in the emerging the ”Internet-of-Things (IoT)”. These devices can be endowed with an IP address and outfitted to the objects of thesmart parking and/or its infrastructure to form smart objects. The latter may be interconnected into an IoTplatform with sensor devices used for empty park spot localization, while the RFID devices are used forparking spot authorization, car localization, and theft prevention. The users will use their mobile handholddevices to access the services over local networks and internet. However, integration of these technologiesyields a heterogenous environment with many challenges in terms of deployment, communication andprotocol engineering. Some of these challenges have been addressed in this paper to propose a CarParking Framework (CPF), which is builds around a modular approach to enable a variety of servicessuch as driver guidance, automatic payment, parking lot retrieval, security and vandalism detection. Itspartial implementation as a lab prototype is also described in this paper, where some modules have beentested and evaluated by real experiments.

Smart parking solutions that are closely related to the one proposed in this paper have been presentedin [5], [6] and [7]. These solutions use a single wireless mote per parking lot, which is outfitted with asensor for vehicle detection. The sensor can be a magnetic sensor, an ultrasonic sensor, an optical sensor,etc. When a car is in sensing field of some sensor (a parking spot), the sensor status will change anda signal will be transmitted to the connecting mote. The latter processes the incoming signal to decideabout possible detection. Using a mote in every spot has drawbacks. First, the cost will be high (both interms of financial cost for deployment, and energy consumption when in operation). Second, it increases

the collision domain and creates more interference. An interesting study about indoor communicationsin [8] shows the dramatic impact of mobile metallic objects over the mote on wireless communicationchannels, which is a typical situation that takes place when detecting a vehicle that parks over the sensormote. It is thus useful to separate the communication module from the sensing area. To overcome thedrawbacks of using a single mote per parking spot, we propose the clustering of a bunch of sensors andthe use of hybrid wire/wireless communications. Every sensor-board is connected to a micro-controllerunit (MCU), then the bunch of sensors is connected in serial mode to a single mote using serial wiredcommunication. The mote is considered as a master of the communication, and the sensors are consideredas slaves. Wireless communication is used to connect the mote to the rest of the network. This patternconsiderably reduces the cost and the interference, as it will be illustrated in the upcoming sections.

The rest of the paper is organized as follows: The next section presents the state-of-the-art on car parkmanagement with integration of WSN and RFID. Section III introduces the proposed framework, andSection IV its implementation. Section V presents the interfaces we developed for mobile users. SectionVI is the proof of concepts of the proposed solution with some prototype experimentations and somepreliminary performance evaluation. Finally, Section VII concludes the paper.

II. SATE-OF-THE-ART

Existing car parking management systems can be divided into two categories, i) gate management,vs. ii) lot management. The systems based on gate management provide some basic services for driverssuch as the possibility to check free spot availability, reservation over internet, etc. These services areenabled by the gate management model, where information about the entries/exits are used. In the secondclass, every parking spot can be equipped with a sensor to detect the presence/absence of vehicles. Themain advantage of the gate management model is its low cost and simplicity over lot management model.However, the latter allows to provide more advanced services for drivers, such as driver guidance systemsthat orient the drivers towards the nearest free spot, free lot maps, etc. This is achieved by endowingparking lots with sensors [9], [10]. Table I provide a summarized comparison between the two models.Further, smart car parking based on lot management can be classified into two categories; multi-parkingmanagement vs. mono-parking management (table II). Multi-parking management systems can managedifferent parks in different areas. These areas may be indoor, such as mall or airport parks, or outdoor,such as street parks in urban areas. Multi-parking management may provide guiding capability for driversto the most appropriate car park. Many solutions belong to mono-parking management, while only fewsolutions are proposed for multi-parking management [11]. We are focusing in this paper on the formerclass.

Mono-park solutions are generally targeting indoor parks and focus on a single park management. Theymay provide guiding capability for car drivers within the car park. The authors of [12], [13] propose asystems using image detection. In [5], [6], [8], a wireless sensor network based mono-parking system isdescribed. In this system, wireless sensors are deployed into a car park field, and each parking lot equippedwith one sensor node. All these solutions have used TinyOS (A WSN operating system) for developingcustomized networking stack. ZigBee stack is chosen for many mono-parking solutions [7], [14], [15]due to its efficient energy consumption and flexibility in topology management. In [16], the authorspropose KATHODIGOS, which is a smart parking system that uses wireless technologies and wirelessapplications to acquire information about the status of roadside parking spaces. The obtained information istransmitted to a central information system through gateways. The solution in [17] proposes a frameworkof parking management called iParking, which monitors incomming/outgoing vehicles using a sensornetwork. iParking calculates the number of available parking spaces, and disseminates the information tothe parking’s customers. In [18], the authors describe a street parking system(SPS) based on a wirelesssensor network. The system can monitor the state of every parking space by deploying a magnetic sensornode on the space. To accurately detecting a parking car, a vehicle detection algorithm is proposed.

Lot management Gate managementCost High Low

Guiding possibility Yes NoFree lot maps Yes No

Complexity degree High LowNumber of free lots Yes Yes

TABLE I: Comparison between gate management and lot management

An adaptive sampling mechanism is used to reduce the energy consumption. Nawaz et al. [19] presentParkSense, a smartphone-based sensing system that detects a driver vacates a parking spot. Most ofthe solutions proposed thus far consider the use of either WSN, or RFID. However, we think that theintegration of both technologies in a single framework have potential impacts, as it will enable effectiveobject tracking and environment perception. This represents the purpose of the work presented in thispaper.

Mono-parking Multi-parkingScope Single park Many parks

Type of park indoor (generally) Indoor and outdoorGuiding service Only within a park Between parks

TABLE II: Comparison between multi-parking management and mono-parking management

III. SOLUTION OVERVIEW

Advanced features and services for car park managements are provided through the integration of RFID,WSN and mobile computing technologies in the proposed car parking management framework (CFP). Anactive RFID tag is used for every parked vehicle. The tag can be allocated to a subscribed customers overa long period of time (private parks), or it can be dynamically provided to the transient customers at theentrance. A sensor is installed at every parking lot and connected to a wireless mote. The framework isgeneric and allows the use of any sensing technology that detects vehicles. In the implemented prototype,ultrasonic sensor are used. The mote manages a bunch of sensors connected in serial mode. This is moreefficient than using a mote per lot as done in most state-of-the-art solutions (Fig. 1).

The proposed CPF builds upon a four-layer ubiquitous architecture. A sensor layer is used for parkingspot detection and security. A network layer is layered above the sensor layer to enable disseminationof the information between sensors and gateways. It uses both wired and wireless communication. Amiddleware on top of the network layer stores data and enables visualization. It acts as an interfacebetween the network layer and the application layer (the upper layer), where different services related tothe smart parking are implemented.

The implementation of the CPF includes two parts that are connected using a LAN network (WLAN orEthernet). The first part is the parking manager, where all data are stored. It may provide some informationfor clients over internet, e.g. parking spots availability. The payment service is also provided by the parkingmanager within a park. The level manager is the elementary part that forms the parking manager, whichwill be presented in the next subsection. The second part is the gate manager. Its main function is tocontrol the gate using WSN, and to forward the gate status to the parking manager. WSN is used withinthe level to carry the status of spots and cars to the sink in a level. This would be transmitted using aLAN network. The proposed CPF scales to a multi-level car parks.

a) single mote per spot

a) single mote per bunch of spots

Fig. 1: Mote per spot vs. clustering

A. Level ManagerThe level management is illustrated in Fig. 2. The lots manager can be viewed as the atomic component

in the level (floor) management. A WSN that contains several types of nodes is used. The included nodesare: i) the mote of a bunch of lots that clusters a set of sensor boards (one per lot), ii) a hybrid node,where an RFID reader and a wireless mote are integrated, iii) guiding nodes that are connected to aVMS (Variable Message Screen), and finally, iv) a sink node that is connected to ethernet plug to forwarddata to the parking manager. Event detection at the sensor mote (master) needs to be forwarded to thesink of a level using multi-hop communications. For this purpose, we propose to use our previouslypublished data gathering protocol, LIBP[20], [4]. The main utility of LIBP in car parking management isthe management of energy consumption in heterogeneous wireless sensor networks. In WSN with multipletypes of motes having different levels of consumption, the aim is to ensure the network traffic can bemanaged to achieve balanced lifetime for all the motes in the network. This is a typical utilization in the carparking management, featured by the co-existence of: i)simple node as guiding node with moderate energyconsumption, ii) the master node responsible for parking lots management, which consumes more power,and iii) the hybrid node (integrating the wireless mote and RFID reader) that is the most power consuming.The use of the proposed lots manager pattern, as well as the multi-hop communication paradigm betweenmotes (masters) with an appropriate data gathering protocol (LIBP) provide a scalable level manager.

B. Provided Services1) Guidance Service: An active RFID tag is allocated to every parked vehicle. Active RFID tag has

been chosen given its longer communication range compared to passive RFID tag, as well as the advancedservices it enables. The hybrid nodes (integration of wireless mote and RFID reader) should be placed atstrategic points, where they can detect the coming car and identify it using the RFID tags. Depending onthe type of parking, there are two scenarios for guiding service:

• Predefined parking spots: This is when every entered car has a predefined parking spot. In this case,the hybrid node gets the tag’s ID of the car, then it will retrieve the predefined spot of the car fromthe parking manager using the WSN within the level, as well as the LAN network between levelsand the parking manager. After that, the direction will be displayed on VMS using the guiding node.

• Random parking spots: If the entered car has not a predefined parking spot, then when the hybridnode detects the coming car using the RFID tag, it would retrieve the nearest available (optimal)parking spot from the parking manager. The calculation of the optimal parking spot is made basedon the current location of the car, and the map of free spots in the park that is stored in the server.

Fig. 2: Parking level manager pattern

2) Car Retrieval Service: A common problem for customers is when they forget where they parkedtheir cars. The proposed CPF provides a service that assists in retrieving a forgotten car location usingthe integration of WSN and RFID in the hybrid node, and an active RFID tag to be kept by customers.When a customer requests his car’s spot using a trigger in the active tag, a hybrid node gets the tag’s IDand transmits it to the parking manager. This latter checks the occupied parking spots in the database andin the parking field using the WSN. The response is returned to the appropriate guiding node for displayin the VMS.

3) Illegal Parking Detection Service: The CPF provides a mechanism to solve this problem that maytake place when the customer has a specific parking area. When a car is detected in a parking spot, anevent will be transmitted to the parking manager. This latter will request the ID of the car’s tag from thehybrid node where the car is parked. Using the tag’s ID, the car parking manager can check if this car isauthorized to access this area.

4) Security Service: In the proposed CPF, a driver may either have a priori the RFID tag (case of privateparking), or he may get it at the entrance. The tag must be kept in the car when parking. Meanwhile, thesystem confirms that a car is in a parking spot using sensors, and it matches the ID and the parking spot(refer to Sec. III-B3). Next, the customer can keep the RFID to take advantage of other services such asthe ”find forgotten car service”. However, when a car is leaving a parking spot, the system will checkthe IDs of existing tags in the area using Hybrid nodes. These IDs will be matched with the parking spot,and the parking manager commits this matching using the parking database. This is to confirm whetherthe customer’s tag is in the area or not. In other words, it will check if the driver is in the hybrid noderange, in specific level, or close to a bunch of sensors, etc. The system will trigger a car stolen event ifthe car owner is not close to the car. Timestamping of the events and time synchronization between nodesis needed to assure this service. Machanisms such as [21] [22] may be used.

5) Other Services: The flexibility of the integration networking of WSN, RFID, and handheld devicesenables the CPF to provide many services beyond the above mentioned. The following are some examplesof these services:

• Parking spot reservation: Bunch of sensors can be enhanced with a source of light in each parkingspot. These lights are controllable by the wireless mote. They can provide information about the statusof a spot, e.g., red for occupied, green for empty, yellow for reserved, and blue for out-of-service.

• Payment service: RFID tags have many utilities in the CPF. They can be used for contact-less paymentcalculation. Then the handheld devices may be used for E-payment.

Fig. 3: An overview of the prototype

• Gate management service: Another use of RFID tags is gate management. For example, a gate canbe opened automatically using an RFID reader and the vehicle’s tag at the gate.

• Availability checking over internet: The CPF provides a real-time availability checking over internet.Cars can be checked if they are in spots and drivers if they are in a car parking using their tags.

IV. PROTOTYPE IMPLEMENTATION

We partially implemented and tested the proposed CPF in lab prototype. In this prototype, client devicesconnect via TCP/IP protocol to a parking database, which is updated in real-time with the status of parkinglots (Fig. 3). Mobile applications have been developed for android devices, including customer interfacesand administrator tools.

Ultrasonic sensors are used and connected in serial mode using micro-controllers. All events are firsttransmitted to a master mote, then forwarded to the based station. This latter stores the events in Mysqldatabase for the client application. I2C protocol is used to transmit the event from the parking lots(micro-controllers with ultrasonic sensor) to the master wireless mote.

ARM-based development boards have been integrated to the level gateway, and the proposed portableand efficient design for lot manager pattern has been implemented. In order to test the efficiency of the CPFprototype, real-world experimentation on a bunch of lots has been carried out in our campus at CERISTresearch center. One of the main challenges was to bend the I2C protocol for our application, given thatthis protocol has not been designed to operate at long distances. Some changes to some parameters havebeen necessary, such as the pull resistance and clock frequency. We have used for this experiment:

• Three HC −SR04 ultrasonic sensors (Fig . 4 (a)), one is used per lot. The sensors are connected toMSP430 micro-controllers.

• Two wireless motes, a transmitter and a receiver. The transmitter is the sensor bunch master (connectedto a bunch of HC−SR04 sensors), and the receiver is for connection to the gateway, i.e., it is pluggedto the gateway. LM4F120 microcontroller has been used, along with nRF24L01 radio to form themote (Fig . 4 (b)).

• A PandaBoard: an ARM based board, as a gateway (Fig. 4 (c)).When a car is parked in the free lot, it will be detected by the mote through the appropriate sensor.

The event will be transmitted from the wireless mote (transmitter) to the receiver mote that is connectedto Pandaboard. The event will be stored in the gateway of the level, which is in our case a Pandaboard. The latter provides the service of checking the free parking spot using JSON web server over the localnetwork (WiFi or Ethernet).

a)

b) c )

Fig. 4: The equipment used: a) HC−SR04 ultrasonic sensors, b) Mote (LM4F120 microcontroller withnRF24L01 radio, c) (gateway) PandaBoard

V. MOBILE USER INTERFACE

To make the proposed framework handy and ready to use, we developed appropriate software forAndroid platforms. This includes graphical interfaces that enable the drivers to interact with smart carpark networks over the internet, using mobile phones, tablet, etc., to check availability, make reservation,payment, etc. Accesses to the database is assured using web services and through two modules: i) smartparking module, ii) web site administrator for the management. Different technologies have been used,Android as the underlying operating system, and Java as the programming language with the androidintegrated virtual machine. The user interface is constructed using an XML file, which allows for an SDK(Software Development Toolkit). The paradigm of web services is based on a components’ architecturethat uses internet protocols to manage the communication between components, to interact with a database,to make the link between an external application and the database, to provide access to content whilekeeping it safe. REST (REpresentational State Transfer) has been chosen for its stateless architecture thatgenerally runs over HTTP. REST involves reading a designated web page that contains an XML file. Thelatter describes and includes the desired content. Database server, MySQL, is defined as a managementsystem database. A set of APIs for the development of business-oriented applications is used. The J2EEarchitecture is based on the Java language that allows the deployment of components on various platforms,independently of the programming language. The API that we used are Servlet and JSP, with the web-server Apache Tomcat, which is a server provided by JBuilderX to compile and execute Servlets and JSP. Itcontains a Java virtual machine, and associated elements to provide a complete development environmentfor Java. A snapshot of the front-page graphical interface is given in Fig. 5

VI. EXPERIMENTATIONS

An experimental study has been carried out to quantitatively investigate the impact of the proposedframework in terms of engineering and economic efficiency. Two classes of performance metrics areconsidered. The first class is the engineering efficiency expressed in terms of energy consumption ofa bunch of lots for locating parking spots. The use of a bunch of sensors clustering is justified by theeasiness to use a common power source (usually a battery) for those sensors, given their close positions.This feature limits the wires used for connecting the bunch of sensors in the real deployment. Moreover,the replacement of the batteries per bunch of lots would be more practical than the separate replacementof each single sensor node. The second class is the economic efficiency, which expressed in terms of the

Fig. 5: User interface

deployment financial cost. Usually, an economically efficient deployment is expressed in terms of howmuch does the monitoring of a number of parking spots cost. We have expressed the economic cost inUS Dollar.

1 2 3 4 5 6Number of spots

0

100

200

300

400

500

600

700

800

900

Ener

gy c

onsu

mpt

ion

(mA)

The energy consumption by the spot number and solutions

Serial connectionA wireless mote in each spot

(a)

1 2 3 4 5 6Number of spots

0

50

100

150

200

250

300

Cost ($

)

The cost by the spot number and solutions

Serial connectionA wireless mote in each spot

(b)

Fig. 6: Efficiency vs. Number of spot per cluster, (a) Engineering efficiency, (b) Economic efficiency

We compared the proposed approach of using hybrid wired/wireless communication (integration in of abunch of nodes into a single mote) with the standard approach of deploying a wireless mote per spot usedin [5], [6], [8]. Empirical experimentation using real equipment has been conducted for this comparison.For practical limitations, limited number of nodes has been used in the experiments, but it is obviousthat the difference will increase with the number of nodes. For each number of nodes (from one to six),the two approaches are compared. Fig. 6 (a) depicts the cumulative energy consumption. It numerically

demonstrates that the proposed solution is more engineering efficient than using one sensor per parkingspot. Fig 6 (b) shows the cumulative economic cost vs. the number of sensor nodes. The figure alsoconfirms the relative economic efficiency of the proposed framework. From these results, it is clear thatthe difference with the standard approach considerably rises with the increase of the number of nodes.This confirms that the proposed framework scales much better than the single sensor per parking spotbased solutions.

VII. CONCLUSION

The recent advances in Radio Frequency Identification (RFID), Wireless Sensor Networks (WSNs), andhandheld/mobile computing have led to a new information technology (IT) era where devices built aroundthese technologies are deployed in our daily life environments, to provide application over the internetand form the internet of things. Amongst the variety of enabled applications, smart car park managementhas been considered in this paper, where the integration of the aforementioned technologies is expected tooffer invaluable services that undutifully will help reduce traffic jam in modern cities and it consequences.

CPF, a novel smart car park framework that uses the integration of wireless sensor network (WSN),radio Frequency identification (RFID), mobile computing technologies has been proposed. The frameworkis scalable and provides advanced services. It is a modular solution that uses multiple patterns, whichdivides a parking to levels, a level to bunches of lots, and a bunch of lots to lots. Smooth integrationof RFID and WSN, along with the sensor clustering with a single mote per bunch of spots, representthe most relevant features and contribution of the proposed CPF. A prototype implementation with realequipments has been presented and used for some quantitative empirical evaluation, where appropriateapplications for mobile devices have been developed. The results show that the clustering of a bunchof spots with a single mote considerably reduces the cost, as well as the energy consumption. The gainbecomes more important as the number of clustered spots increases, which confirms the scalability.

The prototype proposed in this paper has been tested in outdoor setting using parking lot of ourcampus in CERIST, Algiers. The experimental testing of our prototype in indoor settings is an avenuefor future work. The challenges associated with such experimentation might require the use of QoSrouting approaches such as proposed in [23], and complex routing architectures where networks of hybridsensor/RFID devices are layered above networks of hybrid wired/wireless network.

REFERENCES

[1] W.-C. Hu, Y. Zuo, Y. Zuo, and H.-J. Yang, “Contemporary issues in handheld computing research,” International Journal of HandheldComputing Research, vol. 1, no. 1, pp. 1–23, 2010.

[2] L. M. Borges, F. J. Velez, and A. Lebres, “Survey on the characterization and classification of wireless sensor network applications,”IEEE Communications Surveys and Tutorials, vol. 16, no. 4, pp. 1860–1890, December 2014.

[3] D. Delen, R. Sharda, and B. C. Hardgrave, “The promise of rfid-based sensors in the perishables supply chain,” IEEE Wireless Commun.,vol. 18, no. 2, pp. 82–88, 2011.

[4] A. Bagula, D. Djenouri, and E. Karbab, “Ubiquitous sensor network management: The least interference beaconing model,” in IEEEPIMRC, 2013, Sept 2013, pp. 2352–2356.

[5] Z. Y. C. J. Tang, V W. S., “An intelligent car park management system based on wireless sensor networks,” in 2006 1st InternationalSymposium on Pervasive Computing and Applications, Proceedings, Callaghan, V and Hu, B and Lin, Z and Zhang, H, Ed. IEEE,2006.

[6] Z. H.-S. Y. T.-X. L. Z.-J. BI Yan-Zhong, SUN Li-Min, “A parking management system based on wireless sensor network,” ACTAAUTOMATICA SINICA, vol. Vol. 32, No. 6, 2006.

[7] P. K. et al, “A prototype parking system using wireless sensor networks,” International Journal of Computer Communication andInformation System ( IJCCIS), vol. Vol 2. No 1., 2010.

[8] B. J. R. et al, “Car-park management using wireless sensor networks,” in 31st IEEE LCN, Elmallah, E and Christensen, K and Frank,M, Ed., 2006, pp. 588–595.

[9] L. C.-L. J Ming-Shen, Y K Shiuh, “Modular rfid parking management system based on existed gate system integration,” WTOS, vol. 7,no. 6, Jun. 2008.

[10] I. N. P Zeydin, “Smart parking applications using RFID technology,” in Proceedings of the 1st RFID Eurasia Conference, Ozok, AFand Ustundag, A, Ed., 2007, pp. 121–123.

[11] Z. H. S. X. L Rongxing, L Xiaodong, “SPARK: A New VANET-based Smart Parking Scheme for Large Parking Lots,” in IEEEINFOCOM, 2009.

[12] B. B. Z. A. A. Z. A A Yass, N M Yasin, “New design for intelligent parking system using the principles of management informationsystem and image detection system,” International Conference on Computer Engineering and Applications, vol. vol.2, 2011.

[13] K. T. D.B.L. Bong and K. Lai, “Integrated approach in the design of car park occupancy information system (coins),” IAENGInternational Journal of Computer Science, 2008.

[14] N. N. Z. R. K. F. M.Y.I. Idris, E.M. Tamil, “Parking guidance system utilizing wireless sensor network and ultrasonic sensor,” InformationTechnology Journal, 2009.

[15] S. H. S Shim, S Park, “Parking management system using zigbee,” IJCSNS International Journal of Computer Science and NetworkSecurity, vol. 6, no. 9B, 2006.

[16] KATHODIGOS - A Novel Smart Parking System based on Wireless Sensor Networks. EDIS - Publishing Institution of the Universityof Zilina, 2013.

[17] D. S. . P. R. Chinrungrueng, J., “iparking: A parking management framework,” Wireless Commun., pp. 63 – 68, 2011.[18] J. Gu, Z. Zhang, F. Yu, and Q. Liu, “Design and implementation of a street parking system using wireless sensor networks.” in

INDIN. IEEE, 2012, pp. 1212–1217. [Online]. Available: http://dblp.uni-trier.de/db/conf/indin/indin2012.html[19] S. Nawaz, C. Efstratiou, and C. Mascolo, “Parksense: A smartphone based sensing system for on-street parking,” in ACM MobiCom,

2013, pp. 75–86.[20] A. Bagula, D. Djenouri, and E. Karbab, “On the relevance of using interference and service differentiation routing in the internet-of-

things,” in Internet of Things, Smart Spaces, and Next Generation Networking, ser. Lecture Notes in Computer Science. SpringerBerlin Heidelberg, 2013, pp. 25–35.

[21] D. Djenouri, “R4syn : Relative referenceless receiver/receiver time synchronization in wireless sensor networks,” IEEE Signal Process.Lett., vol. 19, no. 4, pp. 175–178, 2012.

[22] D. Djenouri, N. Merabtine, F. Z. Mekahlia, and M. Doudou, “Fast distributed multi-hop relative time synchronization protocol andestimators for wireless sensor networks,” Ad Hoc Networks, vol. 11, no. 8, pp. 2329–2344, 2013.

[23] D. Djenouri and I. Balasingham, “Traffic-differentiation-based modular qos localized routing for wireless sensor networks,” IEEE Tran.on Mob. Comp., vol. 10, no. 6, pp. 797–809, 2011.