Design Methodology of Energy Consumption for ...

3 downloads 2202 Views 735KB Size Report
energy harvesting schemes such as solar panels [2], but the amount of harvested energy is ... Dr. J. Laassiri is with the University Ibn Tofeil, Faculty of Sciences.
JOURNAL OF COMPUTING, VOLUME 4, ISSUE 5, MAY 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

267

Design Methodology of Energy Consumption for Wireless Sensor Networks using Renewable Energy KRIT SALAH-DDINE, LAASSIRI JALAL, EL HAJJI SAID Abstract—In recent years, the number of wireless sensor networks (WSNs) deployments for real life applications has rapidly increased. Still, the energy problem remains one of the major barrier somehow preventing the complete exploitation of this technology. WSN uses a software technique to monitor device usage patterns, and combines it with hardware power information in runtime. Hardware systems have been recently proposed to increase the autonomy of embedded systems. WSNs are typically powered by batteries with a limited lifetime and, even when additional energy can be harvested from the external environment (e.g., through solar cells or photovoltaic mechanisms), it remains a limited resource to be consumed judiciously. Efficient energy management is thus a key requirement for a credible design of a wireless sensor networks. One key design challenge is how to optimize the efficiency of WSNs. This paper proposes a new design architecture and implementations of power management dc-dc buck-boost converter targeted for optimization of consumption of wireless sensor network applications. Key issues in WSNs are discussed. We operated real applications on the WSNs, The proposed system can operate at 2.4GHz and dissipates a power of 15 and 25 uW, respectively, from a 2.5 V supply. The noise performance at input, output and the frequency response is presented. Index Terms—Wireless Sensor Network (WSNs), Software, Hardware, Embedded Systems, Power Management, dc-dc buck-boost converter.

——————————

—————————— essential functionality required for any practical system [6]. Various approaches are found for runtime and off1 INTRODUCTION line energy monitoring of sensor nodes. Dunkels [7] proposed a runtime software-based energy estimation technique and Landsiedel [8] proposed off-line ower management is an important issue in Wireless techniques. These approaches may perform incorrect power profiling due to unstable power models and Sensor Networks (WSNs) [1] since micro sensor nodes inaccurate pre-measured power data for individual are typically powered by energy-limited batteries. devices. To overcome the limitations of software-based Although sensor network technology has recently been monitoring, hardware-assisted power monitoring applied to various applications, battery lifetime is still a techniques were proposed. Trathnigg [9] and Jiang [10] major limiting factor for the widespread use of sensor suggested power measuring hardware that should be applications. Alternatively, researchers have studied connected to sensor node to collect data. An on-board energy harvesting schemes such as solar panels [2], but hardware-based power monitoring technique was also the amount of harvested energy is limited and the system proposed. Stathopoulos [11] suggested EndScope using has similar problems to those found in batteries. Power the current monitor IC [12]. EndScope provides real-time management to extend battery lifetime, therefore, energy usage information for individual devices but the becomes more important and active research has been system is not suitable for resource-constrained sensor done in recent years. nodes. Fonseca [13] proposed Quanto, which traces the An appropriate power management technique for sensor energy consumption of a device using iCount [14] that is nodes is to detect an idle device operation and change it combined with a software technique. iCount uses a PFM to reduce unnecessary energy consumption [3]. The (Pulse Frequency Modulation)-type DC–DC switching energy consumption of a sensor node is affected by node regulator and timer on a microcontroller, which counts type, deployed environment, and application [4]. In the switch cycle of the regulator. For continuous energy addition, batteries have chemical characteristics that monitoring, however, iCount uses a system timer on the affect battery lifetime of sensor nodes [5]. Runtime energy microcontroller, which would affect the application monitoring is, therefore, an important issue for the performance. Recent advances in CMOS technology have efficient power management of sensor nodes, and an enabled the development of a micro-sized battery monitor IC [15]. The hardware has a small form factor ———————————————— and consumes little power; hence, it can readily be • Dr. S. Krit Author is with the University Ibn Zohr, Polydisciplinary Faculty integrated into a sensor node. Using a battery-monitoring of Ouarzazate, Morocco • Dr. J. Laassiri is with the University Ibn Tofeil, Faculty of Sciences IC is a common technique to manage batteries in typical Deppartement of Informatics kenetra, Morocco. embedded devices. In Wireless Sensor Networks, the • S. El Hajji is with the University Mohamed V, Faculty of Sciences conventional belief is that using a hardware IC would

P

Deppartement of Mathematics Informatics, Rabat, Morocco

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 5, MAY 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

involve additional cost, power, and access time overhead. However, the recent widespread use of batteries in micro sensor nodes will accelerate the use of a battery monitor IC [16]. In this paper, we describe hardware architecture for micro sensor nodes. Our primary objective is to provide an energy monitoring architecture that monitors power consumption of resource-constrained sensor nodes in real-time. In particular, dc-dc buck-boost converter is designed to provide a method for monitoring individual devices in sensor nodes. Our system provides a software technique to handle hardware-assisted power monitoring information. For a practical use of the system, we use a battery-monitoring IC, which is commonly used in mobile embedded systems. The software technique provides a power trace of individual hardware components in a sensor node by using incomplete power information from the battery monitoring IC. The remainder of the paper is organized as follows: In Section 3, we describe the architecture level optimization chose a battery monitor IC as a hardware assistant and discuss its limitations. We present the system design of sensor nodes in Section 4. In Section 5, we show the applications of WSNs in a real environment. In section 6 we present the circuit overview. The evaluation results are shows in Section 7. We conclude the paper in Section 8.

2 SOURCE OF ENERGY CONSUMPTION IN WIRELESS SENSOR NETWORKS In order to design a low power wireless sensor network, first step is to analyze the power dissipation characteristics of wireless senor node. Each node in the network is consists of four components: a sensor which connects the network to physical world, computation part which is consists of microcontroller or in some application microprocessor and is responsible for control of the sensors and communication, a transceiver for communicating between nodes and base station, and a power supply which is usually a battery. There are wide ranges of choices for each part of the node and choosing a right device will affect the energy consumption.

268

Despite the energy efficiency of specific hardware platforms, Sensor network lifetime can be significantly enhanced if the software of the system, including different layers and protocols are designed in a way that lower the consumption of energy [2]. In the following parts, at first the most important points in choosing the hardware for the node will be discussed and then we will go through several techniques which affect different aspects of WSN software, in order to have low power network.

3

ARCHITECTURE-LEVEL PTIMIZATIONS

Fig. 2 shows an integrated WSN architecture (i.e., a WSN integrated with external networks) capturing architecture-level optimizations. Sensor nodes are distributed in a sensor field to observe a phenomenon of interest (i.e., environment, vehicle, object, etc.). Sensor nodes in the sensor field form an ad hoc wireless network and transmit the sensed information (data or statistics) gathered via attached sensors about the observed phenomenon to a base station or sink node. The sink node relays the collected data to the remote requester (user) via an arbitrary computer communication network such as a gateway and associated communication network. Since different applications require different communication network infrastructures to efficiently transfer sensed data, WSN designers can optimize the communication architecture by determining the appropriate topology (number and distribution of sensors within the WSN) and communication infrastructure (e.g., gateway nodes) to meet the application’s requirements. An infrastructure-level optimization called bridging facilitates the transfer of sensed data to remote requesters residing at different locations by connecting the WSN to external networks such as Internet, cellular, and satellite networks. Bridging can be accomplished by overlaying a sensor network with portions of the IP network where gateway nodes encapsulate sensor node packets with transmission control protocol or user datagram protocol/internet protocol (TCP/IP or UDP/IP).

Fig. 1 Overview of sensor node hardware components Fig. 2. Wireless sensor network architecture.

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 5, MAY 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

269

5 Since sensor nodes can be integrated with the Internet via bridging, this WSN-Internet integration can be exploited to form a sensor web. In a sensor web, sensor nodes form a web view where data repositories, sensors, and image devices are discoverable, accessible, and controllable via the World Wide Web (WWW). The sensor web can use service-oriented architectures (SoAs) or sensor web enablement (SWE) standards [17]. SoAs leverage extensible markup language (XML) and simple object access protocol (SOAP) standards to describe, discover, and invoke services from heterogeneous platforms. SWE is defined by the OpenGIS Consortium (OGC) and consists of specifications describing sensor data collection and web notification services. An example application for a sensor web may consist of a client using WSN information via sensor web queries. The client receives responses either from real-time sensors registered in the sensor web or from existing data in the sensor data base repository. In this application, clients can use WSN services without knowledge of the actual sensor nodes’ locations. Another WSN architectural optimization is tunneling. Tunneling connects two WSNs by passing internetwork communication through a gateway node that acts as a WSN extension and connects to an intermediate IP network. Tunneling enables construction of large virtual WSNs using smaller WSNs [18].

4

SENSOR NODE OPTIMIZATIONS

COMPONENT-LEVEL

COTS sensor nodes provide optimization opportunities at the component-level via tunable parameters (e.g., processor voltage and frequency, sensing frequency, duty cycle, etc.), whose values can be specialized to meet varying application requirements. Fig. 3 depicts a sensor node’s main components such as a power unit, storage unit, sensing unit, processing unit, and transceiver unit along with potential tunable parameters associatedwith each component [18]. In this section, we discuss these components and associated tunable parameters.

WSN APPLICATIONS

The tasks for WSN applications depend mainly on the sensors and actuators available in nodes, since the innetwork processing and communication capabilities allow a rich set of application functionality. Currently, the actuators are limited mostly to servo drives and different types of switches, while the physical quantities that can be measured with already existing technologies are diverse. These include for example temperature, humidity, pressure, acceleration (vibration), sound, light (luminance, image), magnetic fields (compass), location, chemical compositions, and mechanical stress [19, 20, 21]. The typical WSN application in the literature are: Industrial monitoring and control: the replacement of traditional cabling in machine surveillance and maintenance systems is the main application for WSNs. Further, WSNs can be benefited for managing logistics [19, 20]. Military: military applications were the driver for the first WSN research initiatives [21, 22,23]. While the scope has expanded rapidly, military is still one of the main application areas. Usage scenarios cover for example intelligence, surveillance, reconnaissance and targeting. Personal security and asset management: WSNs are envisioned to replace or extend existing wired alarm systems in home, office, and other public environments such as airport and factories. Compared to wired systems, WSNs allow faster deployment, more flexibility, and larger area coverage. Traffic control: WSNs make it possible to extend the traffic monitoring and control systems farther away from the most critical points. Temporary situations such as roadworks and accidents can be covered in place. A far reaching vision embeds a WSN node to every vehicle. These nodes share the traffic information, generate warning of accidents and jams, and guide drivers in route selections. Health care: Biomedical sensor networks can be used to gather physiological data directly from patients. Other health care related application areas are drug administration, and tracking of doctors and patients in hospital premises.

6

CIRCUIT OVERVIEW

Scavenging power via solar panels proved to be a difficult task. For example, an inexpensive thin film solar panel produced about 2.5 mW/cm2 in full sun, but only about 15 µW/cm2 indoors under flourescent lighting. Also, its operating voltage dropped from 3 V outdoors to around 1 V indoors. Thus, any usable power circuitry must efficiently store energy with a wide range of input voltages and currents. A conceptual I–V curve for a solar cell is shown in Figure 4.

Fig. 3. Sensor node architecture with tunable parameters.

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 5, MAY 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

Fig. 4 Conceptual solar cell I–V curve.

At some point on the curve with voltage Vm and current Im, the solar cell produces maximum power Pm. Ohm’s law gives the load resistance necessary to operate at this maximum power point. Circuits driven by the solar cell should present this resistance to the solar cell so that power is not wasted. The initial power circuit design was suggested by a solar battery charge circuit [21]. Figure 5 shows a simplified version. This circuit uses the capacitor to keep the solar panel operating at near its maximum power point for high efficiency. A comparator

Fig. 5 Solar dc-dc converter circuit.

constantly monitors the capacitor voltage, comparing it to a reference voltage Vref . When the capacitor voltage reaches Vref , the comparator enables a DC-DC boost converter that produces the battery charging voltage. Due to hysteresis, the comparator turns off the boost converter when the capacitor voltage has dropped to some minimum voltage. Additional circuitry not shown in the figure monitors the battery voltage and adjusts the DC-DC converter output voltage for proper charging. The comparator relies on the battery to provide its supply voltage. Since most comparators do not guarantee proper operation with an inadequate supply voltage, the circuit may have strange behavior if the battery’s voltage drops too low. For example, if the DCDC converter can operate at a lower voltage than the comparator, it might erroneously turn on before the comparator is producing a valid output. Therefore, a battery, at least partially charged, is an essential component of the circuit. If the battery were allowed to discharge to a voltage below the comparator’s minimum operating voltage, the circuit would not necessarily be able to recharge it.

270

The stochastic sensor node( SSN) power supply circuit is conceptually similar to this battery charger. In the SSN, a comparator enables the DC-DC converter to power the logic when the reservoir capacitor reaches a voltage near Vm, the maximum power voltage. Instead of using a rechargeable battery, however, the SSN uses a non rechargeable lithium coin cell, which has much higher energy density. Also, instead of using a discrete comparator, the SSN monitors the capacitor voltage in software using the microcontroller’s analog-to-digital converter (ADC) and a programmable threshold. Because the SSN has full control over

Fig. 6 Power supply circuit schematic.

the power supply in software, more exotic power management schemes than simple thresholding could be implemented without changing the hardware. The addition of a non rechargeable battery does limit the operational lifetime of the SSN. Because the circuit was designed for low power consumption, however, a standard cell can easily power the circuit in its inactive mode for several years, Figure 6 shows the SSN power supply circuit. The buck boost converter is DCDCBUCK_1.

7

SPICE COMPATIBLE BUCK BOOST AND IMPLEMENTATION

MODEL

Figure 6 shows the SSN power supply circuit. The buck boost converter is DCDCBUCK_1, a Maxim buck-boost. It creates a higher voltage than its input by switching current on and off through inductor L1, then filtering the resulting voltage spikes. When used as configured for the SSN, it boasts efficiency near 90%. The inductor L1 is a PTAVAL=10uH, chosen primarily for its small size, and the capacitor C1 is a Metal oxide C2=22uF, chosen for its low series resistance R2=100mohm. This specific part was selected mainly for its low forward voltage drop. Connection of the boost converter’s feedback (FB) pin to the output voltage pin selects its 3.3V output mode. When in shutdown mode, which the microcontroller selects through U3’s SHDN pin, the DCDC converter’s output voltage is slightly below the input voltage applied to the inductor connection (LX) pin. Also, the boost converter has no overvoltage regulation. When not in shutdown mode, if its input voltage is higher than 3.3 V, the output voltage will follow the input voltage. Using this scheme, the circuit can run from the battery in

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 5, MAY 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

271

inactive mode, then enable the boost converter during its active cycle.

8

EXPERIMENTAL RESULTS

Figure 7 presents the graph that is composed of tree curves, one dedicated to the battery charge vdd period (during 200us/div up to 5V) its discharge (during 1ms down to 2.1 voltage level) and the second is the current load IL to using pulses generated by the SystemC framework simulating the work of a single node with the 100% duty cycle, the third present the output voltage Vout. The battery discharging is quite smooth at 3.1 to 2.1 voltage range. However, starting from 2.1 V the discharging curve slumps To carry out the energy consumption experiments we were charging the battery model during 20000 seconds up to 5 V (we imply two cylindrical type NiMH batteries connected in series). The battery model capacity was defined as 100mAh. The input discharge pulses for the battery were generated by sensor node state simulator presented in this paper. Figure 8 shows the switching on/off output of the two drivers of the dcdc buck-boost converter. Figure 9 illustrate the Simulation results of PWM DCDC buckboost converter with different current load and different input voltage. Figure 7 presents the battery run time under specified duty cycle. Moreover, in comparison to both of 50% duty cycle experiments this mode is more energy efficient in terms of utilization of battery charge unit per second. Besides, two various experiments with 50% battery discharge re-veal that short time pulses lead to the battery life time increase. This fact can be explained by the relaxation effect of the battery. The experimental results presented in Figure 9 give a general idea how the discharge pulses rate has an influence on the system long-term operation.

Fig.7 batter charge, Iload and Vout

fig. 8 drivers output voltage

Fig. 9 Simulation results of PWM DCDC buck-boost converter with different current load and different input voltage.

9

CONCLUSION

We have presented a SystemC-based methodology for power consumption evaluation of an individual node of a wireless sensor network that supports the PBD paradigm. We first gave a general overview of the methodology. Then we have applied the methodology to perform power consumption estimation of a case study application running on a OMNET++ mote in terms of a battery lifetime. We show that with our framework it is possible to capture the influence of the relaxation effect to the battery lifetime by increasing the performances of the dc-dc buck-boost converter. As for our future work we plan to include simulation of heterogeneous networks by extending the presented methodology with an additional network layer and provide ability to model and simulate a network of nodes. Furthermore, a simpler (linear) customizable battery model will be implemented/added as a part of the framework. It will support quicker analysis/evaluation of power consumption and lifetime estimation. The other interesting area is the support of already available WSN frameworks and OSs. Our main focus is on OMNET++due to its popularity in the WSN community. Currently we are investigating ways to link OMNET++ applications with our service layer.

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 5, MAY 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

REFERENCES [1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12] [13]

[14]

[15]

[16]

D. Estrin, A. Sayeed, M. Srivastava, Wireless Sensor Networks, in: Proceedings of Eighth International Conference on Mobile Computing and Networking (MobiCom02 Tutorial), USA, September 2002. A. Kansal, J. Hsu, S. Zahedi, M.B. Srivastava, Power management in energy harvesting sensor networks, ACM Transactions on Embedded Computing Systems 6 (4) (2007). A. Sinha, A. Chandrakasan, Dynamic power management in wireless sensor networks, IEEE Design & Test of Computers 18 (2) (2001) 62–74. P. Dutta, J. Hui, J. Jeong, S. Kim, C. Sharp, J. Taneja, G. Tolle, K. Whitehouse, D. Culler, Trio: Enabling Sustainable and Scalable Outdoor Wireless Sensor Network Deployments, in: The fifth International Conference on Information Processing in Sensor Networks (IPSN 2006), Nashville, Tennessee, USA, April 2007. S. Choi, H. Cha, R. Ha, A Selective DVS Technique based on Battery Residual, Microprocessors and Microsystems, Elsevier Science 30 (1) (2006) 33–42. S. Choi, B. Song, H. Cha, R. Ha, Energy-Aware Pipeline Monitoring System Using Piezoelectric Sensor, IEEE Sensors (in press). A. Dunkels, F. Österlind, N. Tsiftes, Z. He., Software-based online energy estimation for sensor nodes, in: Proceedings of the Fourth Workshop on Embedded Networked Sensors (Emnets IV), Cork, Ireland, June 2007. O. Landsiedel, K. Wehrle, S. Gotz, Accurate Prediction of Power Consumption in Sensor Networks, in: The Second IEEE Workshop on Embedded Networked Sensors, Sydney, Australia, May 2005. T. Trathnigg, R. Weiss, A Runtime Energy Monitoring System for Wireless Sensor Networks, in: Proceedings of the 3rd IEEE International Symposium on Wireless Pervasive Computing, Santorin, Greece, May 2008. X. Jiang, P. Dutta, D. Culler, I. Stoica, Micro Power Meter for Energy Monitoring of Wireless Sensor Networks at Scale, in: Proceedings of the Sixth International Conference on Information Processing in Sensor Networks(IPSN), Cambridge, Massachusetts, April 2007. T. Stathopoulos, D. McIntire, W. Kaiser, The Energy Endoscope: Real-time Detailed Energy Accounting for Wireless Sensor Nodes, in: Proceedings of the Seventh International Conference on Information Processing in Sensor Networks(IPSN), St. Louis, Missouri, USA, April 2008. Ti INA219 current monitoring IC, http://focus.ti.com/docs/prod/folders/print/ina219.html. R. Fonseca, P. Dutta, P. Levis, I. Stoica, Quanto: Tracking Energy in Networked Embedded Systems, in: Proceedings of the Eighth USENIX Symposium on Operating System Design and Implementation (OSDI’08), San Diego, CA, USA, 2008. P. Dutta, M. Feldmeier, J. Paradiso, D. Culler, Energy Metering for Free: Augmenting Switching Regulators for Real-Time Monitoring, in: Proceedings of the Seventh International Conference on Information Processing in Sensor Networks(IPSN), St. Louis, Missouri, USA, April 2008. DS2782 battery monitor IC, . J. Rahmé, K. Agha, A state-based battery model for nodes’

[17]

[18] [19]

[20]

[21] [22]

[23]

272

lifetime estimation in wireless sensor networks, in: Proceedings of the tenth ACM international symposium on Mobile ad hoc networking and computing, New Orleans, USA, 2009. Mahalik, N. (2007). Sensor Networks and Configuration: Fundamentals, Standards, Platforms, and Applications, Springer. Karl, H. &Willig, A. (2005). Protocols and architectures for Wireless Sensor Networks, JohnWiley and Sons, Inc. Akyildiz, I. F., Melodia, T., and Chowdhury, K., "A Survey on Wireless Multimedia Sensor Networks," Computer Networks (Elsevier) Journal, March 2007. D. Estrin, R. Govindan, J. Heidemann, and S. Kumar, “Next centry challenges: Scalable coordination in sensor networks” in Proc. 5th annual ACM/IEEE International Conference on Mobile Computing and Networking (Mobicom’99), Seattle, WA, USA, August 15-20 1999, pp. 263-270. H. Karl and A. Willig, Protocols and Architectures for Wireless Sensor Networks. John Wiley & Sons Ltd, 2005. Hai-Ying Zhou, Dan-Yan Luo, Yan Gao, De-Cheng Zuo, Modeling of Node Energy Consumption for Wireless Sensor Networks, Wireless Sensor Network, 2011, 3, 18-23. Krit Salah-ddine, Jalal Laassiri and El Hajji Said, " Power Aware Simulation Framework for Spice Compatible Battery Model of Wireless Sensor Networks", Journal Of Computing, Volume 4, Issue 4, April 2012.

Salah-ddine Krit received the B.S. and Ph.D degrees in Microectronics Engineering from Fes Sidi Mohammed Ben Abdellah university, Fez, Morroco. Institute in 2004 and 2009, respectively. During 2002-2008, he is also an engineer Team leader in audio and power management Integrated Circuits (ICs) Research. Design, simulation and layout of analog and digital blocks dedicated for wireless sensor networks (WSN) and satellite communication systems using CMOS technology. He is Member of the International Association of Engineers (IAENG) and European Alliance for Innovation (EAI), currently He is a professor coordinator of professional license of informatics with Polydisciplinary Faculty of Ouarzazate, Ibn Zohr university, Agadir, Morroco. His research interests include wireless sensor Networks (Software and Hardware), microtechnology and nanotechnology for wireless communications.

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 5, MAY 2012, ISSN 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

Jalal Laassiri received his Bachelor’s degree (License es Sciences) in Mathematics and Informatics in 2001 and his Master’s degree (DESA) in computer sciences and engineering from the faculty of sciences, university Mohammed V, Rabat, Morocco, in 2005, and he developed He received his Ph.D. degree in computer sciences and engineering from University of Mohammed V, Rabat, Morocco, in Juin, 2010. He was a visiting scientific with the Imperial College London, in London, U.K. He is Member of the International Association of Engineers (IAENG), He joined the Faculty of Sciences of Kénitra, Department of Computer Sciences , Ibn Tofail University, Morocco, as an Professor in October 2010, His current research interests include Software and Systems Engineering, UML-OCL, BMethod, ..

273

Said El Hajji, Professor of Higher Education at Mohammed V - Agdal University, chief of Laboratory MIA, Faculty of Sciences, Rabat, Morocco. http://www.fsr.ac.ma/mia/elhajji.htm.