Transactions Template JOURNAL OF ENGINEERING RESEARCH AND TECHNOLOGY, VOLUME 2, ISSUE 3, SEPTEMBER 2015 175 MEMS Based Energy Harvesting Controler Using Fuzzy Logic Basit Ali 1 , Muhammad Waseem Ashraf 1 , Shahzadi Tayyaba, Muhammad Faisal Wasim 1 1 GC University Lahore, Pakistan 2 The University of Lahore, Pakistan *Email: Muhammad.Waseem.Ashraf@gmail.com Abstract—This paper presents the design and simulation of MEMS based energy harvester controller. Energy harvesting from physical motion such as finger motion, heart beating, walking and running are becoming so important now-a-days. In this study outputs voltage and current of the MEMS based energy harvester can be controlled by using fuzzy logic based control system. A new way of controlling the outputs by changing the inputs has been proposed in this study. The system has been developed in MATLAB using fuzzy logic Mamdani model.Two inputs pressure and area have been selected. The three membership functions to the input parameters have been assigned. The two outputs like voltage and current are selected for output power. Three membership functions are also assigned to the outputs. The system works according to the rules defined in the fuzzy inference system (FIS). The results according to suggested rules belonging to assigned MFs have been displayed in surface viewer.Different rules of combinations were defined in MATLAB rule editor and we used AND logic for simula- tion. The rsults obtained from fuzzy logic controller have been verified by using Mamdani’s formula for specific values of the inputs and the outputs. As there is only 1% error for both the outputs, current and voltage, this shows that the system performs very well. Index Terms — Fuzzy logic, Fuzzy rule editor, Fuzzy Inference system, MEMS, Membership functions, MATLAB. I Introduction Vibrations are one of the most common phenomena that exist at all the time and at everywhere and in every field of life.Some common and simple natural acts occurring in nature that can produce vibrations are figure motion ,heart beating, walking, running and any mechanical instrument in action. The mechanical actions like running cars or any other function of the human body is the best source to produce vibrations which can easily be converted into electrical energy through MEMS based energy harvesters. This ambient energy would be useful for wearable devices, household applications, sensors for medical implants, compute or communicate practically everywhere.Very low power 10-100μW is requied for the normal operation of VLSI design and lithium-ion batteries can produce a power 160W.h/kg but it is very large in size and has a limited life [1]. The MEMS based energy harvesting devices approch in different fields eliminating the need for wiring, chemical batteries or power sources which are bulkier and higher in cost. Therefore, MEMS based energy harvester becomes more appealing or even essential. MEMS based energy harvesters are so good that we can produce a power of 1µW only by using a device of dimensions 1cm×1cm and this power can easily be converted into 1mW if it is given to a little capacitor. [2]. Lin et al. proposed an intelligent fuzzy logic system whose basis was artificial neural network. It was seen that by linking the self-organized and supervised system the results taken were astonishing. The system considered was user friendly and could be easily understood by anyone. The system was seen to be benevolent, solid and provide the best results [3]. Castro presented a work on the topic how does fuzzy logic is an approximate method and why it has an edge over other logics. He also described why does fuzzy logic shows great performance while other logics cannot do so. People mostly criticize the performance of fuzzy controller but in his work he proved that the fuzzy controller had great impact on daily life problems.Castro had shown both quantitative and qualitative approach to the fundamental problems[4].Tang et al. proposed a fuzzy logic based system for creating genetic algorithm. He used a method in which a system was divided into two different inputs. Fuzzy membership functions were used to define the parameters of inputs. Fuzzification was done on the input and rules were defined, then defuzzification was done on the outputs. Fuzzy logic based system was used due to its unique way of attempting complex and non-linear problems [5]. Sue et al. [6] (2011) particularly investigate and characterize the energy harvesters that can be used to produce power from human body. They also reviwed the currently available MEMS based energy harvester. They evaluate and briefly described power gain, different methods for tunning the frequency and showed that micro-energy harvesters are biologically harmless. Liu et al. [7] (2012) fabricated a new piezoelectric cantilever at microlevel for harvesting vibrational energy at very small frequencies and low accelerations. They obtained a maximum potential difference 42 mV and a power of 0.31 μW g −2 by acceleration of 0.06g. Elizabet et al. [8] (2012) studied the enhancement of the mechanical response of MEMS based energy harvester by optical excitation. In this way they http://link.springer.com/search?facet-author=%22Huicong+Liu%22 Basit Ali, et.al. / MEMS Based Energy Harvesting Controler Using Fuzzy Logic (2015) 176 gave a pathway for the moving structures to response with heating effect. They showed how a MEMS based connected structure can responds mechanically to the temperature change of the element. The heating purpose they used infrared radiations. Liu et al. [9] (2012) presented a piezoelectric energy harvester (PEH) system with a large operating bandwidth. They showed that their device could produce an output power of 34 to 100 nW. Dhakar et al. [10] (2014) presented a triboelectric energy harvesting devices. The maximum output power measured from the device was observed to be 0.69 μW. Jia et al. [11] (2014) reported a piezoelectric MEMS cantilever vibrational energy harvester. In this way they were able to produce 0.7μW with 7g and 2.56 μW at 3 ms -2 . Cao et al.[12] (2011) designed a piezoelectric cantilever. They used finite element method (FEM) for simulation. They verified their results and showed that the optimized cantilevered piezoelectric energy harvesters could produce a 56V peak open-circuit voltage. The proposed method would be suitable for optimization design of piezoelectric energy harvester. Bala et al. [13] (2014) described an electrode position optimization in magnetoelectric sensors based on piezoelectric bilayer cantilever substrates. They applied the Finite element method (FEM) for simulations.A 15% higher signal voltage across the piezoelectric layer was obtained for optimally positioned electrodes with a simple layered cantilever and an insulating magnetostrictive material. They also described that the signal voltage was increased 25% for a trenched cantilever. Kellogg et al.[14] (2011) studied piezoelectric energy harvester. They said that by increasing the length of the cantilever the stress level of the cantilever increased and in this way power output of each piezoelectric element increased. Leadenham et al. (2014) [15] described a piezoelectric cantilever for sensing, actuation and energy harvesting. They found the the proposed model and experimental investigation were in close agreement with each other. Mutlaif et al.[16] (2015) presented a mathematical derivations for piezoelectric energy harvester.They used MATLAB and COMSOL Multiphysics software for Simulation.They also studied the the effect of length and shape of the cantilever beam on the output voltage. Rivadeneyra et al. [17] (2015) reported a low frequency <300 Hz vibrational energy harvester due to the fact that many industrial and commercial devices operate at these frequencies. In their paper they investigate the influence of perforating sections of the Si beam had on the resonant frequencies of the cantilever by numerical simulation. Kim et al.[18] (2013) fabricated dual-beam cantilevers on the microelectromechanical system (MEMS) scale with an integrated Si proof mass. They used the finite element method (FEM) with parametric analysis carried out in the design process. According to simulations, the resonant frequency, voltage, and average power of a dual- beam cantilever was 69.1 Hz, 113.9 mV, and 0.303µW, respectively, at optimal resistance and 0.5 g. The harvested power density of the dual-beam cantilever compared favorably with the simulation. Their experimental results for the resonant frequency, voltage, and average power density were 78.7 Hz, 118.5mV, and 0.34 µW. The error between the measured and simulated results was about 10%. The maximum average power and power density of the fabricated dual-beam cantilever at 1 g were 0.803µW and 1322.80 µW cm −3 , respectively. Fuzzy logic is basically a flexible technique and is a numerical representation of system in which answer is just not only high or low, 0 or 1, ON or OFF and True or False. It is a free technique which is not bounded by any specific states. For example in thermally heated metal where value is not just only hot or cold but also between them, this system could be easily developed by using fuzzy logic as it can tell that some part of the metal is at normal temperature. The most common way of using fuzzy logic is to solve it through MATLAB software. In this paper we have done the simulation for an efficient MEMS based energy harvester that can convert mechanical energy into electricity. The rsults obtained from fuzzy logic controller have been verified by using Mamdani’s formula for specific values of the inputs and the outputs. As there is only 1% error for both the outputs current and voltage, this shows that the system performs very well.On the analysis of the results obtained from simulation we will draw the conclusion. II DESIGN METHODOLOGY A Designing in MATLAB FLC is comprises of two inputs with three membership functions and two outputs also with three membership functions. Fig. 1 shows two input variables: Pressure, Area and Voltage and Current as outputs. Figure 1 Fuzzy Logic Controller (FLC) FLC system in Fuzzy Logic Inferring System (FIS) editor could be assigned by several numbers of inputs but here it has two inputs and each input has three membership functions (MFs). The ranges should be selected according to the desired values of input (MFs) and output (MFs). The ranges of inputs and output have been taken (0-100) for both inputs and outputs as shown in table1 Table 1 Ranges selected for inputs and outputs http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://www.ncbi.nlm.nih.gov/pubmed/24569244 http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://www.ncbi.nlm.nih.gov/pubmed/24569244 http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://www.ncbi.nlm.nih.gov/pubmed/24569244 http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://adsabs.harvard.edu/abs/2014SPIE.9057E..02L http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://adsabs.harvard.edu/abs/2014SPIE.9057E..02L Basit Ali, et.al. / MEMS Based Energy Harvesting Controler Using Fuzzy Logic (2015) 177 The figure 2 shows that how the common regions have been differentiated. The first overlapped region of the range 0 to 50 and 0 to 100 is called Region 1 and second overlapped region of range 0 to 100 and 50 to 100 is called Region 2. It is same for the inputs and the outputs. The calculations have been taken according to this regional division . Figure 2 Division of the regions In designing of this system different rules have been estab- lished for the better result. The rules involve the simple If and Then statement and the AND logic. Table 2 Rules for the inputs and output IF THEN Pressure Area Voltage Current L V L L L S M M L La M M M V M M M S M M M La M M H V H H H S M M Fig.3, Fig.4, show membership functions of input variables and Fig.5 and Fig.6 show the membership functions of the output variables Voltage and Current. Figure 3 MFs graph for Area Figure 4 MFs graph for pressure Figure 5 MFs graph for the output voltage Figure 6 MFs graph for the output current Figure 7 surface viewer graph among area, presure and output voltage This graph shows that by increasing the pressure output voltage will increase but for medium to high value of the pressure the output voltage will be high. Similarly by increasing the area the output voltage will increase but for the large value of the area the output voltage will remain medium. Basit Ali, et.al. / MEMS Based Energy Harvesting Controler Using Fuzzy Logic (2015) 178 Figure 8 Surface viewer graphs among area, presure and output current This graph shows that by increasing the pressure the output current wil lincrease but for the medium and high value of the pressure the current wiil remain high. Similarly, by in- creasing area the value of the output current increases but for the large value of area the output current will remain medi- um. B- Algorithm design for Flow Controller System For design algorithm of fuzzy logic controller the %age value of the input and output parameters are as Pressure = 23.5 Area = 75.3 Voltage = 55.6 Figure 9 graphs showing the % age values of the input parameters pressure and area and the corresponding MATLAB SIMULATED values of the output voltage and current The value of the Pressure (23.5) lies in region 1as shown in the fig. 10. Membership functions for region 1 are Low (L ) and Medium (M ). The MFs mf1 and mf2 for these values are mf1 = 50-23.5/50 =0.53 and mf2 = 1 - mf1 = 1-0.53 =0.47 Figure 10 Graph showing the values of the membership functions mf1 and mf2 for the Pressure. For Area (75.3) values lies in the region 2 as shown in the fig. 11. Membership functions for region 2 are Very Small (V) and Small (S). The MFs mf3 and mf4 for these values are mf3 = 100 – 75.3/50 = 0.494 mf4 = 1 – mf3 =1- 0.494= 0.506 Figure.11 Graph showing the values of the membership functions mf3 and mf4 for the Area For Voltage (55.6) value lies in region 2 as shown in fig. 12. Membership functions for region 1 are Low(L) and Medium(M). The MFs mf5 and mf6 for these values are mf5 = 100 – 55.6/50 = 0.888 mf6 = 1 - mf5 = 1- 0.888 = 0.112 Figure 12 graph showing the values of the membership functions mf5 and mf6 for the Voltage Selected rules for fuzzy logic controller according to value of input parameters (Pressure =23.5, Area =75.3) are listed in Table 3. Table 3 Used for Selected Rules The value of pressure is lying in Region 1 P= 23.7; Area is in Region 2 A= 75.3, Voltage is in region 2 V = 55.6 and cur- rent is in region 2 I=55.6. Pressure the 1st input of the sys- tem; whose value lays in Region 1 in MF graphs. Mfs are Low (L) and High (H).The mfs mf1 and mf2 for these values Basit Ali, et.al. / MEMS Based Energy Harvesting Controler Using Fuzzy Logic (2015) 179 are mf1 = 0.53 and mf2 = 0.47. The 2nd input parameter for the system is Area; whose value lays in Region 2 of MF graphs. Mfs are: Small (S) and Large (L). The mfs mf3 and mf4 for these values are mf3 = 0.494 and mf4 =0.506. The fisrt output parameter is Voltage whose value lies in region 2 of MF graphs.Mfs are Medium (M) and High (H).The mfs mf5 and mf6 for these values are mf5 = 0.888 and mf6 = 0.112.Table 4 shows the singleton values for this system: Table 4 shows the singleton values Table 5 shows the rules corresponding to the mfs Calculations using the Mamdani’s Formula By using the formula for Mamdani’model output is calcu- lated for both the conditions as The value of pressure is lying in Region P= 23.7; Area is in Region 2 A= 75.3, Voltage is in region 2 V = 55.6 and cur- rent is in region 2. Pressure the 1st input of the system; whose value lays in Region 1 in MF graphs. Mfs are Low (L) and High (H).The mfs mf1 and mf2 for these values are mf1 = 0.53 and mf2 = 0.47. The 2nd input parameter for the system is Area; whose value lays in Region 2 of MF graphs. Mfs are: Small (S) and Large (L). The mfs mf3 and mf4 for these values are mf3 = 0.494 and mf4 =0.506. Where Ri are rules of table 5 and Si are singleton values of table 3.Here singleton values corresponds to 2 different var- iables of current 0.5 for Medium and 1 for high. Hence Σ Si × Ri = S1× R1 + S2 ×R2+ S3 × R3+ S4× R4 =0.50×0.494+0.5×0.112+0.5×0.506+1×0.112=0.247+0.056 +0.253+0.112=0-668 ΣRi = R1 + R2 + R3 + R4 = 0.494+0.112+0.506+0.112 =1.224 Flow controller = [ ΣRi × Si / ΣRi ] = 0.668/ 1.224 = 0.546 MATLAB SIMULATION VALUE= 0.556 CALCULATED VALUE= 0.546 Difference= 0.556-0.546 = 0.010 Percentage error will be only 1% which is very small; therefore, the proposed system will performed well. The value of pressure is lying in Region P= 23.7; Area is in Region 2 A= 75.3, Voltage is in region 2 V = 55.6 and cur- rent is in region 2. Pressure the 1st input of the system; whose value lays in Region 1 in MF graphs. Mfs are Low (L) and High (H).The mfs mf1 and mf2 for these values are mf1 = 0.53 and mf2 = 0.47.The 2nd input parameter for the system is Area; whose value lays in Region 2 of MF graphs. Mfs are: Small (S) and Large (L). The mfs mf3 and mf4 for these values are mf3 = 0.494 and mf4 =0.506. The second output parameter is Current whose value lies in region 2 of MF graphs.Mfs are Medium (M) and High (H).The mfs mf5 and mf6 for these values are mf5 = 100 – 55.6/50 = 0.888 mf6 = 1 - mf5 = 1- 0.888 = 0.112 Hence Σ Si × Ri = S1 × R1 + S2 × R2+ S3 × R3+ S4× R4 =0.50×0.494+0.5×0.112+0.5×0.506+1×0.112=0.247+0.056 +0.253+0.112=0.668 ΣRi = Ra + Rb + Rc + Rd = 0.494+0.112+0.506+0.112 =1.224 Flow controller = [ ΣRi × Si / ΣRi ] = 0.668/ 1.224 = 0.546 MATLAB SIMULATION VALUE= 0.556 CALCULATED VALUE= 0.546 Difference= 0.556-0.546 = 0.010 Percentage error will be only 1% which is very small; therefore, the proposed system will perform well. III RESULTS AND DISCUSSIONS Fuzzy logic (FL) based control MEMS energy harvester is being proposed here for the control of outputs voltage and current of the MEMS based energy harveste. The given sys- tem contains FL controller which has two inputs (pressure and area) and 2 outputs (voltage and current).AND logic and the Mamdani’s model has been used here the results of whom are given below: MATLAB SIMULATION VALUE= 0.556 CALCULATED VALUE= 0.546 Difference= 0.556-0.546 = 0.010 Rules Pres- sure Area Voltage Singleton Value R1 L S M 0.5 R2 L La M 0.5 R3 M S M 0.5 R4 M La H 1 Rules Membership Functions R1 mf1˄mf3˄mf5 = 0.53˄0.494˄0.888= 0.494 R2 mf1˄mf3˄mf6 = 0.53˄0.494˄0.112= 0.112 R3 mf1˄mf4˄mf5 = 0.53˄0.506˄0.888 = 0.506 R4 mf1˄mf4˄mf6 = 0.53˄0.506˄0.112 = 0.112 Basit Ali, et.al. / MEMS Based Energy Harvesting Controler Using Fuzzy Logic (2015) 180 Percentage error will be only 1% for both the out puts Volt- age and current, which is very small; therefore, the pro- posed system will performed well. IV Conclusion In this study outputs voltage and current of the MEMS based energy harvester can be controlled by using fuzzy logic based control system. A new way of controlling the outputs by changing the inputs has been proposed. The system has been developed in MATLAB using fuzzy logic Mamdani mod- el.Two inputs pressure 75.5 and area 23.5 have been selected. The three membership functions to the input parameters have been assigned. The two outputs like voltage 55.6 and current 55.6 are selected for output power. Three membership func- tions are also assigned to the outputs. The system works ac- cording to the rules defined in the fuzzy inference system (FIS). The results have been displayed in surface view- er.Different rules of combinations were defined in MATLAB rule editor and we used AND logic for simulation. The rsults obtained from fuzzy logic controller have been verified by using Mamdani’s formula for specific values of the inputs and the outputs. As there is only 1% error for both the out- puts, current and voltage, this shows that the system performs very well. Reference [1] S.J.Roundy, ―Energy Scavenging for Wireless Sensor Nodes with a Focus on Vibration to Electricity Conver- sion‖.PhD thesis University of California,Berkely (2003). [2] Zhong Lin Wang, Youtube video on piezoelectrici- ty….energy harvesting at the nanolevel. [3] Chin-Teng Lin, C. S. George Lee, ―Neural-Network- Based Fuzzy Logic Control and Decision System‖ IEEE Transactions on Computers, Vol. 40, No. 12, December 1991. [4] J.L.Castro―Fuzzy Logic Controllers Are Universal Ap- proximators‖IEEE Transactions on Systems, Man, and Cy- bernetics, Vol 25, No. 4, April 1995. [5] K.S.Tang, K.F.Man, Z.F.Liu, S.Kwong ―Minimal Fuzzy Memberships and Rules Using Hierarchical Genetic Algo- rithms‖ IEEE Transactions On Industrial Electronics, Vol. 45, No. 1, February 1998. [6] C.Y.Sue, N.Chyuan, ―Human powered MEMS-based energy harvest devices‖. Science direct Applied Ener- gy,Volume 93, pp 390–403( 2012). [7] H.Liu, C.Lee, T.Kobayashi, C.J.Tay And C.Quan, ―In- vestigation Of A MEMS Piezoelectric Energy Harvester Sys- tem With A Frequency-Widened-Bandwidth Mechanism Introduced By Mechanical‖.Microsystem Technologies (2012), Vol.18, Issue 4, pp497-506(2012). [8] E.A.Dobisz, L.A.Eldada, ―Study of two-dimensional device-error-redundant single-electron oscillator system‖. Nanoengineering: Fabrication, Properties, Optics, and De- vices IX, Vol. 8463 (2012). [9] H.Liu, C.Lee, T. Kobayashi, C. J. Tay and C. Quan, ―A new S-shaped MEMS PZT cantilever for energy har- vesting from low frequency vibrations below 30 Hz‖.Smart Matrials and Structures Vol.21 No.3 (2012). [10] L. Dhakar, F. E. H. Tay and C.Lee, ―Investigation of contact electrification based broadband energy harvesting mechanism using elastic PDMS microstructures‖, Vol.24, No.10.(2014). [11] Jia, Y. Seshia, A. A, ―White Noise Responsiveness of an AlN Piezoelectric MEMS Cantilever Vibration Energy Harvester‖ Journal of Physics: Conference Series, Volume 557, Issue 1(2014). [12] C.Junyi, Zhou, S, Ren, X. Cao, Binggang, ―Piezoelec- tric cantilevers optimization for vibration energy harvest- ing‖.Third International Conference on Smart Materials and Nanotechnology in Engineering. Vol. 9,pp 8409 (2011). [13] Bala, U.B.Krantz, C.Matthias, Gerken, Martina, ―Electrode position optimization in magnetoelectric sensors based on magnetostrictive-piezoelectric bilayers oncantile- ver substrates‖. Ultrasonics, Ferroelectrics, and Frequency Control, IEEE Transactions on Vol.61, Issue 3 (2014). [14] Kellogg, A.Rick, A.Sumali and Hartono, ―Piezoelectric energy harvester having planform-tapered in- terdigitated beams”.United state patent Kellog et al, Patent No.: US 7,948,153 B1 (2011) [15] Leadenham, Stephen,Erturk, Alper, ―Global nonlinear electroelastic dynamics of a bi morph piezoelectric cantilever for energy harvesting, sens- ing, and actuation‖.Spie digital library Active and Passive Smart Structures and Integrated Systems (2014), Vol.9057, Issue 02, doi 10.1117/12.2045161,(2014). [16] A.Muthalif, A.Gani, ―Optimal piezoelectric beam shape for single and broadband vibration energy harvesting: Modeling, simulation and experimental results‖.Elsevier (2014). [17]R.Almudena,S.Rueda,J.Manuel,O'Keeffe,Rosemary,B. Jesus,J.Nathan,M.Alan,L.V.J.Antonio,―Tunable MEMS Pi- ezoelectric Energy Harvesting Devices‖ Microsystems Technologies (2015), DOI: 10.1007/s00542-015-2455-1. [18]K.Moonkeun, L.S.Kyun, Y.Y.Suk, J. Jaehwa, M. Nam Ki, K.K.Ho, ―Design and Fabrication of Vibration Based Energy Harvester Using Microelectromechanical System Piezoelectric Cantilever for Low Power Applications‖. Journal of Nanoscience and Nanotechnology, Vol. 13, No. 12, pp. 7932-7937 December (2013). http://www.sciencedirect.com/science/journal/03062619 http://www.sciencedirect.com/science/journal/03062619 http://www.sciencedirect.com/science/journal/03062619/93/supp/C http://adsabs.harvard.edu/cgi-bin/author_form?author=Jia,+Y&fullauthor=Jia,%20Y.&charset=UTF-8&db_key=PHY http://adsabs.harvard.edu/cgi-bin/author_form?author=Seshia,+A&fullauthor=Seshia,%20A.%20A.&charset=UTF-8&db_key=PHY http://adsabs.harvard.edu/cgi-bin/author_form?author=Cao,+J&fullauthor=Cao,%20Junyi&charset=UTF-8&db_key=PHY http://adsabs.harvard.edu/cgi-bin/author_form?author=Zhou,+S&fullauthor=Zhou,%20Shengxi&charset=UTF-8&db_key=PHY http://adsabs.harvard.edu/cgi-bin/author_form?author=Ren,+X&fullauthor=Ren,%20Xiaolong&charset=UTF-8&db_key=PHY http://adsabs.harvard.edu/cgi-bin/author_form?author=Cao,+B&fullauthor=Cao,%20Binggang&charset=UTF-8&db_key=PHY http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://www.ncbi.nlm.nih.gov/pubmed/24569244 http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://www.ncbi.nlm.nih.gov/pubmed/24569244 http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://www.ncbi.nlm.nih.gov/pubmed/24569244 http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=58 http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=58 http://ieeexplore.ieee.org/xpl/tocresult.jsp?isnumber=6746310 http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://www.osti.gov/doepatents/biblio/1017264 http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://www.osti.gov/doepatents/biblio/1017264 http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://adsabs.harvard.edu/abs/2014SPIE.9057E..02L http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://adsabs.harvard.edu/abs/2014SPIE.9057E..02L http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://adsabs.harvard.edu/abs/2014SPIE.9057E..02L http://science.gov/scigov/link.html?type=RESULT&redirectUrl=http://adsabs.harvard.edu/abs/2014SPIE.9057E..02L http://malrep.uum.edu.my/rep/Author/Home?author=Abdul+Muthalif%2C+Asan+Gani http://dx.doi.org/10.1007/s00542-015-2455-1 http://www.ingentaconnect.com/content/asp/jnn