Abstract
We investigate the evolution of nonlinear dynamic behaviors of two polarization components (x-PC and y-PC), as well as the interplay of polarization bistability, frequency detuning and injection strength in the vertical cavity surface emitting laser with optical injection. Specifically, by encoding two logic inputs and one clock input in the amplitude of the light from a sampled grating distributed Bragg reflector laser, and by decoding two output logic responses from the x-PC and y-PC emitted by the laser, we demonstrate two parallel data-selection computing. The correct logic output encoded in two emitted PCs response for as short as 100 ps bit time and the response bit time of the correct logic output encoded in the y-PC may be 67 ps by the optimization of the injection strength. The probability of a correct response is controlled by the interplay of the bit time, the injection strength and noise strength, and is equal to 1 in a wide region of the injection strength and noise strength. The chaotic data-selection computing in an optically VCSEL offer interesting perspectives for applications where noise is unavoidable and fast switching is required.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
As we all known, chaotic lasers have been generalized to apply some special fields such as chaotic radar ranging [1], optical reservoir computing [2, 3], chaotic secure communication [4–6], chaotic neural network [7, 8], physical random number [9–11]. They have also potential important applications in chaotic computation [12–14]. Optical chaos computing is a new approach to implementing a logical optic circuit. Since the chaotic laser systems exhibit different behaviors and patterns, they have potential ability to implement different types of optical logic function, and can be used to perform universal computing with appropriate combinations, showing that all possible functions can be programmed. Therefore, it is prospective that optical chaos computing can provide a new optical computer architecture. Moreover, since laser chaos has very abundant dynamic behavior, unpredictability and flexible transformation of behavior mode, compared with typical optical logic computing realized by using nonlinear effects of fiber [15], periodically poled lithium niobite waveguide [16] and semiconductor optical amplifier [17], optical chaos logic computing has more security, flexibility, anti-jamming ability and lower power cost.
Vertical cavity surface emitting lasers (VCSELs) show more advantages than edge-emitting lasers such as small volume, low cost, small threshold current, easy integration and so on, showing that they have great potential application in integrated optical chaotic computing system. When VCSELs are subjected to optical injection or optical feedback, they easily generate chaotic dynamic behaviors of two mutually orthogonal polarization components [18–26]. Using polarization bistability and abundant modes embedded in the chaotic VCSEL system, basic and complex optical logic functions can be programmed. Recently, some theoretical works about optical chaos computing focus mainly on some basic logic operations with the bit time of several nanoseconds. For example, using the polarization bistability of the chaotic VCSEL with optical injection in different setups, Masoller C et al. realized some basic chaotic logic operations, such as AND, OR and NAND [27]. Yan S L et al. implemented some basic logic operations based on the drive-response chaotic laser system, such as XNOR, NOR and so on [28]. Singh and his co-workers verified numerically the phenomenon of logical stochastic resonance in a polarization bistability laser [29]. Our previous works have been focused on controllable basic optical chaotic logic operations with the bit time of 10 ns in an optically injected VCSEL and an optical feedback injected one [30, 31]. In optical chaos computing architecture and the processing of chaotic signals in optical chaotic secure communication network, optical chaotic data-selection, as a complex chaos computing, plays an important role. It is noted that optical chaotic data-selection can be generalized to apply in more complex optical combinational chaotic computing and optical sequential chaotic computing, cooperating with other basic optical chaotic computing. The implementations of these complex chaotic logic computing are conducive to the construction of optical computing architecture, and will promote the practical process of optical chaotic network communication systems.
For all we know, optical chaotic data-selection computing is rarely concerned and no researches on it have been reported. Motivated by this, we propose a new theory mechanism and a novel scheme for the realization of high-speed optical chaotic data-selection computing with the bit time of picosecond magnitude in an optically injected VCSEL, controlling the data-selection logic operation between the frequency detuning and three logic chaotic signals including two logic input signals and one clock signal.
2. Theory and model
Figure 1 shows the implementation of two parallel optical chaotic data-selection computing. Here, a sampled grating distributed Bragg reflector laser (SGDBR), as a tunable laser, produces different central frequencies under different bias currents; the optical isolator 1 (IS1) is used to avoid the light from the fiber beam splitter 1 (FBS1) to feedback into the SGDBR. The IS2 is used to avoid the light from the fiber polarization beam splitter 2 (FPBS2) to feedback into the VCSEL. The neutral density filter 1 (NDF1) and the NDF2 are used to control the strength of two input optical signals, respectively. The strength of the optical clock signal is adjusted by the NDF3. The polarization controllers (PC1, PC2 and PC3) are used to ensure that the output light () and two ones ( and ) from the FBS2 have the same polarization. The time delays of different light paths between FPBS1 and FC2 are set as the same, and those between FBS6 and VCSEL are considered to be identical. The light from the SGDBR is separated into two beams by the 1×2 FBS1. One of the beams is injected into the 1×
2 FBS5 and the other is separated into two beams again by the 1×2 FBS2. The two beams of light from the 1×2 FBS2 are then injected into the 1×2 FBS4 and 1×2 FBS3, respectively. Two logical inputs I1 and I2 are encoded in the amplitudes of two beams of light separated equally by the FBS4 and FBS3, respectively. One beam of light from the FBS1 is separated equally into two beams by the FBS5. One beam of light from the FBS5 is encoded into one optical clock signals, which is defined as I3. The logic signals I1, I2 and I3 are coupled into one beam by the 3×1 fiber coupler 1 (FC1). To ensure that arbitrary polarized light from the FC1 can be parallelly injected into the x-polarization component (x-PC) and y-PC of the VCSEL, it is necessary to convert the arbitrary polarized light into linearly polarized light by some optical passive devices placed between the FC1 and the FC2. Here we consider the converted linear polarization light is the x-PC. The arbitrary polarized light from the FC1 is divided into the x-PC and y-PC by the FPBS1. The x-PC is directly injected into the FC2. The y-PC is firstly converted to the x-PC by the faraday rotator (FR1) and half wave plate (HWP1), then injected into the FC2. Thus, the light from the FC2 is ensured to be the x-PC, which is split into two beams by the FBS6. One beam is directly injected into the x-PC of the VCSEL, the other is converted into y-PC by the FR2 and HWP2, and then injected into the y-PC of the VCSEL. Here, the two chaotic PCs emitted by the VCSEL are encoded into the logical outputs Y1 and Y2. In particular, the frequency detuning between the SGDBR and the VCSEL is considered as the control logic signal C. Here, the output amplitudes , and are changed with . Simultaneously, the bias current μ0 of the SGDBR is varied with the change of these output amplitudes. As a result, is further varied with the change of μ0. We consider μ0 as square wave, which the low and high levels are set as μ01 and μ02, respectively. Under , the low levels of and C are and C1, respectively. When , the high levels of them are and C2, respectively. Meanwhile, μ01 and μ01 are varied with the low and high levels of three input signals (I1, I2 and I3), respectively. If the control signal C has the data-selection logic relation with the three signals I1, I2 and I3, it is possible that and . However, it is very difficult that the data-selection logic operation that is directly realized since that the C only depends on the injection current μ0. Tosolve this problem, we propose a scheme to indirectly implement the data-selection operation as follows. The amplitude of one beam of light from the FBS4 is converted into the current signals i1 by the photo-detector 1 (PD1). The noise in it from the PD1 is filtered by the low pass Bessel filter 1 (LPBF1). The i1 is amplified by the electric amplifier 1 (EAM1). The current signals i2 and i3 converted by the and , respectively, are processed in the same way. The i1, i2 and i3, in turn, are encoded into three electric logic inputs of the electronic data selector (EDS) such as , and . The logic sets of the signals , and are synchronized with those of the signals I1, I2 and I3. The output of the EDS is considered as μ0, which is the injection current of the SGDBR. The μ0 is encoded into the logic output Y0of the EDS. Here, is encoded in the low-level current μ01, is encoded in the high-level current μ02. Using the EDS, we obtain . If , , we obtain and ; when , , we have and . These show that C is logically synchronized with Y0. Thus, we further obtain when the logic sets of I1, I2 and I3 are synchronized with the those of , and . In the following, we elaborate the realization of two parallel data-selection computing in detail.
Based on the spin-flip model of VCSEEL presented by Miguel et al [32], the rate equations for the optically injected VCSEL are modified as [33]:
where x and y represent the x-PC and the y-PC, respectively; E is the normalized amplitude, , with g being the differential material gain; A is the slowly varying amplitude; N is the total carrier concentration; n is the difference in concentration between carriers with spin-up and carriers with spin down; k is the field decay; α is the line-width enhancement factor; γp is the linear birefringence; γa is the linear dichroism; γe is the nonradiative carrier relaxation rate; γs is the spin-flip relaxation rate; the normalized injection current ; Γ is the field confinement factor to the active region; U is the injected current; e is the electron charge; V is the volume of the active layer; N0 is the half the transparency carrier density; Kx is the injection strength of the x-PC and Ky is the injection strength of the y-PC; Einj= is the sum of , and , where Ainj is the slowly varying amplitude of the injected field. is the spontaneous emission coefficient, which will also be referred to as the noise strength;ξx and ξy are two independent Gaussian white noises with the mean value of 0 and the variance of 1, and their correlation coefficients are =. is the detuning between the frequency of the injected field and the reference frequency of the VCSEL. The optical frequency of the injected field =193.6 THz (the corresponding wavelength is 1550 nm); the reference optical frequency is defined as +, where and are the optical frequency of the x-PC and the y-PC of the free-running VCSEL.3. Results and discussions
From Fig. 2, it is observed that the output x-PC is chaotic state when Einj is between 0 and 0.06, or 1.42 and 1.7, or 1.74 and 3 under GHz; the output y-PC appears chaotic state while Einj ranges from 0 to 0.32, or from 1.36 to 3. Under GHz, the x-PC is chaotic state when Einj varies from 0.46 to 3, and if Einj is between 0 and 3, the y-PC is chaotic state. To determine the value of Einj that is encoded into logical inputs, for GHz and GHz, we calculate the bistability evolutions of the two output PCs with Einj, as shown in Fig. 3. From this figure, one sees that under GHz, the bistability-loops of the x-PC and y-PC locate in the space of Einj between 1.42 to 1.64, where two output PCs are chaotic states. But for GHz, the bistability-loops occur at Einj that varies from 0 to 0.174.
We suppose that Einj equals to the sum of three square waves, i.e., Einj = + + , where , are encoded into logic inputs I1 and I2, respectively; is used to encode into clock signal I3. Since logical input can be either 0 or 1, there exist eight distinct sets: (0,0,0), (0,0,1), (0,1,0), (0,1,1), (1,0,0), (1,0,1), (1,1,0), (1,1,1). We can encode the eight logic inputs with four-level signals EinjI, EinjII, EinjIII and EinjIV, where EinjI (EinjII - ) represents the set (0,0,0); EinjII denotes the sets (0,0,1), (0,1,0) and (1,0,0); EinjIV (EinjII + ) indicates the sets (0,1,1), (1,0,1) and (1,1,0); Einjvi (EinjII + ) shows the set (1,1,1). The four-level signal is constant during a bit duration T. Here, T is set to 100 ps or 67 ps, correspondingly, the rate of the chaotic data-selection can be about 10 Gb/s or 15 Gb/s. Since that under GHz and GHz, the two output PCs exhibit chaotic states in the parameter place of Einj from 1.42 to 1.7 (see Fig. 2), we take , , and as the four-level signals (see Fig. 3). Here, when , and are equal to 0.48, I1, I2 and I3 are considered as 0; if , and are equal to 0.54, I1, I2 and I3 become 1. In addition, when GHz, C = 1; C = 0 under GHz. To encode the logical output, we use the threshold mechanism to determine two logic outputs from the VCSEL. The threshold mechanism can be established when the conversion between the output x-PC and y-PC from the VCSEL is induced by the variation of the frequency detuning or the bias current μ0 between the low-level and high-level. The threshold mechanism is further explained as follows. Suppose that the total sampling time t is equal to L times the bit duration T, i.e., . The number of sampling times in each T is defined as M, and equal to , where h is the time iteration step. The mean value of the output x-PC and y-PC in the ith T (i=0, 1, 2, …, L, the same below) are defined as and , respectively, which are expressed as
where and are the amplitudes of the output x-PC and the y-PC at the jth sampling time in the ith T, respectively. To find the appropriate threshold to judge two logical outputs, we need to calculate the minimum value of and maximum value of when the output of the VCSEL is dominated by the x-PC, and compute the minimum value of and maximum value of while the output of the VCSEL is dominated by the y-PC. For this purpose, under C = 0, the number of the bit duration is considered as n0, and it is set as n1 when C = 1. Here, . Suppose that if C = 0, the output x-PC is dominant and the output y-PC is suppressed, we obtain the minimum value of and maximum value of , respectively, as follows, where the subscript xd and ys mean the dominated output x-PC and the suppressed output y-PC, respectively. Under C = 1, when the output y-PC is dominant and the output x-PC is suppressed, the maximum values of and the minimum of are, respectively, calculated by and where the subscript xs and yd mean the suppressed output x-PC and the dominated output y-PC, respectively. From Eqs. (5)-(10), the thresholds of the output x-PC and y-PC are obtained as follows,For another case, under C = 0, suppose that the x-PC is suppressed and the y-PC is dominant in the output of the VCSEL, we obtain
For C = 1, if the output of the VCSEL is governed by the x-PC, and its output y-PC is suppressed, the minimum value of and maximum value of are, respectively, computed by
From Eqs. (13)-(16), we obtain the thresholds of the two PCs as follows,
More significantly, in the eight distinct sets of three input signals I1, I2 and I3, suppose that the number of the mth input set is lm (m=1, 2, …, 8, the same below). Correspondingly, there are the response logic outputs with the number of lm in the output x-PC and y-PC. Therefore, for the mth input set, the minimum and maximum of are, respectively, obtained in the output x-PC by
Similar, in the output y-PC, we obtain
In particular, to obtain two identical logical output response encoded in the output x-PC and y-PC, the threshold judgement rule for the output x-PC is opposite to that for the output y-PC, due to the conversion between the output x-PC and y-PC. Based on the above considerations, for the mth input set, if > or , ; when < or , . While < or , ; under > or , . In our simulations, , h = 1 ps, n0 = n1 = 640, lm = 160. According to the above-mentioned encoding principle of the logic inputs and outputs, we present the implementation of optical chaotic logic selection operation, as shown in Fig. 4, where the amplitude of the injected light field varies with fast four-level signals. From Figs. 4(b) and 4(f), it is found that under . Table 2 shows the combination of the logical inputs, clock signal, control signal and logical outputs under T = 100 ps. As further shown in Fig. 4, the x-PC is dominant and the y-PC is suppressed in the output of the VCSEL under C = 0. But for C = 1, the x-PC is suppressed and the y-PC is dominant. Under the condition, according to Eqs. (5)-(12), we obtain the thresholds as follows: = 0.5653 and = 0.6821 under T = 100 ps; = 0.8439 and = 0.6004 when T = 67 ps. From Table 2 and Figs. 4(c) and 4(d), it is further seen that under T = 100 ps, if (I1,I2,I3) = (0,0,0), = 0.5788 and = 0.5248; under (I1,I2,I3) = (0,0,1), = 0.6992 and = 0.5062; when (I1,I2,I3) = (0,1,0), = 0.6873 and = 0.5290; while (I1,I2,I3) = (1,0,1), = 0.7864 and = 0.5215. It is obtained that Y1 = 0 and Y2 = 0 since > and < (m = 1, 2, 3, 6). In addition, if (I1,I2,I3) = (0,1,1), and are found to be 0.5379 and 0.8396, respectively. Under (I1,I2,I3) = (1,0,0), and are of 0.5517 and 0.8352; when (I1,I2,I3) = (1,1,0), and becomes 0.5214 and 0.8614; while (I1,I2,I3) = (1,1,1), and are of 0.4660 and 0.8792, respectively. Thus, we obtain and , owing to the case that < and > (m = 4, 5, 7, 8). It is concluded that and , showing that two parallel optical chaotic data-selection operations can be implemented. Under T = 67 ps, Table 3 presents the combinations of inputs and outputs for chaotic logic selection operation. The evolutions of the logic data-selection operation are displayed in the right column of Fig. 4. From Table 3 and Figs. 4(e)-4(h), the evolutions of the logic data-selection operation have similar behaviors with those under T = 100 ps. However, there exit some error bits in the output of Y1
owing to the faster chaotic polarization fluctuation of the response output for the x-PC caused by the shorter injected bit duration. Under the condition, using the threshold mechanism provided in Eqs. (5)-(12), the partial logic outputs are wrongly encoded in the output x-PC. As a result, the case that cannot be rightly realized. By comparison, the data-selection computing that can be rightly implemented. In order to further evaluate the output performance of digital chaotic data-selection under T = 100 ps and 67 ps, we further discuss the success probability in detail as follows.
It is noted that the data-selection logic computing depends on some key parameters of the system such as the bit duration, the injection strength and the noise strength. In other words, these parameters seriously influence the reliability of the data-selection logic operation. For this purpose, we introduce the success probability P to describe the reliability. It is defined as the ratio between the number of the correct bits and that of total bits of the chaotic logic output. Firstly, we calculate the evolutions of the success probability in the parameter space of Kx and T, as well as Ky and T when Kx = Ky. The results are displayed in Fig. 5. From Fig. 5(a), one sees that under T = 100 ps, the output Y1 shows error-free bits when Kx and Ky increase from 19.6 ns to 50 ns. With the further increase of Kx and Ky, the shorter bit duration of Y1 with error-free bits can be obtained. For example, there are no error bits in the output Y1 when Kx and Ky are more than 50 ns. As seen from Fig. 5(b), when Kx and Ky are between 16 ns and 37.2 ns, the output Y2 appears error-free under T = 100 ps. But for T = 67 ps, the output Y2 with error-free occurs in a narrow region, where Kx and Ky range from 20 ns to 23.6 ns.
To observe the influence of the noise strength on the success probability, we calculate the dependence of the success probability on the noise strength . From Fig. 6(a), it is found that the P of the Y1 is equal to 1 when < 3.2 × 109. As shown in Fig. 6(b), the P of the Y2 keeps to be 1 when < 304 × 109. These show that no error bits occur in the Y1 and Y2 under high noise strength. Therefore, two parallel chaotic data-selection computing have excellent anti-interference ability.
4. Conclusions
In conclusion, based on the chaotic polarization in an optically injected VCSEL, we propose a novel scheme for the realization of two parallel optical chaotic data-selection computing in which the bit duration is picosecond magnitude, controlling indirectly the logic data-selection relation between the frequency detuning and three signals including one optical clock signal and two optical input ones. Moreover, the success probabilities of two parallel chaotic data-selection operation are further explored in the parameter spaces of the injection strength and the bit duration. It is found that the response of the correct logic outputs encoded in the output x-PC and y-PC is short as 100 ps bit time, and that of the correct logic output encoded in the output y-PC may be short as 67 ps by the optimization of the injection strength. Finally, the success probabilities for two data-selection computing have strong robust to the noise strength. An attractive advantage of the VCSEL-based data-selection operation for practical applications is the very short bit time needed to produce the correct operation with probability equal to 1 (in our simulations, about 67 ps - 100 ps). In addition, the chaotic data-selection computing in an optically VCSEL has potential applications in optical chaotic secure communication and optical chaos computing architecture where noise is unavoidable.
Funding
National Natural Science Foundation of China (NSFC) (61475120); Major Projects of Basic Research and Applied Research for Natural Science in Guangdong Province (2017KZDXM086).
References
1. K. Myneni, T. A. Barr, B. R. Reed, S. D. Pethel, and N. J. Corron, “High-precision ranging using a chaotic laser pulse train,” Appl. Phys. Lett. 78(11), 1496–1498 (2001). [CrossRef]
2. J. Y. Gao, J. H. Huang, Z. R. Zheng, Y. Jiang, Y. Zhang, and G. X. Jin, “Hybrid bostable system and its application to dynamic memory,” Opt. Eng. 34(3), 790–794 (1995). [CrossRef]
3. A. Smerieri, B. Schneider, F. Duport, M. Haelterman, and S. Massar, “All-optical reservoir computing,” Opt. Express 20(20), 22783–22795 (2012). [CrossRef] [PubMed]
4. N. Jiang, A. K. Zhao, C. P. Xue, J. M. Tang, and K. Qiu, “Physical secure optical communication based on private chaotic spectral phase encryption/decryption,” Opt. Lett. 44(7), 1536–1539 (2019). [CrossRef] [PubMed]
5. A. Argyris, D. Syvridis, L. Larger, V. Annovazzi-Lodi, P. Colet, I. Fischer, J. García-Ojalvo, C. R. Mirasso, L. Pesquera, and K. Alan Shore, “Chaos based communications at high bit rates using commercial fibre-optic links,” Nature 438(7066), 343–346 (2005). [CrossRef] [PubMed]
6. Q. Li, T. Deng, Z. M. Wu, and G. Q. Xia, “Security-Enhanced Bidirectional Long-Distance Chaos Secure Communication,” Chin. J. Lasers 45(1), 0106001 (2018). [CrossRef]
7. C. T. Lin and C. P. Jou, “Controlling chaos by GA-based reinforcement learning neural network,” IEEE Trans. Neural. Netw. 10(4), 846–859 (1999). [CrossRef]
8. S. Y. Xiang, Y. H. Zhang, J. K. Gong, X. X. Guo, L. Lin, and Y. Hao, “STDP-Based Unsupervised Spike Pattern Learning in a Photonic Spiking Neural Network With VCSELs and VCSOAs,” IEEE J. Sel. Top. Quantum Electron. 25(6), 1700109 (2019). [CrossRef]
9. X. J. Yao, X. Tang, Z. M. Wu, and G. Q. Xia, “Multi-channel physical random number generation based on two orthogonally mutually coupled 1550 nm vertical-cavity surface-emitting lasers,” Acta Phys. Sinica 67(2), 024204 (2018).
10. S. Y. Xiang, A. J. Wen, W. Pan, L. Lin, H. Zhang, H. Zhang, X. X. Guo, and J. F. Li, “Suppression of Chaos Time Delay Signature in a Ring Network Consisting of Three Semiconductor Lasers Coupled With Heterogeneous Delays,” J. Light. Technol. 34(18), 4221–4227 (2016). [CrossRef]
11. Y. Wang, S. Y. Xiang, B. Wang, X. Y. Cao, A. J. Wen, and Y. Hao, “Time-delay signature concealment and physical random bits generation in mutually coupled semiconductor lasers with FBG filtered injection,” Opt. Express 27(6), 8446–8455 (2019). [CrossRef] [PubMed]
12. K. Murali, S. Sinha, and I. R. Mohamed, “Chaos computing: experimental realization of NOR gate using a simple chaotic circuit,” Phys. Lett. A 339(1–2), 39–44 (2005). [CrossRef]
13. M. Beyki and M. Yaghoobi, “Chaotic logic gate: A new approach in set and design by genetic algorithm,” Chaos, Solitons & Fractals 77, 247–252 (2015). [CrossRef]
14. B. Kia, A. Dari, W. L. Ditto, and M. L. Spano, “Unstable periodic orbits and noise in chaos computing,” Chaos: An Interdiscip. J. Nonlinear Sci ., 21(4), 074520 (2011). [CrossRef]
15. L. L. Li, J. Wu, J. F. Qiu, B. B. Wu, K. Xu, X. B. Hong, Y. Li, and J. T. Lin, “Reconfigurable all-optical logic gate using four-wave mixing (FWM) in HNLF for NRZ-PolSK signal,” Opt. Commun. 283(19), 3608–3612 (2010). [CrossRef]
16. Y.B. Tang, Y. P. Tang, H. W. Jiang, W. F. Ji, Y. J. Wu, and X. F. Chen, “Proposal for simultaneous all-optical AND, NOR, and XNOR logic gates using QPM cascading nonlinear effects in two PPLNs,” Chin. Opt. Lett. 11(6), 66–69 (2013).
17. F. Nian, “The Characteristic of Polarization Chaos and Coherence of Semiconductor Optical Amplifier-Based Fiber Ring Laser,” Acta Opt. Sinica 28(1), 128–131 (2008). [CrossRef]
18. Z. Q. Zhong, Z. M. Wu, and G. Q. Xia, “Experimental investigation on the time-delay signature of chaotic output from a 1550 nm VCSEL subject to FBG feedback,” Photonics. Res 5(1), 6–10 (2017). [CrossRef]
19. N. Jiang, C. Xue, D. Liu, Y. Lv, and K. Qiu, “Secure key distribution based on chaos synchronization of VCSELs subject to symmetric random-polarization optical injection,” Opt. Lett. 42(6), 1055–1058 (2017). [CrossRef] [PubMed]
20. C. H. Uy, D. Rontani, S. Breuer, and M. Sciamanna, “Non-local correlations via chaotic itinerancy in VCSEL with optical feedback,” Opt. Express 25(6), 6914–6926 (2017). [CrossRef] [PubMed]
21. Y. Y. Xie, J. C. Li, C. He, Z. D. Zhang, T. T. Song, C. J. Xu, and G. J. Wang, “Long-distance multi-channel bidirectional chaos communication based on synchronized VCSELs subject to chaotic signal injection,” Opt. Commun. 377, 1–9 (2016). [CrossRef]
22. X. J. Yang, J. J. Chen, G. Q. Xia, J.G. Wu, and Z. M. Wu, “Analyses of the time-delay signature and bandwidth of the chaotic output from a master-slave vertical-cavity surface-emitting laser dynamical system,” Acta Phys. Sinica 64(22), 224213 (2015).
23. B. B. Su, J. J. Chen, Z. M. Wu, and G. Q. Xia, “Performances of time-delay signature and bandwidth of the chaos generated by a vertical-cavity surface-emitting laser under chaotic optical injection,” Acta Phys. Sinica 66(24), 224206 (2017).
24. P. Xu, G. Q. Xia, Z. M. Wu, Q. Li, X. D. Lin, X. Tang, L. Fan, and T. Deng, “Circular Polarization Switching and Ploarization Bistability of Optically Pumped 1300 nm Spin Vertical-Cavity Surface-Emitting Lasers,” Acta Phys. Sinica 45(4), 0401002 (2018).
25. Q. Liang, L. Fan, J. Y. Yang, Z. M. Wu, and G. Q. Xia, “Narrow-linewidth Photonic Microwave Acquisition Based on an Optically Injected 1550 nm Vertical-cavity Surface-emitting Laser under Optoelectronic Negative Feedback,” Acta Phys. Sinica 46(3), 0314001 (2018). [CrossRef]
26. W. Y. Yang, G. Q. Xia, Y. S. Hou, Z. F. Jiang, T. Deng, and Z. M. Wu, “Experimental Investigation on Nonlinear Dynamics of a Multi-transverse Mode 1550 nm Vertical-cavity Surface-emitting Laser Subject to Parallel Optical Injection,” Acta Phys. Sinica 47(7), 0714002 (2018). [CrossRef]
27. J. Zamora-Munt and C. Masoller, “Numerical implementation of a VCSEL-based stochastic logic gate via polarization bistability,” Opt. Express 18(16), 16418–16429 (2010). [CrossRef] [PubMed]
28. S. L. Yan, “All-optical and combinational optoelectronic logic gates using chaotic synchronization of coupling-feedback semiconductor lasers and amplitude modulation (in chinese),” Chin. Sci. Bull 56(16), 1264–1271 (2011). [CrossRef]
29. K. P. Singh and S. Sinha, “Enhancement of "logical" responses by noise in a bistable optical system,” Phy. Reve. 83(4 Pt 2), 046219 (2011).
30. D. Z. Zhong, W. Luo, and G. L. Xu, “Controllable all-optical stochastic logic gates and their delay storages based on the cascaded VCSELs with optical-injection,” Chin. Phys. B 25(9), 342–354 (2016). [CrossRef]
31. D. Z. Zhong, W. Luo, G. L. Xu, and Z. Z. Xiao, “Recongurable dynamic all-optical chaotic logic operations in an optically injected VCSEL,” Chin. Phys. B 26(12), 261–271 (2017). [CrossRef]
32. J. Martin-Regalado, F. Prati, M. San Miguel, and N. B. Abraham, “Polarization properties of vertical-cavity surface- emitting lasers,” IEEE J. Quantum Electron 33(5), 765–783 (1997). [CrossRef]
33. M. F. Salvide, C. Masoller, and M. S. Torre, “All-optical stochastic logic gate based on a VCSEL with tunable optical injection,” IEEE J. Quantum Electron. 49(10), 886–893 (2013). [CrossRef]