This PDF file contains the front matter associated with SPIE Proceedings Volume 7632, including the Title Page, Copyright information, Table of Contents, and the Symposium and Conference Committee listings.

High bit rates optical communication systems pose the challenge of their tolerance to linear and nonlinear fiber impairments. Coherent optical receivers using digital signal processing techniques can mitigate the fiber impairments in the optical transmission system, including the chromatic dispersion equalization with digital filters. In this paper, an adaptive finite impulse response filter employing normalized least mean square algorithm is developed for compensating the chromatic dispersion in a 112-Gbit/s polarization division multiplexed quadrature phase shift keying coherent communication system, which is established in the VPI simulation platform. The principle of the adaptive normalized least mean square algorithm for signal equalization is analyzed theoretically, and at the meanwhile, the taps number and the tap weights in the adaptive finite impulse response filter for compensating a certain fiber chromatic dispersion are also investigated by numerical simulation. The chromatic dispersion compensation performance of the adaptive filter is analyzed by evaluating the behavior of the bit-error-rate versus the optical signal-to-noise ratio, and the compensation results are also compared with other present digital filters.

We propose a novel multi-wavelength regeneration scheme in which timing walk-off on different WDM channels is controlled to avoid the overlapping of the channels in time domain such that effective mitigation of the nonlinear interchannel impairments can be obtained, and the number of supported channels in the regeneration is increased by applying a bidirectional fiber configuration. In our scheme, optical 2R regeneration is successfully demonstrated for four 10-Gb/s RZ-OOK WDM channels spaced by 300 GHz. Both the scalability and the cascadability of our proposed regenerator were also evaluated in numerical simulations.

We proposed and experimentally demonstrate the generation of a novel star-16QAM using a dual-parallel Mach-Zehnder modulator (DPMZM) followed by a phase modulator (PM). Coherent detection and off-line processing are used to recover constellation diagrams. We also study optical delay of a novel star-16QAM signal through a silicon microring resonator. Delay time of ~ 30ps is observed by comparing the eye diagram of the star-16QAM signal on-resonance with that off-resonance.

Asynchronous optical under-sampling in combination with subsequent digital signal processing is an economical approach for characterization of ultrahigh speed optical data signals. In this paper, we propose a digital signal processing technique for bit pattern monitoring of ASK signals, which effectively mitigates the random walk clock drift between the data signal and the sampling source in such asynchronously under-sampling systems. The method is verified by measurements on 40 Gb/s NRZ and 160 Gb/s RZ data signals. A quantitative analysis of the displayed timing jitter is performed.

We present a novel hardware architecture for digitally adaptive feed-forward equalization (FFE) suitable to compensate the inter-symbol interference (ISI) caused by chromatic (CD) and time-varying polarization mode dispersion (PMD) in intensity modulated optical links with direct detection (IM/DD). Existing analog tapped delay lines for realizing the equalization filter at a bit rate of 40 Gbit/s commonly use external manual or random dithering approaches for tap weight adjustment^1,2. While manual tap weight adjustment is impractical for systems with randomly time-varying behavior, random dithering of the tap weights to find the optimal setup shows adaptation times above 1s which exceeds the measured PMD variations in installed fibers³ (~10ms) by far. Our solution follows a completely digital implementation approach and it can be scaled to various bit rates using distributed arithmetic (DA) and some parallelization techniques. The digital adaptation unit, which employs a simplified Least-Mean-Square-Algorithm (LMS)⁴, is directly implemented together with the FFE. Measurements in our hardware-in-the-loop testbed with a Virtex-II field programmable gate array (FPGA) from Xilinx have demonstrated that it is able to track time-varying optical channels well within 1 ms at a bit rate of 10.7 Gbit/s.

An all-optical regenerator based on single pump fiber-optic parametric amplifier (FOPA) is demonstrated and assessed experimentally. The static power transfer function of the FOPA is measured, and a clear saturation behavior is observed with the saturation much easier to achieve at a larger the pump power. The regenerator performance is further investigated by changing parameters of the input signal. Results show that the SNR and timing jitter of the regenerated signal is maintained around 15 dB and 2 ps regardless of the transmission distance of the input signal, and above 7 dB SNR improvement can be achieved when the input SNR is within the range of 3 to 10 dB. The regeneration performance is also tested by varying the input signal wavelength, and at least 6 dB SNR improvement is obtained within the range of the gain bandwith of the single pump FOPA.

In this paper, a novel phase regenerator with semiconductor optical amplifier is proposed. By theoretical analysis and simulation, the optimum parameters of phase regenerator are obtained. By comparing the phasor diagrams, received data and Q-factors before and after regeneration, the effectiveness of the phase regenerator is proved. The results show that nearly ideal regeneration is achieved and Q-factors are gotten great improvement.

In this paper, all-optical OFDM transmitter and receiver with 5×20Gbit/s were simulated. The 100-Gbits/s signal was successfully transmitted over 40-km-SMF without dispersion companion in the simulation experiment. We have analyzed the principle of orthogonal carrier generation, and obtained the optimal configuration for two, three and five orthogonal carriers. Through designing proper system parameters, the clearly eye diagrams were obtained. The BER curves of five orthogonal signals were detected.

We report an experiment of endless control of PMD compensation in 25-Gb/s CSRZ-DQPSK system using a DSP based PMD compensator. The control algorithm used is a modified particle Swarm Optimization. The PMD compensator can track the average SOP variation of 65 rad/s without any lost of optimum tracking.

An endless polarization stabilizer based on DSP system is reported. It can transform the state of polarization (SOP) of optical signal to any desired SOP and maintaining output signal at the desired SOP. Feedback controlling method is applied to the proposed polarization stabilizer. We design a target function relates to current SOP and the desired SOP in transmission link. It has only one extremum when SOP in transmission link is exactly the same with the desired SOP. Particle swarm optimization algorithm is employed to search the extremum point. To investigate performance of the proposed polarization stabilizer, HP11896A polarization controller is used as polarization scrambler. We test performance of the polarization stabilizer under rate 1 (~16 rad/s), 4 (~33 rad/s), and 8 ( ~65 rad/s), separately. Under the existing condition of our laboratory, the developed polarization stabilizer has the ability of stabilization up to 65 rad/s SOP changes (rate 8).

Phase shaped binary transmission (PSBT) is a kind of duobinary modulation formats. It has narrower spectrum structure. In this paper, the principle of PSBT format is investigated and its performance in chromatic dispersion, nonlinearity and polarization mode dispersion (PMD) are studied. The PMD compensation system is set up and the results show that PMD tolerance is enhanced effectively after compensation.

This talk focuses on the actual field implementation results of 40G in a video network scenario. OSNR, DCM /PMD requirements are described in the actual overlay environment of current 10G infrastructure. Results are extended to 100G where coherent detection in single carrier and OFDM approaches are compared and summarized. Standards requirement of 40G to achieve datacom compatibility with 10GE is also described. We conclude with the necessary ingredients of the ecosystem required for a realizable migration from 40G to 100G.

Multiple channels of ADCs for high bit rate CO-OFDM system is proposed by jointly processing outputs of all channels with a simple algorithm. The required sampling rate of ADCs is reduced lower than Nyquist rate.

The edge-triggered generation of baseband ultra-wideband (UWB) monocycle pulse using a dual-parallel Mach-Zehnder modulator is proposed and experimentally demonstrated. By setting three biases of the modulator at maximum, minimum and linear transmission points respectively, a monocycle UWB pulse sequence with the fractional bandwidth of ~ 155% is generated. The pulse sequence is differential encoded and the pulse bandwidth is determined by the transition time of the drive edge due to the nonlinear biases of the modulator. Furthermore, by using a basic digital signal processing (DSP) algorithm to calculate the bit error rate (BER), the transmission performance over a 25km-fiber is studied.

The optical pulse trains at a multiplied repetition frequency is generated by Fractional Talbot effects in optical fiber. The principles of Talbot effects and Fractional Talbot effects in fiber are analyzed in this paper. The 20GHz, 40GHz and 50GHz optical pulse trains are obtained from the original pulse trains with 2.8ps width at 10GHz repetition frequency. An optical pulse at 40GHz repetition frequency is also produced by upshifting the repetition frequency of 5GHz pulse trains. The degradation of 40GHz and 50GHz pulse trains is discussed by spectrum analysis of the pulse trains and some methods are proposed to enhance the performance of the pulse trains.

We propose a new automatic optical polarization demultiplexing scheme for polarization-division-multiplexed (PDM) signals, which uses the tag light transmitted in a different but close wavelength to the data signal as the feedback control. The effectiveness of this scheme is demonstrated in a 10-Gb/s PDM optical system over 50-km SMF link.

We propose a novel equalization and timing recovery scheme, which adds an adaptive butterfly-structured equalizer in an all-digital timing recovery loop, for polarization multiplexing (POLMUX) coherent receivers. It resolves an incompatible problem that digital equalizer requires the timing recovered (synchronous) signal and Gardner timing-error detection algorithm requires the equalized signal because of its small tolerance on dispersion. This joint module can complete synchronization, equalization and polarization de-multiplexing simultaneously without any extra computational cost. Finally, we demonstrate the good performance of the new scheme in a 112-Gbit/s POLMUX-NRZ-DQPSK digital optical coherent receiver.

In order to solve the phase rotating problem for constant modulus algorithm (CMA) in polarization division multiplexing system with optical coherent receivers, a modified constant modulus algorithm based on decomposing cost function into real and imaginary parts is proposed in this paper. Simulation results show that, the proposed algorithm can compensate any angle phase distortion arising from channel, and deal with phase offset caused by intradyne receiver, with better performance comparing the conventional CMA.

Received signal from weather satellite is overlaid with WDM RoF technique on 10.7Gb/s (OTU2) data stream. For extending the signal transmission distance between the antenna and the receiver, a low noise amplifier is designed by using an EP-HEMT to achieve lower noise and higher gain so that the received signal level is as low as -130dBm. The received signal is amplified by using a laser driver which is also designed by using the EP-HEMT. A low noise and linear 1550nm DFB laser is modulated by the laser driver. The converted optical signal is overlaid on a 10.7Gb/ (OTU2) data streams by using a 1310/1550nm WDM . A clear image from the weather satellite is received after 40km SMF. There was no degradation in the eye diagram and total jitter of the 10.7Gb/s data signal.

We propose and demonstrate a simple and flexible all-optical NRZ-DQPSK receiver based on optical detuned filtering. The proposed demodulator is quite flexible to input bit rates, which means it can be used for different bit rates without any change. With two given optical band-pass filters, simulation and comparison show that NRZ-DQPSK signals at 20, 40 and 80Gb/s (10, 20 and 40G symbol rate) can be demodulated to in-phase and quadrature data successfully by adjusting the detuning of two given filters.

There are three key modules in coherent optical code division multiple access (OCDMA) system: source of ultra-short optical pulse, en/decoder and receiver. Two-user 2.5Gbps OCDMA experiment is shown to achieve 100km transmission using Gain Switch Laser Diode ,63-chip Equivalent phase shift super structured fiber Bragg grating (EPS-SSFBG) phase en/decoder, and threshold adjustable receiver with broadband decision. The match between optical pulse source and en/decoders is also an important issue in OCDMA system design. Half band application for EPS-SSFBG en/decoder is proposed and testified in the experiment. Two receiver schemes, which are bit rate receiver and broadband decision receiver, are compared in experiment. Relationships between BER and decision level&time are observed in the experiment, error free transmission was achieved in the range of 10ps decision time and 17% decision level. Results of the experiment show that OCDMA system with fewer users can be put forward with fairly well performance and acceptable expenditure.

A WDM-compatible spectrally phase encoded-optical CDMA scheme is experimentally demonstrated. The system is based on the use of chromatic dispersion, chosen to achieve a first-order Talbot effect, to obtain a frequency-to-time mapping that allows subsequent coding/decoding of the signal spectrum by applying electro-optic phase modulation in the time domain. Independent coding of each of the spectral lines from a pulsed source is achieved, thus guaranteeing an efficient use of the spectrum. The use of electrooptic modulators to impose the codes enables a simple and rapid reconfiguration of the used codes. Results are given for a system transmitting at 10 Gbps within a 0.7 nm optical window using Hadamard codes.

We have demonstrated a reconfigurable time domain spectral phase encoding scheme for coherent optical code-divisionmultiple- access application. The proposed scheme is based on the concept of temporal pulse shaping dual to spatial arrangements. It uses Fiber Bragg Gratings as dispersive elements and electro-optic modulators. The data speed is 1.25 Gbps and the code is introduced at 10 Gcps, using a subset of the Hadamard codes with a length of 8 chips within a 0.7 nm optical window. The system is electrically reconfigurable and compatible with fiber systems and permits scalability in the size of the codes by modifying only the phase modulator velocity.

A novel two-level signaling technique in OOC-based fiber-optic CDMA systems is proposed. Users of the system are categorized into two classes. Users of class 1 and 2 transmit the optical pulses at power level P and 2P respectively. We study two kinds of receiver structure for the proposed two-level system. For the first structure we use ordinary AND logic gate and for the second one we use optical ADC with AND logic gate structure. We obtain the system performance and we show that for the two-level signaling multi-access interference is considerably suppressed. Two-level signaling increase system performance and capacity and also presents different QoS for the users in two classes.

Colorless optical network unit (ONU) is a very important concept for the wavelength division multiplexing (WDM) based passive optical networks (PON). We present a novel scheme to construct non-wavelength-selective ONUs in WDM/OCDMA-PON by making use of the broad spectrum band of superstructure fiber Bragg gratings (SSFBGs). The experiment results reveal that the spectrum-sliced encoded signals from different wavelength channels can be successfully decoded with the same SSFBGs, and thus the proposed colorless ONU scheme is proved to be feasible.

An extended reach EPON repeater is one of the solutions to effectively expand FTTH service areas. In this paper, we propose a reconfigurable multi-port EPON repeater for effective accommodation of multiple ODNs with a single OLT line card. The proposed repeater, which has multi-ports in both OLT and ODN sides, consists of TRs, BTRs with the CDR function and a reconfigurable electrical matrix switch, can accommodate multiple ODNs to a single OLT line card by controlling the connection of the matrix switch. Although conventional EPON repeaters require full OLT line cards to accommodate subscribers from the initial installation stage, the proposed repeater can dramatically reduce the number of required line cards especially when the number of subscribers is less than a half of the maximum registerable users per OLT. Numerical calculation results show that the extended reach EPON system with the proposed EPON repeater can save 17.5% of the initial installation cost compared with a conventional repeater, and can be less expensive than conventional systems up to the maximum subscribers especially when the percentage of ODNs in lightly-populated areas is higher.

Ethernet is the dominant transport technology for Local Area Networks. Efforts are now under way to use carrier-grade Ethernet in backbone networks of different service providers. With the advent of applications such as IPTV and Videoon- Demand, there is need for techniques to route multicast traffic over the Ethernet backbone networks. Here, we address the problem of Routing and Wavelength Assignment (RWA) of a set of multicast requests in a Multi-Line-Rate Ethernet backbone network with the objective of minimizing the cost of setting up the network, in terms of the Service Provider's Capital Expenditure (CAPEX). We present an Auxiliary Graph based heuristic algorithm that routes each multicast request on a light-tree structure, and assigns minimum cost wavelengths along the route. We compare the properties of the algorithm to the optimal solution given by a mathematical model formulated as an Integer Linear Program (ILP), and show that they compare very well. We also find that the algorithm is most cost-effective when the incoming requests are processed in descending order of their bandwidth requirements.

To support multicast in the optical layer, a wavelength cross connect node need to be equipped with light splitters. Wavelength conversion can improve the fiber link utilization and reduce the blocking probability for both unicast and multicast traffic. However, it is of high complexity to integrate light splitters and wavelength converters in a single node. This paper investigates a new resource allocation strategy in optical multicast network: a subset of optical nodes are multicast-capable, while the other nodes are wavelength convertible. The optimal configuration of the wavelength crossconnect nodes is investigated to produce the minimal blocking probability for dynamic multicast traffic.

Effect of seasonal temperature fluctuations on performance 160 Gbps optical fiber networks with chromatic dispersion compensation is studied. Total compensation is obtained by adjusting both length of compensating fiber and compensation slope factors of transmission fiber that vary with seasonal temperature. Based on numerical models on temperature dependence and measurements data, numerical simulation studies have been performed. The obtained results show quantitative compensation requirements for seasonal thermal effects for the range of temperatures considered during seasonal changes in Muscat area of more than 60 degrees. The results show also that NRZ coding format is more tolerant to seasonal temperature fluctuations than RZ coding format.

We describe a method to determine the channel capacity of an arbitrary multilevel modulation scheme by modeling the fiber-optic channel as a dynamical nonlinear ISI channel with memory. We also propose a multilevel turbo-equalization scheme that is able closely to approach the channel capacity. We show that with this scheme we are able straightforwardly to upgrade currently installed 10 Gb/s optical transmission systems to 100 Gb/s, and even with small memory assumption we are able to achieve 100 Gb/s per DWDM channel transmission over 9600 km.

We experimentally demonstrate a multi-pulse all-optical time domain continuous Fourier transformation (MOFT) based on a time lens. With two of MOFT devices, one is at transmitter, the other is at receiver, we experimentally show a successful all-optical OFDM fiber transmission system with 20Gbit/s 200km transmission without any dispersion compensations.

A novel optical single-sideband directly detected orthogonal frequency division multiplexing (DD-OFDM) system based on balance receiver consisting of two PINs is proposed and demonstrated. By introducing phase shift of π to one arm of the proposed receiver, inter-modulation distortion (IMD) near the optical carrier can be eliminated efficiently. The simulation results show that our system improves spectral efficiency to about 1.54 bps/Hz with 4-quadrature amplitude modulation (4QAM) mapping, and the optical carrier and signal power ratio (CSPR) has negligible impact on the OSNR performance, which can be improved by 3 dB with negligible error vector magnitude (EVM) penalty. Additionally, the optimum CSPR can be controlled by adjusting them respectively. 5 Gbit/s data with 4QAM mapping and 6.25% cyclic prefix (CP) inserted is successfully transmitted over 320 km of standard single mode fiber (SSMF) and our system exhibits a negligible penalty. Higher spectral efficiency can be achieved with more complex modulation formats and polarization multiplexing.

In an effort to meet core, regional, and metro network requirements, 40G DWDM suppliers have employed numerous modulation formats. Each modulation format brought its own unique set of optical, electrical, and optoelectronic component requirements, which challenged component suppliers to prioritize research and development efforts. In addition to a highly segmented market, a lack of investment in standardization has resulted in a highly fragmented market for suppliers of 40G technology. We explore how the 100G market may differ.

Even though OFDM is a well established modulation format in the wireless community, it is relatively new in fiberoptics. In this paper we give an overview of the past, present and future of optical OFDM for long-haul fiber-optic transmission systems and discuss several application spaces for the commercialization of optical OFDM.

Principle of 7 × 5.71 Gbit/s on-off-keyed (OOK) over duty cycle division multiplexing (DCDM) technique is presented. The 40 Gbit/s DCDM signals are then recovered with a receiver operates at 5.71 GHz clock using seven sampling circuits. With seven-user DCDM we show that spectral width of return-to-zero signal can be reduced around 42.8 %. The 40 Gbit/s OOK using DCDM required an optical signal-to-noise ratio of around 31.6 dB and tolerate around 50 ps/nm of chromatic dispersion. In addition, performance of the system is compared against other multiplexing techniques and modulation formats.

We show 1-Tb/s single-channel CO-OFDM transmission consisting of continuous 4,104 spectrally-overlapped subcarriers generated using a novel device of recirculating frequency shifter (RFS). The 1-Tb/s CO-OFDM signal with a spectral efficiency of 3.3 bit/s/Hz is successfully received after transmission over 600-km SSMF fiber without either Raman amplification or dispersion compensation.

We propose a cost-efficient method of multiple receiver optical OFDM system that reduces the ADC requirement and FFT size of receiving. Aided by signal predistortion, aliasing free signal can be retrieved independently and directly at the low bandwidth receivers. Simulation and experiment results are given.

We demonstrate a 160 Gb/s orthogonal frequency division multiplexing (OFDM) system using an all-optical symbol generator based on planar light circuit (PLC) technology. Excellent bit error rate (BER) is observed after long-distance transmission. The proposed symbol generator fundamentally eliminates the processing speed limits introduced by electronics and is suitable for high integration, making it physically realizable to build high-speed all-optical OFDM systems with a large number of subcarriers.

We show numerical simulation results of three different timing synchronization methods in Optical Orthogonal Frequency Division Multiplexing (OOFDM) system using Intensity-Modulation and Direct-Detection (IM/DD) in Multi-Mode Fiber (MMF) channel and do comparative study of the three algorithms. Based on Park's algorithm, we present a modification to it that can obtain higher time estimation accuracy in OOFDM system using IM/DD.

We investigate SPM suppression for coherent optical OFDM systems utilizing three PAPR reduction methods including the clipping, the selective mapping and the partial transmit sequence.

We review existing BER estimation methods and propose alternative methods to assess the performance of multilevel modulation formats with both direct and coherent detection. The impact of digital signal processing (DSP) on the BER estimation procedure is discussed for the latter case. The different approaches are illustrated by simulating exemplary transmission systems.

Optical coherent detection with digital signal processing (DSP) attracts more and more interests owing to its high performance. Due to a free-running local oscillator (LO) on the receiver, a frequency offset (FO) to the optical carrier brings phase distortion with which data can not be correctly decided and decoded. All-digital feed-forward FO estimation algorithms for optical (D)QPSK coherent receivers, such as PADE, have advantages of low complexity and stable-running wide range of estimation. However, it only converges under the condition that the initialed FO is near enough to the real FO while possible FO with typical tunable lasers can be very large. In order to solve this problem, we present a novel FO estimation scheme by BER aiding. This scheme has been demonstrated to be accurate and effective to break the restriction of PADE by simulation.

This paper analyses and achieves the accurate computational expression of nonlinear phase noise cos(Δφ) in DPSK receivers using optical Mach-Zehnder interferometer (MZI) demodulation and balanced detection of 40Gbit/s system. It had been used in the BER computation expression considering all the noises to achieve a final accurate computation expression of BER.

As coherent optical orthogonal frequency-division multiplexing (CO-OFDM) is sensitive to noises, so the channel estimation becomes a key issue for the performance. There are mainly two OFDM channel estimation methods, Least Square (LS) and Linear Minimum Means Square Error (LMMSE) included. LS estimator is used commonly in the CO-OFDM systems due to its simple realization and low complexity. However, the impact of noise is not considered, so the performance will deteriorate when large noise occurs. On the contrary, LMMSE estimator can compensate the losses more accurately than LS. But classic LMMSE estimator is difficult to realize for prior information about the channel needed in advance. In this paper, an improved LMMSE estimator for CO-OFDM systems has been introduced for the first time. Simulation shows that it can bring 2dB SNR gains compared with LS for BER = 10^-4 .

Two novel optical phase modulation formats, the staggered differential phase-shift keying (SDPSK) with return-to-zero (RZ) and carrier-suppressed return-to-zero (CSRZ) shape, are proposed and compared by numerical simulation for 100 Gbit/s medium-range transmission. RZ-SDPSK and CSRZ-SDPSK signals are generated by cascading two phase modulators and a dual-arm Mach-Zehnder modulator. The demodulation of two phase formats can be achieved on 1 bit rate through only one Mach-Zehnder delay interferometer and a balanced receiver. The transmission performance of the extracted RZ and CSRZ clock signals are compared.

As illustrated by optical fiber and optical amplification, optical telecommunications have appeared for the last ten years as one of the most promising candidates to increase the transmission capacities. More recently, the concept of optical transparency has been investigated and introduced: it consists of the optical routing of Wavelength Division Multiplexed (WDM) channels without systematic optoelectronic processing at nodes, as long as propagation impairments remain acceptable [1]. This allows achieving less power-consuming, more scalable and flexible networks, and today partial optical transparency has become a reality in deployed systems. However, because of the evolution of traffic features, optical networks are facing new challenges such as demand for higher transmitted capacity, further upgradeability, and more automation. Making all these evolutions compliant on the same current network infrastructure with a minimum of upgrades is one of the main issues for equipment vendors and operators. Hence, an automatic and efficient management of the network needs a control plan aware of the expected Quality of Transmission (QoT) of the connections to set-up with respect to numerous parameters such as: the services demanded by the customers in terms of protection/restoration; the modulation rate and format of the connection under test and also of its adjacent WDM channels; the engineering rules of the network elements traversed with an accurate knowledge of the associated physical impairments. Whatever the method and/or the technology used to collect this information, the issue about its accuracy is one of the main concerns of the network system vendors, because an inaccurate knowledge could yield a sub-optimal dimensioning and so additional costs when installing the network in the field. Previous studies [1], [2] illustrated the impact of this knowledge accuracy on the ability to predict the connection feasibility. After describing usual methods to build performance estimators, this paper reports on this impact but at the global network level, quantifying the importance to account for these uncertainties from the early network planning step; it also proposes an improvement of the accuracy of the Quality of Transmission (QoT) estimator to reduce the raise of planned resources due to these uncertainties.

Nonlinear effects induced by Kerr effect in optical fibers result in the severe inter-channel crosstalk and degrade WDM transmission system performances. The electrical post-compensation through backward propagation is demonstrated for nonlinear impairment compensation of FWM via phase diversity detection, whose required bandwidth is less than that of heterodyne detection. By 20-km transmission of dispersion-shifted fiber (DSF), the generated FWM components with both of the original optical signals whose wavelength are around 1550nm are detected by phase diversity detection with one local oscillator. The detected signals are compensated by off-line signal processing using backward propagation, and after compensation there is over 6dB improved for the power difference between signals and FWMs.

Numerical simulations were conducted to study the performance of a 112Gbit/s coherent polarization multiplexing (PolMux) quadrature phase-shift keying (QPSK) system. We addressed the bit resolution, bandwidth of analog-to-digital converters (ADC) in the presence of all chromatic dispersion, polarization-mode dispersion and imperfect laser frequency offset. Both return-to-zero (RZ) and non-return-to-zero (NRZ) QPSK waveforms are taken into account, while the filter type of ADC is assumed to have a Gaussian frequency response. An ADC with 6-bit resolution is found to be sufficient for 112Gbit/s PolMux QPSK transmission. And simulation results show that the ratio of optimum bandwidth of ADC to symbol rate is 0.5 for both RZ- and NRZ-QPSK formats when the 1st-order Gaussian frequency response of ADCs is assumed.

The inconsistency of device performance leads to the power imbalance among all channels on optical beamforming networks. The impact of power imbalance is analyzed and an improvement scheme is presented in this paper. Both experiment and simulation results prove the power imbalance seriously impairs the system functions and a suitable power-balanced element can improve the whole system performance remarkably.

In-phase/quadrature-phase (IQ) imbalance can result in severe performance degradation in optical direct-detection orthogonal-frequency-division-multiplexing (DD-OFDM) systems. We build two optical back-to-back DD-OFDM systems, which implement double-sideband (DSB) and single-sideband (SSB) modulation, respectively. The tolerance to IQ imbalance of these two systems is analyzed and compared using error vector magnitude (EVM) and symbol error rate (SER). We find that, in the back-to-back case, the DSB system has stronger robustness to IQ imbalance than the SSB System. We further build two optical DD-OFDM systems each with 40-km transmission, which respectively implement DSB and SSB transmission. Similarly, we analyze and compare the tolerance to IQ imbalance of these two systems in terms of EVM and SER. We find that, however, in the case of 40-km transmission, the SSB system has stronger robustness to IQ imbalance than the DSB system. As a result, we conclude that, in the case of transmission, SSB modulation can enhance the tolerance to IQ imbalance of DD-OFDM systems.

The paper discusses the challenges of using radio-over-fiber systems to distribute multi-gigabit-per-second wireless signals at mm-wave frequencies. We propose possible solutions to the challenges, and demonstrate the potential of simple radio-over-fiber system architectures to support multi-standard wireless communication at data speeds exceeding 14Gbps using the 60 GHz band.

We demonstrate error free transmission of a single 2.5Gbit/s NRZ data channel at 1550nm over 1 km of free space, and the results of 3Gbit/s experiment are given at the same time. The transmitted beam is detected by avalanche photodiode (APD), Bit Error Rate (BER), Signal-to-Noise Ratio (SNR) and Jitter of the detected signals are measured and compared.

Reconfigurable shutter-based free-space optical switching technologies using fiber ribbon and multiple wavelengths per fiber for Storage Area Networks (SANs) application are presented and demonstrated.

An indoor wireless communication system using commercial white LED is presented in this paper. OOK (On-Off Keying) modulation is adopted in the transmitter module. Data transmission is realized based on RS-485 protocol. A DC-balanced coding with error check and correction abilities is applied to avoid blinks and improve the performance of the system. The system can keep 10-7 error bit rate within 2.5m reach at a rate up to 115200 bps.

Using optical processing techniques, we experimentally enhance the physical layer security of optical communication systems. We exploit optical encryption using fiber nonlinearity to achieve real time data encryption. By implementing interleaved waveband switching modulation and variable two-code keying to the system, the security of the data is further enhanced. Based on spread spectrum, we also demonstrate optical steganography such that the stealth signal is transmitted underneath system noise. Optical steganography in WDM and optical CDMA systems is experimentally demonstrated. We also propose and study optical CDMA-based backup channels that improve service availability without wasting the bandwidth in the backup channel. The multi-layered security provided improves the confidentiality and availability of the network.

We investigate amplitude clipping for electrical-dispersion-predistorted 10-Gbit/s optical fiber transmission systems. Simulation results show that amplitude clipping can effectively suppress SPM-induced nonlinear impairment.

We theoretically and experimentally study of the four-wave mixing (FWM) between co-propagated pump and signal in a distributed Raman amplifier (DRA) over a 25 km of non-zero dispersion shifted fiber (NZDSF) with zero dispersion wavelength of 1495 nm. The 100 mW pump Fabry-Perot (FP) spectra centered at 1440 nm is reproduced via FWM around the single-wavelength probe signal around 1550 nm. The maximum FWM generated power ratio is found in the matching of the group delay between the signal and the two adjacent modes with larger powers in the pump Fabry-Perot laser diode. The zero-dispersion wavelength of NZDSF can also be determined at the middle of the center wavelength of pump and signal waves where maximum FWM occurs at minimum phase mismatching between two 14xx nm FP pump wavelengths and two 15xx nm signal wavelengths.

In this paper, we present a report about the impact of the distributed second order Raman amplification on Brillouin Scattering in a long haul 100km fiber system experimentally. The second order lasers at 1486nm created by the primary pump at 1395nm amplify the signal centered at 1550nm and extend the Brillouin scattering distance. Experimental results show that, both the Brillouin scattering efficiency at high powers and the Brillouin threshold are reduced.

Spectrum compression based on filters for improving bandwidth efficiency is presented and demonstrated experimentally in 40Gb/s OTDM system. Even after 100km transmission, demultiplexing and clock recovery can be implemented successfully, and the data rate-to-bandwidth ratio has been improved to 4 times.

Radio-over-Fiber (RoF) systems have been actively researched to provide future wireless broadband services. In this paper we review the schemes and strategies that have been carried out in Radio-over-Fiber networks over the years to realize high-performance RoF links.

Radio-over-fiber (RoF) system, which is a promising integration solution of wireless and optical systems, has been attracting increasing attention for its significance in providing broadband wireless access with increased capacity and mobility of the access network and reduced cost of the system. 16-Star and 16-Square QAM signals, carrying four information bits by a single symbol, are attractive modulation format in radio over fiber system because of high spectral efficiency and low symbol rate. In this paper, using a quad-parallel Mach-Zehnder modulator followed by a single-drive Mach-Zehnder modulator, we propose a scheme to generate a 16-Star/Square QAM radio frequency (RF)-signal in radio-over-fiber system. The RF carrier is obtained by quintupling the frequency of the employed local oscillator signal, thus our proposal is cost-effective.

We demonstrate a new scheme of 60GHz multichannel mm-waves generation in DWDM-ROF based on spectrum broadening, superstructure fiber Bragg grating (SFBG) comb filtering and two wavelength waves heterodyning. We defined carrier-to-carrier rate(CCR) to evaluate the performance of our scheme through analyze its impact on the output RF power. This scheme has excellent advantages, for instance, simple structures, multi-channel carrier generation, multi-frequency mm-wave generation, easily controlled frequency by choosing suitable parameters.

An analytical model to describe the BER and power penalty of a radio-over-fiber system with different modulation schemes is derived. It proves DPSK can provide 3-dB higher receiver sensitivities than OOK, and the optimal radio-over-fiber structure is DPSK-PM or DPSK-IM, which is depended on the frequency of subcarrier and optical fiber transmission distance.

In this paper, a dual-stage optical single sideband (DSSB) modulation scheme to combat chromatic dispersion for the terahertz wave over fiber (ToF) systems application is proposed. In the scheme, terahertz(THz) wave is obtained via carrier-suppression double sideband modulation, one sideband of terahertz wave then undergo the first single sideband modulation and the second base band SSB modulation in the process of data modulation, which achieves cascade single sideband (SSB) modulation. As can been seen from the simulation result, for mid-frequency carrier of 0.1THz, the BER performance without dispersion compensation achieves 1x10^-11 and the sensitivity of the receiver is -26.6dBm with a data rate of 5Gb/s after 100km transmission. Through comparing the power penalty of double side band (DSB) and DSSB, we demonstrated the optical DSSB modulation has better dispersion tolerance. This paper also shows the contrastive curves of frequency selective fading for the SSB and DSB modulated signals, which definitely demonstrates the existence of dispersion fading during the DSB transmission as well as the effective suppression of the fading implemented by the SSB modulation. The novel scheme's preference is excelled than the conventional scheme's, which is demonstrated by these theory and simulation.

This paper introduces two schemes to generate millimeter-wave (mm-wave) BPSK and M-ASK signals in Base Station in Radio over Fiber (RoF) system based on cascaded Dual-Drive Mach-Zehnder modulators (DD-MZM). A BPSKmodulated mm-wave RoF system is proposed by controlling the amplitudes of baseband digital signal. Another scheme to generate M-ASK-modulated mm-wave in BS is based on commanding the power of Intermediated Frequency (IF) signal. Both schemes can make the remote local reference signal available in the BS, which makes the system costefficient. The proposed schemes are verified by theoretical derivation and simulation.

In this paper two different paradigms to realize a scalable all-optical packet switch with label swapping will be reviewed. All the functions required for switching the packets are based on all-optical signal processing without any electronic control. This allows very low latency and potential photonic integration of the systems. We report for both techniques experimental results showing the routing operation of the 160 Gb/s packets and beyond. We will discuss and compare both techniques in term of devices and bit-rate scalability, latency, power consumption, power penalty performance and cascadability as key parameters for the realization of an all-optical packet switch.

Rapidly growing demands triggered unparallel needs for fast-reliable-reconfigurable, and secure networks. Current solutions are govern by the reappearance of electronic bottleneck which imposes many limitations. To fulfil these demands will require low cost, high speed, high granularity access networks which can not be built using merely current technology and approaches to the data and signal processing. It is believed ultrafast all-optical signal processing may offer some alternatives by supplementing the lack of high-speed electronics. In this paper, we will review some advanced previously demonstrated all-optical techniques which may help to address the electronic bottleneck which is the major limiting factor in communication networks today.

We demonstrate a reader for digital lightpath labels that can be clipped onto a fiber to allow identification of lightpaths in the fiber without disturbing service traffic. Label error-rate measurements of labeled signals at ~10Gb/s confirm that simple fiber bends serve as effective clip-on power taps.

A simultaneous optical signal extracting and erasing scheme is presented by utilizing four-wave mixing (FWM) in optical fiber. Two synchronized gate and negation gate pulses are used as control signal, which are coupled with data stream and produce FWM effect in high nonlinear fiber (HNLF). Through filtering the idler light generated between each control signal and data stream, simultaneous signal extracting and erasing functions can be obtained. Experimental results are successfully demonstrated with NRZ data format at 10 Gbit/s. This system is able to operate at higher bit rates and other data modulation formats.

A all optical header extraction scheme based on semiconductor optical amplifier and Mach-Zehnder-Interferomete (SOA-MZI) with asymmetric control light was proposed and demostrated in this paper. By theoretical analsys, we found the proposed scheme can implement the optical pakcet header extraction more effectively. Numerical analsys and simulation show that more than 18dB contrast ratio of the extrcacted header at 20Gbit/s and the suppressed payload at 80Gbit/s can be achieved. In addition, the parameters of the SOAMZI and the input packet are discussed to optimize the performance of the scheme and the results also indicate that the performance of the scheme is influenced greatly by the parameters of SOA.

An all optical header extraction scheme based on semiconductor optical amplifier and Mach-Zehnder-Interferometer (SOA-MZI) with asymmetric control light was proposed and demonstrated in this paper. By theoretical analysis, we found the proposed scheme can implement the optical header extraction more effectively. Numerical analysis and simulation show that more than 18dB contrast ratio of the extracted header at 20Gbit/s and the suppressed payload at 80Gbit/s can be achieved. In addition, the parameters of the SOA-MZI and the input packet are discussed to optimize the performance of the scheme and the results also indicate that the performance of the scheme is influenced greatly by the parameters of SOA.

Different options for the evolution of a WDM PON toward 10 Gb/s channel rate are compared and experimentally investigated. Although today 10 Gb/s is not yet cost effective for a fiber access network, it can become a viable option in the next years, pushed by the demand of new, bandwidth hungry services and enabled by recent technology advancements.

We propose an optical switch control procedure for high-performance and cost-effective 10 Gbps Active Optical Access System (AOAS) in which optical switches are used instead of optical splitters in PON (Passive Optical Network). We demonstrate the implemented optical switch control module on Optical Switching Unit (OSW) with logic circuits works effectively. We also propose a compact optical 3D-CSP (Chip Scale Package) to achieve the high performance of AOAS without losing cost advantage of PON. We demonstrate the implemented 3D-CSP works effectively.

This work proposes a new wavelength converted broadcast-selective (WCBS) buffer structure. It can deliver each packet to different delays, which enables flexible management of the buffered packet flows. By rearranging the packets leaving queue, the highest priority packets can leave first and in preemption process without being affected by the former queue. Furthermore, we use the WCBS buffer structure to construct an OPS node in output-buffered pattern. The packet loss and delay performances are studied under different traffic loads at different WCBS dimensions. The result shows that the performance of high priority packets is improved effectively.

A novel scheme entitled as Additive Increase Multiplicative Decrease control with the Number of Burst Control Packet (AIMD-NBCP) is proposed. The basic idea of AIMD-NBCP is to count the number of the burst control packet (BCP) at the core node and estimate the average load by which to adjust the burst transmitting rate of the edge node with the AIMD scheme. Simulation results show that AIMD-NBCP scheme is efficient at keeping the network stable and enhancing the performance of deflection routing in terms of good-put, burst loss probability as well as fairness index.

We present a CLOS architecture for optical packet switching that allows for large throughput, low latency and contention resolution in the optical domain. The switch is scalable in terms of bit-rate and data-format and port-number. We describe devices and subsystems that support this architecture.

Novel all-optical implementations for logic AND and XOR gates are experimentally demonstrated at 2.5Gb/s based on cavity-enhanced four-wave-mixing (CE-FWM) combined with optical injection locking phenomena in monolithically integrated semiconductor ring lasers (SRLs). All-optical AND gate with simultaneous wavelength conversion have been achieved using the CE-FWM. XOR function is demonstrated using two parallel SRLs, each performing one part of XOR Boolean equation A ⊕B = A B¯ + AB¯ . Error-free operation with extinction ratio higher than 10dB confirms the suitability of SRLs for all-optical logic applications.

We propose and demonstrate a multi-level all-optical format conversion from NRZ signal to RZ signal, using a semiconductor optical amplifier (SOA) and a detuned optical filter. The three-level NRZ signal and a synchronous optical clock signal are injected into an SOA, which acts as a nonlinear element and causes the spectrum of the input NRZ signal to be broadened due to the cross phase modulation (XPM) effect. A detuned filter is thus used to extract the specific part of the broadened spectrum to obtain three-level RZ signal. The proposed conversion can be achieved with error free at 10G baud rate.

We propose a method to realize frequency up-conversion of UWB monocycle pulse using pulsed-pump fiber optical parametric amplifier (OPA). The spectrum of the amplified signal contains many discrete frequency components which are separated by the modulation frequency of the pump. Each frequency components contain the same spectral information as that of the original signal. By selecting the first-order or higher-order frequency components of the amplified signal and beating in the photodetector, up-converted signal at different frequencies are obtained. We demonstrate frequency up-conversion of baseband UWB monocycle pulse from 3-GHz to 19-GHz in the experiment and frequency up-conversion of pseudo-random binary sequence (PRBS) signal from 3-GHz to 60-GHz in the simulation.

A microwave photonic interference mitigation filter is proposed and experimentally demonstrated. The structure is based on a recirculating delay line loop comprising a semiconductor optical amplifier (SOA) and a tunable narrowband optical filter. Converted signal used as negative tap is generated through wavelength conversion employing cross-gain modulation of amplified spontaneous emission spectrum of the SOA. The converted signal circulating in the RDL loop realizes a high quality factor (Q) response after photo-detection. A bandpass response with negative coefficients combined with a broadband allpass response achieves a notch response with flat passband.

A novel approach is proposed to reduce the cross phase modulation between multiple clock pump in the multi-wavelength optical decision based optical fiber parametric amplifier. The dispersive media is used to preprocess the pump clock, and then the clock pump is changed from intensity modulation into phase modulation. In our experiment, cross phase modulation is obviously restrained between the adjacent channels.

The architectures of wired/wireless signal transmission based on RoF access network are proposed and their experimental results are shown. Both a wired and a wireless signals with a same data and heterogeneous data in different schemes are simultaneously generated using an injection locking technique and optical carrier suppression modulation based on a Fabry-Perot laser as a cost-effective WDM optical source. A reflective semiconductor optical amplifier (RSOA) is employed for the purpose of a colorless and simple base station. 20-km optical transmissions of 1.25-Gb/s wired and wireless data are experimentally demonstrated to verify the proposed RoF systems.

Visible light emitting diode (LED) wireless system based on the existing optical access network is proposed for the first time. Electroabsorption transceiver (EAT) is used to construct LED wireless optical link with optical access network. 5Mbps downlink and uplink transmission is verified through 23Km standard single mode fiber (SSMF) and 40cm wireless channel experimentally.

Compared with time division multiplexing based passive optical network (TDM-PON), wavelength division multiplexing based passive optical network (WDM-PON) is a better technology for the wider bandwidth and one to one connection provided. An efficient evolution from TDM-PON to WDM-PON without changing the current PON is needed. We propose a novel evolution from TDM-PON to WDM-PON based on DPSK/NRZ orthogonal modulation. No change for ONU of the existing TDM-PON is required. The good performance and feasibility of the method after a 20-km transmission is experimentally demonstrated.

Proposed in this paper is a novel flexible scheme of remote node for the full-duplex single fiber transmission systems. Simulations are conducted to validate that proposal based system realizes transmission with baseband signal in downstream and OSSB signal in upstream.

A kind of dynamic wavelength and bandwidth allocation (DWBA) algorithm of WDM-EPON was studied and then improved to guarantee QoS of multiple-services access. MPCP extension with modified REQUEST message and GRANT message are also proposed.The simulation results based on OPNET indicates the improved algorithm performed well in successful QoS assurance.

The simultaneous wired and broadcasting signals transmission based on WDM-PON access network are proposed and their experimental results are shown in this paper. ASE source modulated by broadcasting signal is provided for channels of wired signals. The filtered ASE source is aligned to each channel and modulated by the wired signal at RSOA. Simultaneous downstream transmission of wired and broadcasting signals over 23-km SMF is experimentally verified and mutual interference between wired and broadcasting signals is analyzed.

In this paper, a packet format employed in optical packet switching (OPS) networks is proposed. And the performance of block synchronization of the burst-ode optical packets is analyzed. Based on the analysis, the length of overhead synchronization bits is optimized to achieve high framing efficiency, which can enhance bandwidth efficiency.

Survivability of optical burst switching (OBS) network was concerned. Bidirectional forwarding detection (BFD) protocol which had the ability of fast failure detection was adopted in the failure detection for OBS links and nodes. Moreover, a BFD-triggered failure detection mechanism combines with Fast Reroute (FRR) is proposed and simulated. Simulation results show that the solution stated above can deal with both link failure and node failure in OBS networks.

A novel adaptive assembly algorithm called Cross-counter Balance Adaptive Assembly Period (CBAAP) is proposed in this paper. The major difference between CBAAP and other adaptive assembly algorithms is that the threshold of CBAAP can be dynamically adjusted according to the cross counter and step length value. In terms of assembly period and the burst loss probability, we compare the performance of CBAAP with those of three typical algorithms FAP (Fixed Assembly Period), FBL (Fixed Burst Length) and MBMAP (Min-Burst length-Max-Assembly-Period) in the simulation part. The simulation results demonstrate the effectiveness of our algorithm.

In this paper, a selective burst discarding scheme for deflection routing in OBS networks is proposed. The discarding scheme is based on the number of traversed hops, the number of remaining hops and burst size. The proposed scheme controls superfluous deflection in order to use the network resource efficiently. Simulation results show that the scheme can achieve a high performance.

In order to meet the ever-increasing data and mobility of the access network, we propose a hybrid wire and wireless network. Both high data rate of 10 Gb/s NRZ and OFDM signals were investigated. Results show that less than 1 dB power penalty at the BER measurement of a 25 km standard single mode fiber (SMF) transmission without dispersion compensation.

A multicast scheme in Wavelength-Division-Multiplexing passive optical network (WDM-PON) is realized by superimposing multicast differential phase shift keying (DPSK) signal onto downstream point-to-point Manchester/NRZ signal. A commercial electrical logic XOR gate is used to produce Manchester signal. The proposed system is experimentally demonstrated with 5-Gb/s point-to-point Manchester/NRZ format signal and 1.25-Gb/s DPSK signal. At the receiver side, a Mach-Zehnder delay-interferometer (MZDI) and a low pass filter (LPF) are employed to retrieve the multicast DPSK signal.

We propose and experimentally demonstrate a protection apparatus in ring-architecture time-division-multiplexed passive optical network with self-healing mechanism against the fiber fault promptly. Besides, the proposed system performances are also measured and analyzed.

Recently research on WDM PON has been focused on full-duplex bi-direction propagation in single fiber transmission. As a result the distortion and impairment caused by Rayleigh backscattered light and the reflection of each component cannot be neglected. We present a new flexible scheme of Colorless ONU with RSOA-Assisted Michelson Interferometer (MI), for Rayleigh Backscattering suppression in single fiber WDM-PON systems by generating Optical Single Side Band (OSSB) or Optical Carrier Suppressed (OCS) format signal at upstream. Simulations and discussions have been conducted to validate the proposal. The results are useful for designing cost-effective Multi-Wavelength PON or Point-to-Point Radio over Fiber system.

A 10GHz clock extracted from 80Gb/s OTDM signals is demonstrated experimentally based on Stimulated Brillouin scattering. The recovered clock exhibits excellent stability. In addition, the presented clock recovery scheme is easy to extract clock with any or multiple tributary rates, and the 40GHz and full 80GHz clock are subsequently recovered in simulation.

Single-carrier frequency domain equalization (SC-FDE), as a mature technique in wireless commutation, is widely researched for signal equalization and compensation utilizing high speed electronic devices such as DSP. In this paper, for the first time, we propose and demonstrate a SC-FDE technique in optical communication system. By adopting SCFDE technique, after 50km, 80km, 100km transmission for a 10Gb/s ASK single sideband (SSB) signal on a single mode fiber (SMF), the dispersion of optical signa1 is effectively compensated. The SSB signal is generated by a dual-arm Mach-Zehnder modulator (MZM) cascading a phase modulator (PM) based on Hilbert Finite impulse response (FIR) digital filter. The results demonstrate, in our proposed SC-FDE model, the eye opening and the clock recovery is improved, and the effect of compensation is enhanced as the transmission distance increases.

Differential phase-shift keying (DPSK) modulation format has been widely employed in high speed long haul transmission systems. The DPSK modulation format is a promising technique to increase the system performance, for example: extending the transmission distance, reducing the OSNR requirement and suppressing the nonlinearity. So the optical signal performance monitoring is quite necessary and indispensible for this modulation format. In this paper, a chromatic dispersion (CD) monitoring method for DPSK signals has been demonstrated using an optical delay interferometer with imperfect phase adjustment. We found the eye diagrams shown inclination with the combination effect of phase error and residual CD. This method can not only be used for CD monitoring but also can discriminate dispersion polarity. The CD monitoring range could cover from -425 to +425ps/nm of 10 GHz NRZ-DPSK system.

The OSNR margin improvements due to FEC limit elevation in 112Gb/s PDM-DQPSK systems under two different receiver schemes, direct detection and coherent detection with digital signal processing (DSP), are analyzed. FEC limit elevation provides more benefit to the direct detection scheme than to the coherent detection scheme in 112Gb/s PDMRZ-DQPSK system. For the cases of non-ideal receivers employment and long haul fiber link transmission, FEC performance improvement is more profitable than that for the ideal BtB case.

We compared the bit-error-rate (BER) and throughput performance of coherent time-spreading (TS) optical code division multiple access (OCDMA) systems with PPM (pulse-position modulation) and OOK (on-off keying) signaling. The results indicate that PPM-OCDMA with an appropriate pulse position multiplicity performs better than OOK-OCDMA in the receivers' bandwidth limitation case.

In this paper, experimental and theoretical investigations of TOAD system performance are implemented. 80Gb/s to 10Gb/s all-optical demultiplexing by the TOAD is demonstrated successfully. Based on the travelling wave model, the non-ideal problems encountered in the experiment are analyzed. The results show the position of SOA is no need to be strictly placed beside the loop midpoint, due to the switching window can be periodically generated with asymmetry values of SOA, and the position of switching window can be adjusted by varying the temporal deviation between counter-propagation signals. Furthermore, the unperfected coupling ratio and the synchronization jitter between the signal and control are crucial limits for the demultiplexing application in 160Gb/s OTDM system.

The effect of Mach-Zehnder modulator (MZM) direct current (DC) extinction ratio on single sideband (SSB) radio over fiber (RoF) link is analyzed and simulated, finite extinction ratio and drive signal unbalanced ratio of dual-drive MZM (DD-MZM) are considered. Our results show that power variation of target signal increase and harmonic suppression ratio decrease with the decline of extinction ratio. However, for a nonideal modulator with low extinction ratio, optimum drive signal unbalanced ratio could be applied to obtain a high quality microwave signal.

Based on 40Gbit/s Return-to-zero differential phase-shift keying (RZ-DPSK) DWDM system, the four dispersion management programs are proposed to suppress optical pulse phase jitter induced by kerr nonlinear effect including the self-phase modulation (SPM) and cross- phase modulation (XPM). The tolerance against the kerr nonlinear effect is studied through detailed numerical simulation by Split-Step Fourier Method. The results are showed that there are more than 2dB power penalty difference for different channels in a program and 5dB power penalty difference for different compensation programs. Through appropriate selection, DPSK system must be having 2dB performance improvement.

In this paper, we utilize a new study method that is under independent case of multiple optical orthogonal codes to derive the probability function of MOOCS-OPS networks, discuss the performance characteristics for a variety of parameters, and compare some characteristics of the system employed by single optical orthogonal code or multiple optical orthogonal codes sequences-based optical labels. The performance of the system is also calculated, and our results verify that the method is effective. Additionally it is found that performance of MOOCS-OPS networks would, negatively, be worsened, compared with single optical orthogonal code-based optical label for optical packet switching (SOOC-OPS); however, MOOCS-OPS networks can greatly enlarge the scalability of optical packet switching networks.

A direct detection-optical orthogonal frequency division multiplexing (DD-OOFDM) system is studied. The nonlinearity impact of Mach-Zehnder modulator (MZM) on DD-OOFDM transmission is analyzed. Signal to noise ratio (SNR) after MZM is calculated by theory, and proved by simulation. The nonlinearity impact of fiber is analyzed. Pilot-assist (PA) approach based on linear interpolation algorithm is used to smooth channel response and the optimal fiber launch power and MZM modulated depth in this system is obtained.

Based on one-coincidence sequence (OCS), bipolar one-coincidence sequence (BOCS) is constructed for coherent 2-D optical code-division multiple access system (OCDMA). Compared with OCS, BOCS has much larger code cardinality under the same code length and the same wavelengths. Furthermore, BOCS has the same cross-correlation and auto-correlation performance, and has lower crosstalk value, which can enhance performance of 2-D OCDMA system.

In coherent time-spreading OCDMA, beat noise is the main performance limitation. We propose a novel quasisynchronous coherent time-spreading optical CDMA, where the synchronization among users can be controlled within permissible time delay. Within permissible time delay, the mean intensity of aperiodic cross-correlation equals zero, and the beat noise can be removed. Simulation of four simultaneous users employing LA sequence (156, 8,16) is demonstrated, where the fiber link consists of 10km single-mode fiber, plus 1 km Dispersion Compensating Fiber. It can be seen that, beat noise can be removed in quasi-synchronous coherent time-spreading optical CDMA system.

The effects of pulse shape and pulse width on chip-asynchronous coherent time-spreading OCDMA system are evaluated by the aperiodic cross-correlation function of up-sampled sequence. It can be shown that for chip-asynchronous coherent time-spreading OCDMA, the shorter optical pulse will obtain better aperiodic cross-correlation properties, which can reduce multiple-access interference (MAI) and beat noise (BN). For 127-length Gold sequence, normalized throughput of OCDMA is derived. It can be shown that the shorter optical pulse will obtain larger normalized throughput.

Based on the analysis of the characteristic of ultrashort optical soliton, third-order dispersion (TOD ) and Raman scattering influences were taken into account. The quantitative relationship between the self-frequency shifting with distance, pulse width and TOD are found out using the Numerical simulation and fitting method. The size of the pulse time delay is also studied caused by the self-frequency shifting. The results are obtained the quantitative relationships between the self-frequency shifting and pulse time delay with distance, pulse width and TOD, which are very important for the high-speed optical switch.