DSP-based optical access approaches for enhancing NG-PON2 systems

Signal Shaping to Achieve OOK and PSK Co-Existence for Improved Optical Access Network Performance

“…Given that the OLT sends different data streams to multiple ONUs simultaneously, the data streams must avoid interfering with each other. In constellation sharing, a multilevel modulation signal such as OOK+ASK is used as the modulation format, and a number of bits allocated to a symbol differs with the ONU [10,11]. In the example of 8-star OOK+ASK, each symbol represents 3 bits, where 1 bit is allocated to the amplitude level of the symbol, and 2 bits to the phase information, as shown in Figure 1.…”

“…These modulation formats can co-exist with the existing PON which utilizes On-off Keying (OOK), which allocates 1 bit to 1 symbol. Co-existence can be achieved by constellation sharing; one terminal uses the OOK format (amplitude modulation) to send one data stream while another terminal uses an advanced modulation format to sends different data streams (amplitude and phase modulation) [10,11]. For example, one data stream is transferred by the amplitude information of the modulation format, whereas the phase information expresses the other data stream.…”

Polynomial-Time Complexity Large-Signal Model for DML-Based OOFDM Transmission Systems

“…Equation (3) bears two underlying approximations: the phase modulation due to the transient chirp is considered linear with respect to the optical power (2), and the linearization of the square root in the equation of electric field envelope [6, eq. 3].…”

“…By inspecting (4) we can see that the argument of the Bessel function is always u k ≡ 2m F M k sin θ k . Calculating m F M k as in (2) and using the range of values shown in Table I, it is easy to show that |u k | < 0.6. For this range of u k , the Bessel functions of first kind can be accurately represented by few terms of their Maclaurin series expansion:…”

“…As a matter of fact, for OOFDM signals composed of hundreds of subcarriers, e.g., [1], [9], these large-signal models become prohibitively time-consuming due to their inherent exponential-time complexity with respect to the number of subcarriers. Considering that potential future applications of those models lie also in equalization and interference cancellation techniques (see for instance [2], [3]), efforts toward models combining high accuracy and low complexity are paramount in reducing both hardware costs and power consumption in digital signal processor (DSP).…”