定义：
缩写：EDC 在接收器处采用电子元件来消除光纤通信链路中的色散。

光纤通信系统会被光纤中的色散效应影响。色散会使光纤链路中的比特字节展宽和发生畸变，因此解码信号更加困难。色散补偿通常在光域中进行，也就是在光探测器之前。然而，也可以采用电子色散补偿放大，采用电子学方法来实现色散补偿的目的。
存在两种不同的方案实现电子色散补偿，需要区分它们，因为它们应用于不同类型的数据接收器：
对于直接探测的接收器，色散效应不可能真正移除，因为色散主要是依赖于频率的相位变化，而在探测过程中相位信息就丧失了。然而，还是存在一些方法当色散不是很强时可以减弱这一效应。一般来讲，这些方法是采用抽头延迟线均衡器（横向滤波器），其中电子输入信号经历了不同的时间延迟，然后被放大到适合大小后重新结合。单纯的线性均衡器技术即使在存在诸如自相位调制等非线性畸变的情况下也可以提高信号处理过程。也有非线性均衡技术。例如，非线性判决反馈均衡器（DFE）通过依赖于接收器中得到对于过去的判断可以部分补偿缺失的光谱信息。假设这种系统装置被细致的进行了优化，信号质量可以得到极大的提高，尽管真正的光学色散补偿没有完全被利用。根据基于数字或者类似信号处理过程的反馈，参数被自动改变来减小误码率。即使多模光纤（在短距离光纤链路中会采用）的模间色散效应也可以被减弱。
采用光外差探测（或者零差探测）的接收器会更适宜于采用电子色散补偿方法，因为相位信息并没有消失。将具有合适频率响应的电子滤波器作用于中频信号，可以移除色散效应。
采用有限的成本时，即使未被完全补偿的系统也是有用的。可以将比特速率或者传输距离提高，例如20%，甚至50%。电子色散补偿方法可以采用比较便宜的发射器（例如，直接调制的激光器，而不是采用外加调制器的系统），因此可以极大的节约成本。
电子色散补偿应用时遇到的一个技术难题是如何根据链路的特点而自动调节其参数，因为最理想的参数设置依赖于光纤链路的性质和发射器，而手动优化则性价比不高。尤其是多模光纤系统，最佳参数也会随时间变化。

Acronym: EDC

Definition: a method for mitigating the effects of chromatic dispersion in fiber-optic communication links with electronic components in the receiver

Systems for optical fiber communications can be affected by the effect of chromatic dispersion of the fibers used. Dispersion in a fiber-optic link broadens and distorts the features of the bit symbols, making it more difficult to decode the signal. Dispersion compensation is normally be done in the optical domain, i.e., before photodetection. However, there are also methods of electronic dispersion compensation, utilizing electronics for that purpose.

Two quite different approaches of electronic dispersion compensation have to be distinguished, which are applied to different types of data receivers:

• In a receiver with direct detection, the effect of dispersion cannot be really removed, because it is essentially a frequency-dependent phase change, and the phase information is lost in the detection process. However, there are methods which can at least mitigate the dispersion effect as long as it is not too strong. Typically, such methods rely on tapped delay line equalizers (transversal filters), where portions of the electronic input signal are subject to different time delays and recombined after amplification with suitable levels. Purely linear equalization techniques can improve the signal processing even in the presence of nonlinear distortions [3], such as those arising from self-phase modulation due to fiber nonlinearities.
• There are also nonlinear equalization techniques. For example, nonlinear decision-feedback equalizers (DFE) can partly compensate for lost spectral information by making decision thresholds dependent on past decisions made in the receiver. Provided that the settings of such systems are carefully optimized, the signal quality can be significantly improved, even though the full potential of true optical dispersion compensation can not be reached. The parameters may be adjusted automatically using feedback techniques based on digital or analog signal processing, minimizing the bit error rate. Even the effect of intermodal dispersion in multimode fibers (as used in short-distance fiber-optic links) can be mitigated.
• A receiver using optical heterodyne detection (or homodyne detection) offers a higher potential for electronic dispersion compensation, as the phase information is not lost. When an electronic filter with an appropriate frequency response is applied to the intermediate frequency signal, this can directly remove the effect of chromatic dispersion [1, 2, 4].

Even a non-perfect compensation system can have various benefits, obtained at limited cost. The bit rate or transmission distance can be increased, e.g., by 20% or even 50%. Alternatively, electronic dispersion compensation may allow the use of a cheaper type of transmitter (e.g. a directly modulated laser instead of a system with an external modulator) and thus lead to significant cost savings.

A technical challenge of great importance for applications is automatic adaptation of the parameters of the electronic dispersion compensator to the link properties, because the ideal parameter settings depend on the properties of the fiber link and the transmitter, and manual optimization is not cost-effective. Particularly in systems with multimode fibers, the optimum parameters may also drift with time.

Bibliography