Adaptive Digital Baseband Predistortion for RF Power Amplifier Linearization

Power amplifiers (PAs) are vital components in many communication systems. To be transmitted wirelessly, a signal must be amplified with high fidelity so as to account for attenuation through the channel or propagation medium. The linearity of a PA response constitutes an important factor that ensures signal integrity and reliable performance of the communication system. Unfortunately, PAs are inherently non-linear due to their technology limitations. A significant manifestation of this non-linearity is the excessive distortion of the transmitted signal which can result in symbol recovery errors at the receiving end. Overcoming the PA’s nonlinearity therefore poses a major engineering challenge. A popular approach to PA linearization is to use predistortion [1-3]. This concept implies the creation of a function that represents the inverse of the PA transfer function in order to linearize the overall gain from the input to the output. There are a plethora of linearization techniques that have been reported in literature. Feedforward linearization is a commonly used technique where the spurious, distorted PA output spectrum is modified via two complementary circuits: a) an input signal cancellation circuit, and b) a distortion or error-cancellation circuit [4]. Unfortunately, feedforward linearization has generally moderate power efficiency and suffers from limited bandwidth. Contemporary analog radio-frequency (RF) techniques such as diode predistortion that are primarily operated in an open-loop configuration are only moderately effective, but relatively simple to implement [5]. Currently, digital baseband adaptive predistortion is the popular choice as it permits application of a variety of corrective algorithms and ensures reasonable bandwidths for common cellular phone standards such as Code Division Multiple Access (CDMA) and Global System for Mobile Communications (GSM).