Digital pre-distortion circuit and method, and digital pre-distortion training circuit

A digital pre-distortion method carried out by a digital pre-distortion circuit and capable of compensating for a non-linear characteristic of an analog circuit, comprising the following steps:

• processing a digital feedback signal according to a modified Cholesky Decomposition algorithm to thereby generate a plurality of coefficients in which the digital feedback signal is derived from an output signal of the analog circuit and the output signal of the analog circuit is derived from an original digital signal;
• and
• processing the original digital signal according to the plurality of coefficients under a compensation mode to thereby generate a digital pre-distortion signal,
• wherein the non-linear characteristic of the digital pre-distortion signal is operable to compensate for the non-linear characteristic of the analog circuit and thereby make the output signal of the analog circuit conform to a predetermined characteristic, and the step of processing the digital feedback signal according to the modified Cholesky Decomposition algorithm includes:
• obtaining a pre-distortion transformation matrix according to the digital feedback signal, the order of a memory polynomial and the depth of a memory effect;
• doing a matrix calculation according to the conjugate matrix of the pre-distortion transformation matrix, the pre-distortion transformation matrix and the original digital signal to thereby obtain a calculation result;
• and
• processing the calculation result according to the modified Cholesky Decomposition algorithm to thereby obtain the plurality of coefficients.

A digital pre-distortion training circuit, comprising:

• a pre-distortion training circuit operable to process a digital feedback signal according to a modified Cholesky Decomposition algorithm and thereby generate a plurality of coefficients through carrying out at least the following steps:
• obtaining a pre-distortion transformation matrix according to the digital feedback signal, the order of a memory polynomial and the depth of a memory effect;
• doing a matrix calculation according to the conjugate matrix of the pre-distortion transformation matrix, the pre-distortion transformation matrix and the original digital signal to thereby obtain a calculation result;
• and
• processing the calculation result according to the modified Cholesky Decomposition algorithm to thereby obtain the plurality of coefficients,
• wherein the digital feedback signal is derived from an output signal of an analog circuit and the plurality of coefficients is operable to compensate for a non-linear characteristic of the analog circuit.