In order to use the signal, people have to process it. For example, when the electrical signal is weak, it needs to be amplified; When mixed with noise, it needs to be filtered; When the frequency is not suitable for transmission, modulation and demodulation are required; When the signal encounters distortion, it needs to be balanced; When there are many signal types, identification is required.
signal processing Extract characteristic signals in the event change process, and remove interfere , analysis, synthesis, transformation and operation to obtain information that reflects the nature of event changes or is of interest to the processor. branch Analog signal processing And digital signal processing. [1] Digital signal processing began to develop in the 1960s, beginning with Bell Laboratories and Massachusetts Institute of Technology Simulation of circuit and filter design with electronic computer has established digital filter The foundation of development. In the mid-1960s, he invented fast Fourier transform The calculation speed of Fourier analysis of spectrum analysis has been increased by more than 100 times, so that the purpose of spectrum analysis can be achieved by using electronic computers, the practical foundation of signal and system analysis has been laid, and the basic methods and concepts of digital signal processing centered on digital filtering and fast Fourier transform have been formed. From the 1970s, the term digital signal processing came out in the field of science and technology. Reduce the redundant content in the signal; Filter out intermingled Noise and interfere ； Or convert the signal into easy to process transmission , analysis and identification form For subsequent processing. 
signal processing along with digital computer The rapid development of signal processing theory and method It has also been developed. In front of us, there is a pure mathematical processing that is not subject to physical constraints, namely algorithm, and the field of signal processing has been established. Now, for signal processing, people usually put analog signal Into digital signals, and then use efficient Digital Signal Processor (DSP: Digital Signal Processor) or computer Digital signal processing is performed on it. So, how to carry out digital signal processing?
Generally speaking, digital signal processing involves three steps:
（1） Analog digital conversion (A/D conversion): analog signal It is a process of discretization of independent variable and amplitude at the same time theory The guarantee is the sampling theorem. (II) digital signal processing (DSP): including transform domain analysis (such as frequency domain transformation) digital filtering , identification, synthesis, etc. (III) Digital to analog conversion (D/A conversion): restore the processed digital signal to analog signal 。 Usually, this step is not necessary. There are many successful examples as DSP, such as the invention of medical CT tomography scanner. It is an instrument that uses the phenomenon that various parts of the organism have different X-ray absorptivity, and uses the projection data of scanning in all directions to construct a detection volume profile. FFT in this instrument（ fast Fourier transform ）It plays a role in fast calculation. Later, we have successively developed: positron They have made great contributions to the medical field. Terminology related to signal processing:
one | adapter signal process (ASP) adapter signal processing |
two | process, adapter signal (ASP) adapter signal processing |
three | signal processing signal processing |
four | Native Signal Processing Built in signal processing |
five | Entropic Signal Processing System entropy signal processing system |
six | Advanced Signal Processing System advanced signal processing system |
seven | Adaptor Signal Process Adapter signal processing |
The basic contents of signal processing include transformation, filtering, modulation, demodulation, detection, spectrum analysis and estimation. Transformations such as type Fourier transform, sine transform, cosine transform, Walsh transform, etc; Filtering includes high pass filtering, low-pass filtering, band-pass filtering, Wiener filtering Kalman filtering , linear filtering, nonlinear filtering, and Adaptive filtering Etc; Spectral analysis includes the analysis of deterministic signals and random signals. Usually, the most common research is the analysis of random signals, also known as statistical signal analysis or estimation, which usually includes linear spectral estimation and nonlinear spectral estimation; Spectrum estimation includes periodogram estimation Maximum entropy spectral estimation Etc; With the complexity of signal types, when the signal to be analyzed cannot meet the conditions such as Gaussian distribution and non minimum phase, there is also a method of higher-order spectral analysis. High order spectrum analysis can provide signal phase information, non Gaussian information and nonlinear information; Adaptive filtering and equalization is also a major field of application research. Adaptive filtering includes horizontal LMS adaptive filtering, lattice adaptive filtering, adaptive cancellation filtering, and adaptive equalization. In addition, there are array signal processing and so on. Signal processing is the basic theory and technology of telecommunications. Its mathematical theories include equation theory, function theory, number theory, stochastic process theory, least square method and optimization theory. Its technical pillars are circuit analysis, synthesis and electronic computer technology. Signal processing is closely related to contemporary pattern recognition, artificial intelligence, neural network computing and multimedia information processing, which closely links basic theory with engineering applications. Therefore, signal processing is a subject with both complex mathematical analysis background and broad practical engineering prospects.
The development of signal processing is centered on digital signal processing. This is because signals can generally be expressed in digital form, and digital signals can be calculated or processed by software on an electronic computer. Thus, no matter how complex the operation is, as long as it can be analyzed mathematically and solved optimally, it can be simulated on an electronic computer. If the calculation speed is reasonably fast, it can also be completed in real time with a large-scale dedicated digital signal processing chip. Therefore, Digital signal processing technology It has become one of the most dynamic disciplines in the development of information technology. Signal processing is used in many important application fields with strong penetration. The engineering construction department is used to simulate the seismic performance of large building structures; Mechanical manufacturing industry is used to analyze the model of mechanical structure vibration, so as to improve the vibration performance and structure; Used in aircraft manufacturing to check the transmission characteristics and wear of engines; Space remote sensing is used for the classification of ground vegetation and the distribution of meteorological clouds, and in the medical field for the analysis and diagnosis of B ultrasound, X-ray films and physiological electrical signals; In the field of telecommunications and electronics, digital signal processing is the most direct application.
In the field of telecommunications, the most typical applications of digital signal processing are:
① Speech coding and compression. The digitized voice occupies a very wide frequency band. In order to carry out narrowband transmission and efficient storage, compression is required. Generally, a voice needs 64kbit/s bit rate. Medium speed coding requires that this code rate be compressed to 32kbit/s, 16kbit/s and even 8kbit/s, and still maintain good voice quality. adopt Digital signal processing technology , many adaptive coding schemes have reached Consultative Committee on International Telegraph and Telephone (CCITT). Low speed coding requires the bit rate to be reduced to 4.8 kbit/s, 2.4 kbit/s or even 800 bit/s, which has been well implemented by algorithms and hardware. ② Image coding compression. Whether still images or moving images, or even television images, the amount of data after digital coding is very large. For high-quality transmission, they generally need to be compressed to 1/10~1/100. Various coding methods, such as the so-called wavelet transform method and fractal signal analysis method, have proposed feasible schemes for high compression ratio television coding.
③ Design of shunt and combiner filter banks. The core of the technical realization of time-division/frequency division multiplexing equipment is the shunt filter bank. The design and implementation of shunt filter banks depend entirely on the digital filter This digital filter bank not only has unified performance, stability and reliability, but also Price performance ratio Very high. ④ Adaptive equalization and echo cancellation. In long-distance data communication, equalization and echo cancellation are essential. Using analog devices can not achieve the equalization that can meet the requirements of various channels, and only digital methods can guarantee its performance.
The application of signal processing technology has developed to a wide extent no less than the application of electronic computers. With the continuous discovery of algorithms and the birth of devices, signal processing will become a basic discipline that all telecommunication engineers need to be familiar with. [2]
