A method of simulating channels using digital signal processing technology

As an important part of the communication field, wireless communication has been increasingly widely used. In particular, with the continuous development of digital signal processing technology and the emergence of the concept of intelligent mobile terminals, wireless communication has become increasingly closely related to people's daily lives; therefore, all manufacturers have increased their investment in wireless communication research and development, and wireless communication products are emerging in an endless stream.
Due to the characteristics of wireless communication itself, in the process of developing wireless communication equipment, it is necessary to conduct repeated experiments and verifications in the actual communication environment of the field to ensure the performance and quality of the product. As we all know, in field experiments, a lot of manpower and material resources are required, and it is very difficult to solve and reproduce problems; therefore, it is also very difficult to ensure the development cycle.
In order to overcome the unfavorable factors brought about by field experiments, channel simulators have been increasingly widely used in the development of wireless communication equipment. This article proposes an implementation method based on digital signal processing for the development of channel simulators. Through this method of digital signal processing, the characteristics of wireless channels such as signal attenuation, delay and Doppler frequency shift can be simulated more accurately, and compared with the currently commonly used delay method, the method mentioned in this article can more effectively save processor resources.

1 Channel model
Different from wired channels, in addition to direct waves, wireless channels also have reflections, scattering and diffraction, and the signals will reach the receiving end along different paths and directions. At this time, the received signal will have a delay spread phenomenon; in addition, when the receiver is in a mobile state, the signals of each path will also produce different degrees of Doppler frequency shift, which will cause the signal to produce frequency domain dispersion.
Therefore, based on the multipath channel model and taking into account the time-varying nature of the channel response, the two-dimensional impulse response function of the wireless mobile channel can be expressed as:

In formula (1), ak, τk, φk, and fk are the fading factor, delay, phase and Doppler frequency deviation of the kth path, respectively, and L is the number of paths.
Let the input signal (transmitted signal) be, fc is the carrier frequency, then the output signal (received signal) y(t) is the linear convolution of the input signal and the channel impulse response, as shown in formula (2):

In formula (2), fc is the carrier frequency of the input signal.
It can be seen that the paths in the wireless channel are independent of each other. The multipath channel is formed by attenuating, delaying and Doppler shifting the signals of each path, and then superimposing the signals of each path.

2 Channel simulator design ideas
Generally, in wireless communications, the carrier frequency is very high, and it is impossible to directly digitally process the signal at this frequency. Therefore, the input signal is first down-converted to obtain the baseband signal, and the sampled signal is digitally processed after AD conversion to produce the multipath signal, and then the analog signal is generated by DA conversion, and finally the output signal is generated by up-conversion.
Figure 1 shows the channel system block diagram of the channel simulator.

Figure 1 shows the signal processing process of a channel. The processing part of down-converted and up-converted analog signals can be processed by general processing methods, which will not be elaborated in this article.
The selection of AD analog-to-digital converter should be based on the design requirements of the channel simulator for the signal bandwidth. According to the Nyquist sampling theorem, that is, fs>2f, the AD sampling rate must be at least twice the signal bandwidth; at the same time, the appropriate number of bits should be selected according to the accuracy requirements of the system.
Due to the limitation of the number of bits of the AD analog-to-digital converter, the attenuation range is about 60 dB in the case of 12 bits. The attenuation range can be increased by adding a digitally controlled attenuator at the analog end. However, the complexity of the implementation is also increased.

3 Design of channel simulation and digital processing
Channel simulation is mainly realized through the digital signal processing process. Figure 2 shows the system block diagram of the channel simulation and digital processing part.

As shown in Figure 2, in the channel analog digital processing, the interpolation filter first interpolates the baseband input signal, and the source data queue temporarily stores it. At the same time, the channel model operation unit stores the completed channel parameters (such as attenuation factor, delay, Doppler frequency deviation, etc.) in the channel parameter buffer; the channel simulation operation unit obtains the input signal, channel parameters and WGN for operation, and then sends them to the downsampling filter, and finally outputs the baseband signal.
In addition, the maintenance system, channel configuration system and monitoring system are background user interfaces, which can perform equipment maintenance (such as system upgrades, calibration compensation, etc.), channel configuration, operation status monitoring and control, etc., and can be designed according to actual conditions and needs, which will not be elaborated here. There are many materials on interpolation filters and downsampling filters, which will not be explained here, but in principle, the smaller the system delay generated, the better.
The other parts are explained below.
The main function of the source data queue is to buffer the interpolated baseband data, which is implemented using dual-port RAM. In order to improve the system's support capability and resolution for delay, in principle, the larger the queue capacity, the better. Take 5μs delay, 1ns resolution and 12-bit modulus as an example: (5 000÷1)×2=10 000 bytes. It can be seen that by increasing the capacity of the dual-port RAM, longer path delays can be simulated without occupying processor resources.
The channel model operation unit operates the configured model to generate channel parameters such as attenuation factor, delay and Doppler frequency deviation, and puts them into the channel parameter buffer. Due to the large amount of channel model calculation, DSP high-speed processor + FPGA is used as the main operation unit, and the channel model operation is optimized, depending on the specific supported channel model.
The channel parameter buffer is mainly used to store the calculated channel parameters. The buffer is divided into two blocks, using a ping-pong mechanism, and the size of each buffer space is calculated according to the data processing speed. The
channel simulation operation unit consists of a series of operation units, and its structural block diagram is shown in Figure 3.

In Figure 3, each operation unit is responsible for calculating the channel of a path. The number of operation units is the number of supported paths. Each operation unit is independent of each other and calculates in parallel (reducing system delay). The output result is obtained after summing. Among them, each operation unit obtains signal data from the input signal according to the delay parameter. In addition, each operation unit can be configured on demand according to the actual number of paths.
The above gives the overall design idea of ​​channel simulation digital processing. There are many things to consider in the specific implementation details, such as clock synchronization system, fixed-point accuracy, algorithm optimization, etc. It is best to simulate the system solution before making the implementation design as a reference for the implementation design.

4 Conclusion
The digital signal processing method is flexible in implementation, especially it can generate accurate delays, and can simulate multipath channel models, which brings great convenience to wireless communication R&D personnel; the use of digital signal processing methods is the development direction of wireless channel simulation. With the continuous upgrading of hardware technology, more and more complex channel models can be simulated.
Compared with the implementation method of simulating multipath through a delay device, the method used in this article can simulate longer delays and number of paths with lower device requirements.
However, this method currently also introduces a large system delay, generally around 2μs, which may cause some inconvenience in certain specific application scenarios.