The hottest short wave communication radio based o

  • Detail

Development of short wave communication radio based on OFDM technology

Abstract: This paper studies the application of OFDM modulation technology in short wave communication, puts forward a design scheme of short wave communication radio based on OFDM modulation, and completes the design of software and hardware system. The test results of the experimental prototype show that the performance index of the system can meet the actual requirements

key words: short wave communication; Software radio; Orthogonal frequency division multiplexing technology

I. Introduction

short wave communication is widely used in the field of wireless communication because of its advantages of long communication distance, simple erection and convenient movement. Orthogonal frequency division multiplexing (OFDM) modulation has attracted more and more attention because of its fast transmission rate, high frequency band utilization and strong anti multipath ability. It has also been gradually applied to the field of short wave communication, replacing the original single carrier modulation and non orthogonal multi carrier modulation technology [1]. The short wave communication radio station based on OFDM modulation technology introduced in this paper adopts the idea of software radio, and takes DSP as the control and operation core to complete the OFDM modulation and demodulation of digital signals

II. System model and performance parameters of short wave communication radio

1 The block diagram of the experimental prototype of the basic system model

short wave communication radio is shown in Figure 1. The sender first enters the human voice data through the PC for voice coding and compression, and then transmits the compressed bit data stream to the digital transmitter through the RS-232 interface for OFDM modulation. Finally, the RF module converts the OFDM signal to the RF frequency band and sends it to the wireless channel; At the receiving end, the analog signal transmitted through the wireless channel is first received by the RF module, and then the signal is restored to the baseband signal in the digital receiver for synchronization and OFDM demodulation. Finally, the demodulated bit data is uploaded to the PC through the RS-232 interface, which decompresses and decodes the voice to restore the data to the idiom data

the system parameters of short wave radio are shown in Table 1. The main performance indicators are:

① during 4QAM modulation, the data rate reaches 11.25 Kbps on the 10 kHz signal bandwidth; At 16QAM, it can reach 22.5 Kbps

② when 4QAM modulation is adopted, the bit error rate in AWGN channel with signal-to-noise ratio of 10 dB can reach 10-5

③ when 4QAM modulation is adopted, the bit error rate can reach 10-4 in the short wave channel with a signal-to-noise ratio of 20 dB (the maximum delay of multipath channel is 4 ms)

limited to space, the following mainly introduces the software and hardware design of the two if processing modules of the digital transmitter and receiver, but does not introduce the RF module and voice codec module

III. software and hardware structure of digital transmitter

the structural block diagram of digital transmitter is shown in Figure 2. The compressed voice data is transmitted to the transmitter through RS-232 interface, buffered first, and then sent to DSP for OFDM modulation. Finally, the modulated signal is up converted to if and then sampled

Figure 2 structure block diagram of digital transmitter in the digital transmitter, we use tmsvc5410 chip of TI company to complete the OFDM modulation of the signal. This chip is a 16 bit fixed-point DSP, with 64 K 16 bit ram in the chip, and the maximum working clock can reach 100 MHz

dsp internal signal processing flow is shown in Figure 3. After constellation mapping (4qam/16qam) is performed on the data, the pilot is inserted. Because in our scheme, the data transmission adopts a frame structure, every 20 symbols is a frame, so it is necessary to insert a time pilot in the first symbol of each frame for frame synchronization at the receiving end, and insert a gain pilot in all symbols for signal synchronization and channel estimation at the receiving end, The amplitude of time pilot and gain pilot is times of the signal amplitude, and the phase is randomly distributed

since OFDM modulation can be equivalent to an IDFT, the modulated signal can be expressed as:

in order to eliminate inter symbol interference caused by multipath, cyclic prefix (CP) needs to be added. In fact, the import replacement length of companies with cyclic prefix focusing on medium and high-end elastomer varieties is generally greater than the longest delay time of the channel, and the final output signal is

in order to generate OFDM signals with a bandwidth of 10 kHz and a center frequency of 512 kHz, There are two schemes to choose from: ① it is generated directly during OFDM modulation in DSP. According to Nyquist sampling theorem, the discrete sampling rate at this time is at least 1.024 MHz, which means that the number of IFFT points during OFDM modulation is 16384 points. Although the hardware structure of this method is relatively simple, it has high requirements for DSP operation speed, and tmsvc5410 is not competent; ② Generate an OFDM signal with a discrete sampling rate of 32 kHz in DSP (I zarut banner and Zhejiang Liancheng Electronic Technology Co., Ltd. have reached a resonance on the aluminum deep processing project, and the number of FFT points successfully signed is 512), and then interpolate and filter it. In order to obtain a discrete sampling rate of 1.024 MHz, interpolate at least 32 times the original signal, and finally mix it to 512 kHz, In practice, in order to reduce the design requirements of analog filters after digital to analog conversion, we use 256 times interpolation. Although this scheme has a large amount of computation, it can be completed by using a special up conversion (Duc) chip. HSP50215 produced by Harris company is a single channel modulated up converter with a maximum input data stream of 3 MHz and an output data stream of 52 MHz. It includes a 32-bit programmable carrier digital oscillator (NCO), a 30 bit programmable symbol timing NCO, and a 256 order programmable shaping FIR filter. The maximum interpolation factor is 256, which can fully meet our design needs

IV. software and hardware structure of digital receiver

the design of digital receiver adopts the software radio model of IF bandpass sampling, and its structural block diagram is shown in Figure 4. First, the input if signal is bandpass sampled, and then the baseband signal is recovered by mixing, low-pass filtering and down conversion. Finally, OFDM signal synchronization, channel estimation and demodulation are carried out

1. Bandpass sampling

the input signal is a narrow-band signal with a center frequency of 512 kHz and a bandwidth of 10 kHz. In order to make the recovered OFDM signal consistent with the sampling rate of the transmitted signal to ensure that the actual frequency value corresponding to each subcarrier is consistent, it is first necessary to obtain a discrete baseband signal with a sampling rate of 32 kHz. There are two schemes to choose from: ① direct sampling. According to the sampling theorem, the sampling rate is at least 1.024 MHz. In order to obtain the required baseband signal, 32 times of extraction is required, and finally the required signal is filtered out with a low-pass filter. Considering the large amount of computation, this scheme can be achieved by using special down conversion (DDC) devices (such as HSP50214B of Harris company) like digital transmitters; ② Bandpass sampling, according to the bandpass sampling theorem and the above reasons, the sampling rate must be an integral multiple of 32 kHz. Because there is always carrier deviation in practice, the signal spectrum directly sampled with 32 kHz will produce aliasing, so the sampling rate adopted by this system is 96 kHz, and then the required baseband signal is recovered through mixing, low-pass filtering and 3-fold extraction. After comparing the two schemes, we adopt the latter, because the scheme does not need to use additional down conversion devices, the system structure is relatively simple, and the final signal processing computation is not too complex, which can be completed by DSP

dm signal synchronization and channel estimation

the processing of sampled signals is completed by ADSP21160 of share series of ad company. This chip is a 32-bit dual core floating-point DSP with 250 K 16 bit ram on the chip, and the maximum working clock can reach 80 MHz. It mainly completes the signal preprocessing, that is, the signal is restored to baseband signal, OFDM signal synchronization and channel estimation through mixing, filtering and extraction, and finally the original signal is restored by constellation inverse mapping

for an actual OFDM system, if the influence of time, carrier and sampling rate without synchronization and the random fading of the wireless channel on the signal are considered, the signal received at the receiving end can be written as:

where α l. K means sending signal, n ε Indicates symbol deviation, Δ F represents carrier frequency deviation, ξ Represents the sampling rate deviation, HL, K represents the channel transfer function, NL, K represents additive Gaussian white noise

in order to recover the original signal correctly, the signal must be synchronized and channel estimated first. The signal synchronization is divided into three steps (symbol synchronization, carrier synchronization and sampling rate synchronization). The synchronization algorithm flow is shown in Figure 5

(1) symbol synchronization

symbol synchronization can take advantage of the correlation between the cyclic prefix in each symbol and the signal. Considering that the symbols need to be tracked after coarse synchronization, the requirements for accuracy can be appropriately relaxed for coarse synchronization to reduce the amount of computation during coarse synchronization. We modify the maximum likelihood method (ML) [3] as follows:

where d represents the integer time offset estimate, l represents the length of the cyclic prefix, N indicates the length of valid symbols

(2) carrier synchronization [4,5]

the carrier deviation normalized by the subcarrier interval (62.5 Hz) can be divided into integer part and decimal part. The estimation of the integer offset can use the amplitude information of the pilot signal position in each OFDM frequency domain symbol, and its estimated value can be obtained by the following formula:

where CP represents the pilot set, and D represents the range of searching the integer offset is the front and rear 10 subcarrier intervals

decimal deviation estimation and tracking uses the phase rotation of the corresponding pilot position signal in the two adjacent OFDM symbols, and its estimated value can be obtained by the following formula:

where ng is the length of the cyclic prefix, n is the effective symbol length, α 2( φ k) It is the fading of signal amplitude caused by frequency offset. When the frequency offset is very small, the value is approximately 1, | HK | is the amplitude fading caused by the signal passing through the wireless channel

(3) sampling rate synchronization

the deviation of sampling rate will also cause the phase rotation of the signal, and the size of the rotation is related to the subcarrier number, so its estimation and tracking can also use the rotation value of the phase of the corresponding pilot position signal in the adjacent two OFDM symbols, and its estimation method is shown in formula (6). Since both sampling rate and carrier tracking can be attributed to the tracking of signal phase rotation, in practice, only a phase-locked loop is needed to track the phase change of the signal

(4) channel estimation

the signal will cause random fading of amplitude and phase after passing through the short wave wireless channel. Even if the signal is completely synchronized at the receiving end, the correct signal cannot be recovered without channel estimation. Generally, channel estimation methods can be divided into two kinds [6,7,8]: ① data assistance methods. The auxiliary data can be pilot or training sequence. The former is to insert some pilot signals into each or every several OFDM symbols before modulation, and the latter is to insert a certain length of training sequence at the beginning of each frame or every several frames of modulated OFDM signal; ② Blind estimation, using only the received signal for channel estimation. This paper adopts the channel estimation method based on pilot, and the specific algorithm flow is as follows:

1) for the received pilot signal, the corresponding

is estimated by using the following formula, where k represents the sub carrier sequence number, M represents the pilot sequence number, l represents the number of interpolation, and l=1... M

3) construct a special interpolation filter, which can ensure that the value at the non-zero position remains unchanged when filtering the signal, and use it to filter the signal to obtain the estimated value of the channel transfer function

v. measured signal graphics

the short wave communication radio developed by us includes two parts: digital transmitter and receiver

in the actual performance test, the short wave channel model we use is the model provided by the DRM standard. See literature [9] for the channel model and parameter settings. The short wave channel in Figure 6 refers to the third type of channel provided by the standard

Copyright © 2011 JIN SHI