A method for receiving and processing a UWB preamble signal

Abstract

The application relates to a UWB preamble signal receiving processing method, which comprises the following steps: receiving a discrete digital signal obtained through sampling conversion after front-end processing; in a coarse acquisition stage, a locally reproduced preamble sequence is used to perform convolution processing on the discrete digital signal, coherent accumulation and fast Fourier transform are performed on the signal after convolution processing to obtain a first processing signal, a preamble is detected according to the strongest peak value response of the first processing signal, and frequency offset coarse estimation and CMR coarse estimation are performed when the preamble is detected; in an accurate receiving stage, the signal after convolution processing is subjected to frequency offset coarse compensation processing by using the frequency offset coarse estimation to obtain a compensation signal, interpolation processing is performed by taking the compensation signal as input, coherent accumulation and fast Fourier transform are performed on the signal after interpolation processing to obtain a second processing signal, residual frequency offset estimation and CMR accurate estimation are performed on the second processing signal. The application can improve the reliability of preamble detection and acquisition.

Description

Technical Field

[0001] This invention relates to the field of UWB preamble signal processing technology, and in particular to a UWB preamble signal receiving and processing method. Background Technology

[0002] UWB systems are communication systems that use nanosecond-wide pulses as code chips. Due to the short duration and large bandwidth of UWB pulses, they possess excellent multipath signal discrimination capabilities, enabling precise measurement (centimeter-level) of electromagnetic wave flight time in complex environments such as indoors, and facilitating high-precision ranging and positioning. UWB systems typically have a bandwidth of 500MHz or higher, and their ADC sampling frequency is usually at least 1GHz.

[0003] The UWB preamble consists of a periodically repeated sequence of pseudo-random code (PRN) modulated by the aforementioned pulses, and each such modulated PRN sequence is called a preamble. The PRNs used in UWB preambles possess perfect cyclic autocorrelation characteristics; that is, they exhibit a maximum correlation peak when the time difference between the two functions involved in the correlation operation (both being the PRN itself) is zero, and zero otherwise. Choosing such a PRN and combining it with wide-bandwidth pulse modulation reduces multipath interference to an extremely low level, which is highly beneficial for improving the estimation performance of electromagnetic time of arrival (TOA) and greatly simplifies the channel estimation algorithm. Overall, the carefully designed UWB preamble significantly reduces the design complexity of UWB receivers and improves the performance of UWB ranging and positioning systems. According to the UWB physical layer protocol, the PRN used for the UWB preamble is 31 or 127 units long. After zero-placing (with a zero-placing ratio of 4, 16, or 64), the actual length of the preamble sequence is doubled. For example, a PRN of length 31, after 16 times zero-placing, results in a preamble sequence length of 496 units, which is 16 times longer. The possible values ​​of the sequence are {1, 0, -1}, each corresponding to a chip. Each chip is modulated by a selected narrow pulse (usually a Gaussian pulse) to obtain the preamble. The UWB preamble signal can be composed of 16, 64, 1024, or 4096 consecutive preambles. The UWB receiver needs to detect and estimate the parameters of the preamble signal to achieve ranging, positioning, and subsequent data demodulation and decoding. Current UWB receivers process the preamble signal in two phases: acquisition and tracking. In the initial signal acquisition phase, the presence of the preamble needs to be determined. Once the presence of the preamble is confirmed, the receiver enters the signal tracking phase. During the tracking phase, the receiver needs to perform frequency offset compensation and resampling on the ADC data to keep the local signal and the received signal synchronized, creating conditions for subsequent correlation processing and coherent accumulation to obtain coherent gain. After the receiver completes the reception of all preambles, it processes the CMR (Channel Match Response) represented by the accumulated data to estimate the TOA (Time of Arrival) of the signal.

[0004] The receiver first locates the peak position of the CMR, which represents the location of the strongest path. However, the LOS (direct path) may not be the strongest path. Therefore, starting from the location of the strongest path, it searches backward for the first local peak position greater than a threshold, and determines the area near this peak as the region where the LOS signal is located. This threshold is usually set to a number of times the noise level.

[0005] After finding the approximate location of the LOS signal in the channel impulse response, the next step is to make an accurate time of arrival estimation. The receiver extracts several sample points before and after the local peak point in the CMR. The amplitude of these sample points should be a discrete sample of the UWB pulse envelope. The accurate estimation of the time of arrival is to match the correspondence between the discrete points and the pulse envelope in time.

[0006] However, acquisition or tracking based on a single preamble has a low signal-to-noise ratio, which is not conducive to reliable and effective acquisition, not conducive to accurate and reliable estimation of signal parameters, and not conducive to the detection of direct path signals and TOA estimation.

[0007] The technical terms used in this application are explained as follows:

[0008] UWB: Ultra-Wide Band

[0009] PRN: Pseudo RandomNumbers sequence

[0010] LOS: Line of Sight, direct path

[0011] NLOS: Non-Line Of Sight, also known as a reflection path.

[0012] TOA: Time of Arrival

[0013] RF: Radio Frequency

[0014] ADC: Analog to Digital Converter

[0015] CMR: Channel Match Response

[0016] FFT: Fast Fourier Transform. Summary of the Invention

[0017] The technical problem to be solved by the present invention is to provide a UWB preamble signal receiving and processing method that can improve the reliability of preamble detection and acquisition.

[0018] The technical solution adopted by this invention to solve its technical problem is: to provide a UWB preamble signal receiving and processing method, including the following steps:

[0019] Receive discrete digital signals obtained by sampling and conversion after front-end processing;

[0020] Coarse acquisition stage: The discrete digital signal is convolved using the locally reproduced preamble sequence. The convolved signal is coherently accumulated and subjected to fast Fourier transform to obtain the first processed signal. The preamble is detected based on the strongest peak response of the first processed signal. When the preamble is detected, the frequency offset and CMR are roughly estimated.

[0021] Precise reception stage: The coarse frequency offset estimation is used to perform coarse frequency offset compensation processing on the convolutional signal to obtain a compensated signal. The compensated signal is then used as input for interpolation processing. The interpolated signal is then subjected to coherent accumulation and fast Fourier transform to obtain a second processed signal. The second processed signal is then subjected to residual frequency offset estimation and precise CMR estimation.

[0022] The step of detecting the preamble based on the strongest peak response of the first processed signal specifically involves determining whether the maximum amplitude value in the first processed signal is greater than Th times the average amplitude value of all signals in the first processed signal. If it is greater, then the preamble is detected.

[0023] When performing coarse frequency offset estimation and coarse CMR estimation upon detecting a preamble, the frequency of the maximum amplitude in the first processed signal and its two adjacent frequencies are parabolically interpolated, and the frequency corresponding to the vertex of the parabola is used as the coarse frequency offset estimate; the frequency of the maximum amplitude in the first processed signal is used as the center setting window, and the window is used as the coarse CMR estimate, and no preamble will arrive at the receiver outside the window.

[0024] The convolution processing during the precise reception phase is performed within the window.

[0025] During the precise reception phase, through The frequency offset coarse compensation process is completed, where D1(n) is the compensation signal and D(n) is the signal after convolution. T is a rough estimate of the frequency offset. s This refers to the sampling interval of the analog-to-digital converter in the receiver.

[0026] The number of correlation accumulations in the coarse acquisition phase is less than the number of correlation accumulations in the precise reception phase; the number of Fast Fourier Transform (FFT) points in the coarse acquisition phase is less than the number of FFT points in the precise reception phase.

[0027] Beneficial effects

[0028] Due to the adoption of the above technical solution, the present invention has the following advantages and positive effects compared with the prior art: The present invention divides the preamble signal processing flow into two stages: coarse acquisition and precise reception, gradually improving the accuracy and reliability of the estimation of received signal parameters such as frequency offset, CMR and TOA, and adopting different coherent integration lengths and FFT configurations according to the accuracy of signal parameter estimation, so as to maximize the signal-to-noise ratio and accelerate the convergence speed of estimation accuracy. Attached Figure Description

[0029] Figure 1 This is a processing block diagram of an embodiment of the present invention;

[0030] Figure 2 This is a schematic diagram of the temporal relationship of the preamble convolution result;

[0031] Figure 3 This is a block diagram of the coarse capture stage in this embodiment;

[0032] Figure 4 This is a time-series diagram of the coherent integration results during the coarse acquisition stage in this embodiment;

[0033] Figure 5 This is a time-frequency relationship diagram of the FFT spectrum during the coarse acquisition stage in this embodiment;

[0034] Figure 6 This is a block diagram of the precise reception stage in this embodiment. Detailed Implementation

[0035] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0036] The embodiments of the present invention relate to a UWB preamble signal receiving and processing method, such as... Figure 1As shown, the process includes the following steps: receiving a discrete digital signal obtained by sampling and conversion after front-end processing; a coarse acquisition stage: convolving the discrete digital signal using a locally reproduced preamble sequence, performing coherent accumulation and fast Fourier transform on the convolved signal to obtain a first processed signal, detecting the preamble based on the strongest peak response of the first processed signal, and performing coarse frequency offset estimation and coarse CMR estimation when the preamble is detected; a precise reception stage: performing coarse frequency offset compensation processing on the convolved signal using the coarse frequency offset estimation to obtain a compensated signal, performing interpolation processing using the compensated signal as input, performing coherent accumulation and fast Fourier transform on the interpolated signal to obtain a second processed signal, and performing residual frequency offset estimation and precise CMR estimation on the second processed signal. Specifically:

[0037] The UWB electromagnetic waves received by the antenna are processed by the RF (radio frequency) front-end and then sampled by an ADC to convert them into discrete digital signals suitable for processing by digital ASICs, DSPs, or FPGAs. Next, a locally reproduced preamble sequence is convolved to improve the signal-to-noise ratio and prepare for further signal processing and parameter estimation. L is the length of the preamble in chip units, U is the upsampling factor, P(m) is the chip sequence of the preamble, and r(n) is the received ADC signal. Let n = Li + j, then we have:

[0038]

[0039] Since the preamble signal is periodic with respect to the preamble, the convolution output can be organized using phases and the preamble: each time a preamble period of convolution signal is processed, the next phase begins. Taking a preamble length of 496 chips and an upsampling factor of 2 as an example, the timing relationship is as follows: Figure 2 As shown.

[0040] Before initially connecting to a UWB system, a UWB receiver typically lacks accurate prior information regarding the frequency difference between itself and the transmitter. The UWB physical layer standard specifies a maximum permissible frequency difference of 20 ppm, based on the carrier frequency f. car Taking a 7987.2MHz carrier as an example, the frequency offset can be as high as approximately 160kHz. Using a preamble length of 0.99359µs as an example, the maximum number of preambles for direct coherent accumulation, N, is... max The value should be 3; otherwise, aliasing will occur, causing the frequency offset to be unreliable and potentially resulting in significant sincLoss attenuation.

[0041]

[0042] Frequency offset also causes differences in the sampling clocks at the transmitting and receiving ends, which limits the coherent accumulation time. Taking a chip of 2ns as an example, the maximum deviation of the sampling time in chip units for a preamble is:

[0043]

[0044] Excessively long coherent accumulation times (e.g., more than 50 preambles at a 20ppm frequency offset) will no longer yield significant benefits. However, TOA and CMR estimations require a high signal-to-noise ratio to achieve reliable accuracy, i.e., a longer coherent accumulation time. To address the limitation of frequency offset on the coherent integration length, this implementation divides the preamble signal reception processing into two stages: coarse acquisition and precise reception. When available prior information on the frequency offset is unavailable, long-term coherent accumulation is impossible, and extremely weak LOS cannot be reliably detected and TOA estimated. In this case, the preamble signal processing is in the coarse acquisition stage, where the presence or absence of the preamble is only detected based on the strongest peak response, and the frequency offset and CMR are coarsely estimated. Although the frequency offset estimate obtained in the coarse acquisition stage is not precise, it can already reduce the frequency error from the hundreds of kHz level to the kHz level after compensation, creating conditions for longer coherent accumulation. With the coarsely estimated CMR, subsequent signal reception can be performed within a time window, which is the portion of the preamble duration containing the useful signal, thereby reducing power consumption and improving processing efficiency.

[0045] While the possible frequency offset range of [-160kHz, 160kHz] can be divided into multiple segments during the coarse acquisition stage, and sampling clock offset compensation can be performed to improve the effective accumulation time and enhance the signal-to-noise ratio, this method of segmenting the frequency offset and compensating for it to increase the accumulation time is unacceptable in terms of processing complexity and power consumption, considering the sampling rate at the GHz level. For many power-sensitive mobile or portable UWB receiver devices, signal search and acquisition is the norm and also the state with the highest power consumption; therefore, reducing the power consumption of preamble search and acquisition is a key focus.

[0046] In this embodiment, the coarse capture stage is as follows: Figure 3 As shown. Coarse acquisition relies on the strongest path signal, and detection has much lower signal-to-noise ratio requirements than parameter estimation. Therefore, the smaller accumulation time in the coarse acquisition stage is sufficient for detection requirements compared to the precise reception stage. However, the signal-to-noise ratio provided by a single preamble is usually insufficient, requiring reasonable support for coherent accumulation of multiple preambles. Considering the tolerance for possible frequency offsets up to 160kHz, and taking into account signal processing gain, power consumption, processing efficiency, and flexibility, the number of coherent accumulations can be set to 1–3, with 4, 8, or 16 FFT input samples, padded with the same number of zeros as the samples, corresponding to FFT lengths of 8, 16, and 32. Let the coherent accumulation length be M, then the output of the i-th coherent accumulation of phase j is expressed as:

[0047]

[0048] Taking an FFT input sample size of 8 as an example, the data is as follows: Figure 4 As shown. A 16-point FFT transform is performed, and the spectral data, represented as complex numbers, is as follows. Figure 5 As shown. Then when B m_j A preamble is detected when the maximum amplitude is greater than the mean amplitude Th (a preset relative threshold) times. The frequency index of the maximum value can correspond to an approximate value of the frequency offset, and the phase index can correspond to the arrival time of the strongest path.

[0049] The FFT actually only calculates the spectrum at a set of discrete frequencies. The true frequency offset is usually not aligned with any of these discrete frequencies. If the frequency with the maximum spectrum is used as the frequency offset estimate, the accuracy will be low. Therefore, this implementation uses the maximum spectrum and its two adjacent spectra to perform parabolic interpolation to improve the frequency offset estimation accuracy. That is, the vertex of the interpolated parabola is used as the frequency offset estimate.

[0050] Let the spectrum with the maximum amplitude be B. bmax_pmax Therefore, the CMR is roughly estimated as follows:

[0051]

[0052] Where K is a pre-selected sufficiently large window width, such that outside the window centered on the spectrum with the largest amplitude, it can be assumed that no preamble will arrive at the receiver.

[0053] The process of the precise reception stage is as follows: Figure 6 As shown, during the precise reception phase, the Preamble convolution can be performed within the selected window, that is, convolution processing is only performed on a portion of the phases, thereby reducing computational load and saving resources.

[0054] Let the frequency offset estimate in the coarse acquisition stage be... The coarse frequency offset compensation process is as follows:

[0055]

[0056] T s Here, f is the ADC sampling interval, and f is the sampling rate. s The reciprocal of . The sampling rate corresponding to the frequency offset estimate is:

[0057]

[0058] The corresponding sampling interval is:

[0059]

[0060] In order to obtain For sampling points with a spacing, interpolation is required using the compensation signal D1(n) as input, which can typically be achieved using a Farrow Filter.

[0061] The subsequent processing, from coarse accumulation to residual frequency offset estimation, is similar to coarse acquisition. Since the residual frequency error is in the kHz range after coarse frequency offset compensation, the coarse accumulation time and the number of FFT points can be further increased. For example, coarse accumulation with 4, 8, or 16 preambles can be performed; the number of FFT input samples can be selected as 8, 16, or 32, corresponding to 16, 32, or 64 FFT points, to achieve higher signal-to-noise ratio and frequency offset estimation accuracy. Adding the residual frequency offset estimate to the coarse frequency offset estimate yields a more accurate frequency offset estimate, further reducing the error to the 100Hz level. The improved signal-to-noise ratio makes the CMR estimation less affected by noise and allows potentially weak LOS signals to be significantly larger than the noise and identified and estimated after FFT processing. Let bmax represent the frequency index of the maximum amplitude spectrum; then the LOS signal is the earliest phase-wise spectrum at bmax that is Th times greater than the noise, and its phase index is denoted as pmax.

[0062] Since the signal sampling rate is only twice the signal bandwidth, the estimation accuracy of the TOA phase using pmax is low. It is usually necessary to use three spectra for interpolation, such as the Early-Minus-Late algorithm.

[0063] The preamble signal can sustain up to 4096 preamble codes. After the aforementioned accurate reception, there may be a sufficient number of preamble codes remaining for further accurate reception. Since the frequency offset estimation error is significantly reduced after one round of accurate reception, subsequent accurate reception can further increase the coherent integration, thereby improving the signal-to-noise ratio and making the estimates of TOA and CMR more accurate.

[0064] It's important to note that base station frequencies are typically more accurate than those of mobile devices. This means that the clocks of different base stations are very close to their nominal values, significantly less so than the clock dispersion of mobile devices. Therefore, historical frequency offset information from mobile devices communicating with base stations has high availability. In such scenarios, when the mobile device has usable historical frequency offset information, a coarse acquisition process similar to precise reception can be used. During the acquisition phase, windowing cannot be applied because synchronization with the transmitter has not yet occurred; applying historical frequency offset information to current reception may result in errors ranging from kHz to hundreds of kHz, and the coherent accumulation length should also be within a reasonable range.

[0065] It is not difficult to see that the present invention divides the preamble signal processing flow into two stages: coarse acquisition and precise reception. It gradually improves the accuracy and reliability of the estimation of received signal parameters such as frequency offset, CMR and TOA. Based on the accuracy of the signal parameter estimation, different coherent integration lengths and FFT configurations are adopted to maximize the signal-to-noise ratio and accelerate the convergence speed of estimation accuracy.

Claims

1. A UWB preamble signal receiving processing method, characterized by, Includes the following steps: Receive discrete digital signals obtained by sampling and conversion after front-end processing; Coarse acquisition stage: The discrete digital signal is convolved using a locally reproduced preamble sequence. The convolved signal is then coherently accumulated and subjected to a fast Fourier transform to obtain a first processed signal. The preamble is detected based on the strongest peak response of the first processed signal. When the preamble is detected, a coarse frequency offset estimate and a coarse CMR estimate are performed. Specifically, the frequency of the maximum amplitude in the first processed signal is used as the center to set a window for the coarse CMR estimate. Parabolic interpolation is performed between the frequency of the maximum amplitude in the first processed signal and its two adjacent frequencies, and the vertex frequency is used as the coarse frequency offset estimate. No preamble will arrive at the receiver outside the window. Precise reception stage: The coarse frequency offset estimation is used to perform coarse frequency offset compensation processing on the convolutional signal to obtain a compensated signal. The compensated signal is then used as input for interpolation processing. The interpolated signal is then subjected to coherent accumulation and fast Fourier transform to obtain a second processed signal. The second processed signal is then subjected to residual frequency offset estimation and precise CMR estimation. Specifically, 4, 8, or 16 preambles are coherently accumulated, and the corresponding FFT points are 16, 32, or 64, so that the LOS signal is significantly larger than the noise after FFT processing and can be identified and estimated.

2. The UWB preamble signal receiving and processing method according to claim 1, characterized in that, The step of detecting the preamble based on the strongest peak response of the first processed signal specifically involves determining whether the maximum amplitude value in the first processed signal is greater than Th times the average amplitude value of all signals in the first processed signal. If it is greater, then the preamble is detected.

3. The UWB preamble signal receiving and processing method according to claim 1, characterized in that, The convolution processing during the precise reception phase is performed within the window.

4. The UWB preamble signal receiving and processing method according to claim 1, characterized in that, During the precise reception phase, through Complete the coarse frequency offset compensation process, among which, To compensate for the signal, The signal after convolution processing. This is a rough estimate of the frequency offset. This refers to the sampling interval of the analog-to-digital converter in the receiver.

5. The UWB preamble signal receiving and processing method according to claim 1, characterized in that, The number of correlation accumulations in the coarse acquisition phase is less than the number of correlation accumulations in the precise reception phase; the number of Fast Fourier Transform (FFT) points in the coarse acquisition phase is less than the number of FFT points in the precise reception phase.

Images