An ultra-wideband, high-resolution, flexibly configurable spectrum analyzer
By using star-type taigabit LAN to build a distributed signal acquisition and spectrum computing system in the spectrum meter, the problem of insufficient flexibility and hardware resource utilization in the existing technology is solved, and the real-time ultra-wideband spectrum with high frequency resolution is achieved Analytics and optimized use of hardware resources.
Patent Information
- Application Number
- CN202410965578.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-18
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-07-18
AI Technical Summary
When implementing high-frequency resolution and ultra-wideband spectrum analysis, existing spectrum meters have limited flexibility and insufficient hardware resource utilization, which cannot meet the real-time analysis requirements of ≥100GHz.
The distributed signal acquisition and spectrum calculation system is constructed using star-type taigabit LAN. Through the subsystems of signal acquisition units, switches and spectrum calculation units, the distributed signal acquisition and calculation are realized, and the spectrum calculation and synthesis are used to ensure the hardware. Optimized use of resources.
Real-time analysis of ultra-wideband (≥100GHz) spectrum with high frequency resolution is realized, and the frequency bandwidth and time resolution are flexibly configured, which reduces hardware costs and supports the optimal utilization of hardware resources.
Smart Images

Figure CN118508998B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of radio astronomy, and in particular to an ultra-wideband high-resolution and flexibly configurable spectrum analyzer, which can be used for astronomical research in the terahertz frequency band and realizes ultra-wideband (≥100GHz) spectrum real-time analysis with high frequency resolution. It can also be applied to ultra-wideband real-time spectrum analysis in the fields of communication, measurement and control, and radio monitoring. Background Art
[0002] Radio astronomy mainly conducts scientific research in the time domain and frequency domain. Time domain science mainly studies how celestial radiation changes with time (such as the periodic changes in pulsar radiation and the explosive characteristics of fast radio bursts), and reveals the physical properties of celestial bodies based on such changes; while frequency domain science mainly studies how celestial radiation changes with frequency, and reveals the composition of the universe based on such changes.
[0003] Time domain scientific instruments are related to scientific goals. For example, research on pulsars requires folding the time domain signals and retaining only the folded pulse profile data; while research on fast radio bursts focuses on finding new fast radio bursts and recording the original data before and after the burst for offline analysis.
[0004] The main instrument in frequency domain science is the spectrometer. In order to obtain the fine structure of spectral lines and to accurately characterize the composition of the universe, a spectrometer with high frequency resolution is needed. In order to observe multiple spectral lines simultaneously to improve the efficiency of spectral observation, a spectrometer with high frequency bandwidth is needed. In order to improve the efficiency of searching for unknown spectral lines, a spectrometer with ultra-high frequency bandwidth is needed. For example, in the terahertz band, scientists expect the instantaneous frequency bandwidth to reach ≥100GHz.
[0005] The core of a spectrum analyzer is a customized device based on an analog-to-digital converter (ADC) and a field-programmable gate array (FPGA), where the ADC is responsible for converting analog signals into digital signals and the FPGA is responsible for spectrum calculation. The FPGA is a semi-customized circuit that can be customized through firmware to establish the final function of the FPGA. The function of the FPGA can be changed by loading different firmware, but in order to change the frequency resolution of the spectrum analyzer, it is necessary to design firmware for each required frequency resolution, which has limited flexibility.
[0006] The core of a spectrum analyzer is a customized device based on ADC and DSP microprocessor, where ADC is responsible for converting analog signals into digital signals and DSP is responsible for spectrum calculation. DSP is software-driven and can flexibly configure frequency resolution; however, due to the limited computing power of DSP, it is impossible to achieve true broadband real-time spectrum analysis. At present, frequency slicing or time slicing technology is mostly used to approximate the analysis of broadband spectrum.
[0007] Another type of spectrum analyzer is based on ADC and graphics card (GPU), where ADC is responsible for converting analog signals into digital signals and GPU is responsible for spectrum calculation. GPU is software-driven and can flexibly configure frequency resolution; due to the excellent parallel computing capability of GPU, a single GPU can realize real-time analysis of broadband spectrum (1GHz~10GHz), and multiple GPUs can achieve larger bandwidth in collaboration. However, the current broadband spectrum analyzer based on ADC and GPU has a direct connection between ADC and GPU. When designing the spectrum analyzer, it is necessary to adapt the ADC sampling frequency bandwidth and GPU computing capability, otherwise it will cause waste of ADC or GPU resources. Summary of the invention
[0008] In view of this, the present invention proposes an ultra-wideband high-resolution flexibly configurable spectrum analyzer. The present invention can realize real-time analysis of ultra-wideband (≥100GHz) spectrum with high frequency resolution and achieve optimal use of hardware resources.
[0009] In order to achieve the above object, the technical solution adopted by the present invention is as follows:
[0010] An ultra-wideband high-resolution flexibly configurable spectrum analyzer comprises a signal acquisition subsystem, a first switch, a spectrum calculation subsystem, a second switch and a spectrum synthesis unit;
[0011] The signal acquisition subsystem includes one or more signal acquisition units, the number of which is consistent with the number of RF sub-band signal divisions; each signal acquisition unit includes an analog-to-digital converter, a channelization and packaging module, and one or more network ports;
[0012] The spectrum calculation subsystem includes one or more spectrum calculation units, each of which includes one or more data receiving network cards, one or more GPUs and one or more data sending network cards;
[0013] The spectrum integration unit includes one or more data receiving network cards, one or more CPUs / GPUs, and a display;
[0014] In the signal acquisition unit, the analog-to-digital converter is used to convert the radio frequency sub-band signal into a sub-band digital signal; the channelization and packetization module is used to divide the sub-band digital signal into a plurality of channels, and then packetize the sub-band digital signals of different channels, and at the same time allocate the corresponding data receiving network card of the spectrum calculation unit to the sub-band digital signals after packetization of different channels, and then transmit them to the first switch through the network port;
[0015] The first switch is used to transmit the sub-band digital signals after different channel packetization output by the signal acquisition unit to the spectrum calculation subsystem;
[0016] In the spectrum calculation unit, the data receiving network card is used to receive the sub-band digital signal after corresponding channel packetization transmitted by the corresponding signal acquisition unit, and transmit it to the corresponding GPU. Each GPU receives the sub-band digital signal after partial or full channel packetization transmitted by one or more signal acquisition units, unpacks it and calculates its corresponding spectrum. The frequency bandwidth of the sub-band digital signal processed by each GPU matches the processing capability of the GPU. The GPU transmits the calculated spectrum data to the second switch via the data sending network card.
[0017] The second switch is used to transmit the spectrum data output by the spectrum calculation unit to the spectrum integration unit;
[0018] In the spectrum integration unit, the data receiving network card is used to receive the spectrum data output by the spectrum calculation unit and transmit it to the CPU / GPU. The CPU / GPU is used to integrate all the spectrum data by frequency to obtain full-bandwidth spectrum data, and transmit it to the display for display.
[0019] Furthermore, the port data bandwidth of the first switch and the second switch is greater than or equal to 100GbE; the port data bandwidth of the input ports of the first switch and the second switch is consistent with the port data bandwidth of the output ports of the corresponding connection units, and the port data bandwidth of the output ports of the first switch and the second switch is consistent with the port data bandwidth of the input ports of the corresponding connection units.
[0020] Furthermore, when the network card supports the transceiver duplex mode, the data receiving network card and the data sending network card of the spectrum calculation unit are implemented based on the same network card.
[0021] Further, when the switching capacity of a single switch is sufficient, the first switch and the second switch are implemented based on the same switch.
[0022] Due to the adoption of the above technical solution, the beneficial effects of the present invention compared with the prior art are:
[0023] 1. The present invention constructs a spectrum analyzer based on a star-type terabit local area network. By increasing / decreasing the number of online spectrum analyzer signal acquisition units and spectrum calculation units in the local area network, the effective frequency bandwidth can be flexibly configured; and the spectrum of ≥100GHz can be analyzed in real time.
[0024] 2. In the present invention, the signal acquisition unit and the spectrum calculation unit are not in one-to-one correspondence, which can achieve optimal utilization of hardware resources.
[0025] 3. The present invention is based on GPU computing and integrated spectrum, and can flexibly configure spectrum parameters (such as time and frequency resolution) to achieve real-time analysis of high frequency resolution and / or high time resolution spectrum.
[0026] 4. The present invention is based on GPU computing and integrated spectrum, and its functions can be flexibly upgraded and iterated. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 Schematic diagram of the structure of an ultra-wideband, high-resolution, flexibly configurable spectrum analyzer in an embodiment of the present invention.
[0028] Figure 2 for Figure 1 Schematic diagram of the structure of the signal acquisition unit.
[0029] Figure 3 for Figure 1 Schematic diagram of the structure of the spectrum calculation unit.
[0030] Figure 4 for Figure 1 Schematic diagram of the structure of the spectrum synthesis unit. DETAILED DESCRIPTION
[0031] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.
[0032] An ultra-wideband, high-resolution, flexibly configurable spectrum analyzer, such as Figure 1 As shown, it consists of a distributed signal acquisition subsystem, a switch, a distributed spectrum calculation subsystem and a spectrum integration unit.
[0033] Specifically, 1) build a spectrum analyzer based on a star-shaped terabit local area network; 2) distributed signal acquisition and spectrum calculation; 3) integrate the output of distributed spectrum calculation units to form a high-frequency resolution ultra-wideband spectrum; and 4) graphics display card (GPU) spectrum calculation.
[0034] Among them, the core components of the star-type terabit LAN are general high-speed switches with port data bandwidth greater than or equal to 100GbE, multiple network ports / network cards with data bandwidth greater than or equal to 100GbE, and network data cables.
[0035] Specifically, Figure 1 Switch-1 and switch-2 can be combined, that is, a single LAN can realize all data transmission, saving the hardware cost of the spectrum analyzer;
[0036] like Figure 2 As shown in FIG. 1 , the distributed signal acquisition subsystem is composed of M signal acquisition units. The signal acquisition unit is a customized device, mainly composed of an analog-to-digital converter (ADC), a field programmable gate logic (FPGA), and a single or multiple network ports with a data bandwidth greater than or equal to 100GbE.
[0037] Specifically, the number of signal acquisition units is consistent with the number of RF sub-band signal divisions; the signal acquisition unit ADC samples the input RF sub-band, converts the analog signal into a digital signal and transmits it to the FPGA; the FPGA channels and packages the data, and sends it to the spectrum calculation subsystem through the network port via the star-shaped terabit local area network;
[0038] like Figure 3 As shown in the figure, the distributed spectrum computing subsystem is composed of N spectrum computing units. The spectrum computing unit is a general computing server, mainly composed of one or more network cards with a data bandwidth greater than or equal to 100GbE and one or more GPUs; a single spectrum computing unit can process part or all of the output of one or more signal acquisition units. Since there is no one-to-one correspondence between the signal acquisition unit and the spectrum computing unit, there will be no waste of subsystem resources.
[0039] Specifically, the performance of the signal acquisition unit ADC and the spectrum calculation unit GPU are analyzed: if the processing capability of the signal acquisition unit ADC is stronger than that of the spectrum calculation unit GPU, a single spectrum calculation unit GPU processes part of the output of a single signal acquisition unit; if the processing capability of the signal acquisition unit ADC is the same as that of the spectrum calculation unit GPU, a single spectrum calculation unit GPU processes all the output of a single signal acquisition unit; if the processing capability of the spectrum calculation unit GPU is stronger than that of the signal acquisition unit ADC, a single spectrum calculation unit GPU processes part or all of the output of multiple signal acquisition units; according to the above different situations, the FPGA in the signal acquisition unit is set, and the FPGA distributes the data after packetization of different channels to the corresponding GPU, that is, to the data receiving network card of the corresponding spectrum calculation unit;
[0040] The data receiving network card of the spectrum calculation unit is responsible for receiving digital signals, and the GPU is responsible for calculating the spectrum of the input signal according to given parameters (such as the number of channels and integration time), and then sending the spectrum data to the spectrum synthesis unit through the network card via the star-shaped terabit local area network. Data reception and transmission can be based on the same network card (transmit-receive duplex), saving the hardware cost of the spectrum analyzer;
[0041] like Figure 4 As shown, the spectrum synthesis unit is a general computing server, which is mainly composed of a single or multiple network cards with a data bandwidth greater than or equal to 100GbE, a single or multiple CPU / GPUs and a display. To simplify the system structure, the network card and the GPU can correspond one to one.
[0042] The spectrum synthesis unit receives the output data of the spectrum calculation subsystem through a single or multiple network cards, and synthesizes all the data by frequency to obtain full-bandwidth spectrum data, which is then displayed on the monitor.
[0043] The spectrum analyzer hardware parameters (such as the data bandwidth between the local area network and each subsystem / unit, the data processing capability of each unit, etc.) are determined by the spectrum analyzer parameters (such as the total frequency bandwidth, time / frequency resolution, etc.).
[0044] In summary, the present invention can realize real-time analysis of ultra-wideband (≥100 GHz) spectrum with high frequency resolution and optimize the use of hardware resources.
[0045] Those skilled in the art will appreciate that the embodiments described are intended to help readers understand the principles of the present invention, and should be understood that the scope of protection of the present invention is not limited to the embodiments described. For those skilled in the art, the present invention may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of the claims of the present invention.
Claims
1. An ultra-wideband, high-resolution, flexibly configurable spectrum analyzer. It is characterized in that It includes a signal acquisition subsystem, a first switch, a spectrum calculation subsystem, a second switch and a spectrum integration unit; The signal acquisition subsystem includes one or more signal acquisition units, the number of which is consistent with the number of RF sub-band signal divisions; each signal acquisition unit includes an analog-to-digital converter, a channelization and packaging module, and one or more network ports; The spectrum calculation subsystem includes one or more spectrum calculation units, each of which includes one or more data receiving network cards, one or more GPUs and one or more data sending network cards; The spectrum integration unit includes one or more data receiving network cards, one or more CPUs / GPUs, and a display; In the signal acquisition unit, the analog-to-digital converter is used to convert the radio frequency sub-band signal into a sub-band digital signal; the channelization and packetization module is used to divide the sub-band digital signal into a plurality of channels, and then packetize the sub-band digital signals of different channels, and at the same time allocate the corresponding data receiving network card of the spectrum calculation unit to the sub-band digital signals after packetization of different channels, and then transmit them to the first switch through the network port; The first switch is used to transmit the sub-band digital signals after different channel packetization output by the signal acquisition unit to the spectrum calculation subsystem; In the spectrum calculation unit, the data receiving network card is used to receive the sub-band digital signal after corresponding channel packetization transmitted by the corresponding signal acquisition unit, and transmit it to the corresponding GPU. Each GPU receives the sub-band digital signal after partial or full channel packetization transmitted by one or more signal acquisition units, unpacks it and calculates its corresponding spectrum. The frequency bandwidth of the sub-band digital signal processed by each GPU matches the processing capability of the GPU. The GPU transmits the calculated spectrum data to the second switch via the data sending network card. Specifically, the performance of the analog-to-digital converter of the signal acquisition unit and the GPU of the spectrum calculation unit are analyzed: if the processing capability of the analog-to-digital converter of the signal acquisition unit is stronger than that of the GPU of the spectrum calculation unit, then a single GPU of the spectrum calculation unit processes part of the output of a single signal acquisition unit; if the processing capability of the analog-to-digital converter of the signal acquisition unit is the same as that of the GPU of the spectrum calculation unit, then a single GPU of the spectrum calculation unit processes all the output of a single signal acquisition unit; if the processing capability of the GPU of the spectrum calculation unit is stronger than that of the analog-to-digital converter of the signal acquisition unit, then a single GPU of the spectrum calculation unit processes part or all of the output of multiple signal acquisition units; according to the above different situations, the channelization and packetization module in the signal acquisition unit is set, and the channelization and packetization module distributes the data after packetization of different channels to the corresponding GPU, that is, to the data receiving network card of the corresponding spectrum calculation unit; The second switch is used to transmit the spectrum data output by the spectrum calculation unit to the spectrum integration unit; In the spectrum integration unit, the data receiving network card is used to receive the spectrum data output by the spectrum calculation unit and transmit it to the CPU / GPU. The CPU / GPU is used to integrate all the spectrum data by frequency to obtain full-bandwidth spectrum data, and transmit it to the display for display.
2. The ultra-wideband, high-resolution, flexibly configurable spectrum analyzer according to claim 1, It is characterized in that The port data bandwidth of the first switch and the second switch is greater than or equal to 100GbE; the port data bandwidth of the input ports of the first switch and the second switch is consistent with the port data bandwidth of the output ports of the corresponding connection units, and the port data bandwidth of the output ports of the first switch and the second switch is consistent with the port data bandwidth of the input ports of the corresponding connection units.
3. The ultra-wideband, high-resolution, flexibly configurable spectrum analyzer according to claim 1, It is characterized in that When the network card supports the transceiver duplex mode, the data receiving network card and the data sending network card of the spectrum calculation unit are implemented based on the same network card.
4. The ultra-wideband, high-resolution, flexibly configurable spectrum analyzer according to claim 1, It is characterized in that When the switching capacity of a single switch is sufficient, the first switch and the second switch are implemented based on the same switch.
