A lightweight transmitting end spectrum synthesis method based on tree-shaped sub-channel synthesis

Abstract

This invention discloses a lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis. It employs a multi-level tree-structured multi-rate subband synthesis architecture. The first-stage synthesis is completed through ZOH sample-and-hold double interpolation and ±π / 4 digital frequency complex rotation mixing. Subsequent stages recursively perform double interpolation, HBF half-band filtering, and ±π / 4 digital frequency mixing, resulting in a progressively increasing number of frequency points and a progressively decreasing number of channels. Simultaneously, at each stage, the digital relative frequency difference between the image frequency group and the signal frequency group is maintained at greater than π / 2. Compared to existing technologies, this invention achieves effective image suppression without requiring high-order interpolation filters, only needing short-tap half-band filters, significantly reducing hardware area and power consumption. Each subband occupies only 1 / N of the digital frequency circle, is far from the spectral boundary, has a loose transition band, and a gentle spectral gradient, achieving full-band coverage without reserving blank protection subcarriers. This achieves a lightweight transmitter architecture and efficient utilization of spectrum resources.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, specifically to a lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis. Background Technology

[0002] In modern wireless communication, optical communication, and various digital signal processing systems based on FFT / IFFT, the "IFFT spectrum synthesis + DUC digital upconversion" architecture is commonly used to achieve multi-subband signal synthesis and spectrum shifting. This architecture is a common standard baseband solution for 3G / 4G / 5G base stations, civilian mobile phone terminals, IoT radio frequency equipment, and optical communication transceiver systems. However, it has inherent defects that cannot be eradicated at the architectural level, especially in sparse spectrum and non-uniform heterogeneous spectrum scenarios.

[0003] The classic IFFT spectrum synthesis mechanism essentially requires nearly filling the entire digital frequency circle before performing sample rate boosting and interpolation filtering. This design is a high-risk, high-cost approach that has been universally adopted across the industry for a long time: when the spectrum resources almost fill the frequency domain boundary, it leads to extremely narrow signal transition bands and extremely steep gradients at the spectral edges. To avoid crosstalk between different subbands and suppress severe out-of-band spectral tails and adjacent channel leakage, engineering applications are forced to use extremely high-order, extremely steep interpolation filters, which brings huge multiplication overhead, chip area overhead, and dynamic power consumption overhead; even with high-order steep filtering, out-of-band leakage cannot be eliminated due to the non-ideal brick wall characteristics.

[0004] Therefore, the industry must reserve 10% of the effective spectrum on both sides, totaling 20%, as guard bands. This portion of the spectrum is completely unusable for service data, resulting in a systemic and permanent waste of spectrum resources. More critically, FFT / IFFT is a fixed uniform grid orthogonal transformation architecture, which only supports uniform and equal-width subband division and cannot natively adapt to heterogeneous spectrum scenarios with non-uniform bandwidth, sparse discreteness, and asymmetry. In scenarios such as 5G heterogeneous subbands, 6G ultra-wideband sparse spectrum, 7G intelligent dynamic spectrum, and high-speed optical communication sparse spectrum scheduling, traditional IFFT solutions must globally fill the grid according to the maximum bandwidth requirement. A large number of invalid grids and air interface subbands still need to participate in complete complex number operations and caching, resulting in comprehensive invalid redundancy in computing power, chip area, and power consumption. This is a hard architectural defect that cannot be improved through algorithm optimization. This architectural defect is prevalent in all signal processing scenarios that rely on FFT / IFFT and is a common technical bottleneck across the industry, making it difficult to meet the core iterative requirements of high spectral efficiency and ultra-low power consumption in next-generation communications.

[0005] The existing IFFT+DUC discrete architecture suffers from problems such as high-order filtering dependence, spectrum leakage, and blank carrier redundancy. These are common technical bottlenecks in wireless communication, optical communication, and all FFT-type digital spectrum processing systems. They generally suffer from drawbacks such as large power consumption area, low spectrum utilization, and poor scenario adaptability. They cannot meet the requirements of 6G, 7G, and high-speed intelligent optical communication systems for lightweight and ultra-efficient spectrum processing. There is an urgent need for a general spectrum synthesis scheme with architectural innovation. Summary of the Invention

[0006] This invention aims to provide a lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis. Addressing the problems of traditional IFFT+DUC discrete architectures, such as circuit redundancy, large area and power consumption, fixed spectrum grid, inability to adapt to sparse / non-uniform heterogeneous spectrum, and difficulty in meeting the high spectral efficiency and low power consumption iteration requirements of 5G / 6G / 7G and optical communication, this method integrates the traditional discrete IFFT spectrum synthesis with the DUC up-conversion function. It can replace the traditional commercial standard architecture, abandoning the high-risk design of full-frequency circles in FFT / IFFT and the inherent defects of uniform grids. It achieves overwhelming performance and hardware advantages in sparse and non-uniform mainstream spectrum scenarios, enabling universal, small-area, ultra-low power consumption, and ultra-high spectral efficiency multi-subband signal synthesis.

[0007] To achieve the above objectives, the technical solution adopted by this invention is: a lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis, comprising the following steps:

[0008] S1: Obtain the multi-channel baseband modulation signals to be synthesized, where each of the baseband modulation signals corresponds to a QAM modulation symbol;

[0009] S2: Perform first-level tree synthesis processing on the multi-channel baseband modulation signals, use ZOH sample-and-hold double interpolation to interpolate each of the baseband modulation signals, and perform complex rotation mixing with ±π / 4 digital frequencies to generate multi-channel dual-frequency signals;

[0010] S3: Perform multi-level recursive tree synthesis processing on the multi-channel dual-frequency signals. Each level of recursive processing sequentially performs double interpolation, HBF half-band filtering and ±π / 4 digital frequency mixing, so that the number of frequency points contained in each signal increases by one level and the number of channels decreases by one level.

[0011] S4: In each stage of the multi-level recursive tree synthesis process, the digital relative frequency difference between the image frequency group and the signal frequency group is kept greater than π / 2, so that image suppression can be achieved by using only short-tap half-band filters at each stage;

[0012] S5: After the multi-level recursive tree synthesis process is superimposed and synthesized step by step, a single-channel high-speed time-domain transmission waveform is output, completing the multi-subband spectrum synthesis and rate enhancement.

[0013] Preferably, in S2, the operational expression for the complex rotation mixer is: Where A and B are the QAM modulation values ​​at two frequency points, and n is the sample number; after the complex rotation mixing, useful information is modulated at the positions of the two frequency points ±π / 4.

[0014] Preferably, in S3, the double interpolation is zero-interpolation, that is, zero-value samples are inserted between adjacent samples; the HBF half-band filter is used to filter and suppress the image components generated after interpolation; the ±π / 4 digital frequency mixer is used to shift the filtered signal to the target frequency position.

[0015] Preferably, in S4, the number of taps in the short-tap half-band filter is 7, 15, or 31.

[0016] Preferably, the tree-based synthesis architecture is a tree-based N-type or hybrid synthesis architecture.

[0017] Preferably, the method maintains an oversampling ratio throughout the synthesis process until the final synthesis output, thereby constructing a spectrum in the shape of a brick wall.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0019] This invention combines the traditional discrete IFFT spectrum synthesis with the DUC upconversion function into one, and adopts a multi-level tree-shaped multi-rate sub-band synthesis architecture. The first-level synthesis is completed by ZOH sample-and-hold double interpolation and ±π / 4 digital frequency complex rotation mixing. In subsequent levels, double interpolation, HBF half-band filtering and ±π / 4 digital frequency mixing are recursively executed, so that the number of frequency points increases by a factor of two and the number of channels decreases by a factor of two. At the same time, the digital relative frequency difference between the image frequency group and the signal frequency group is maintained at each level greater than π / 2. Compared to existing technologies, this invention achieves effective image suppression without the need for high-order interpolation filters, requiring only short-tap half-band filters, significantly reducing hardware area and power consumption. Each sub-band occupies only 1 / N of the digital frequency circle, is far from the spectrum boundary, has a loose transition band, and a gentle spectrum gradient, achieving full-band coverage without reserving blank protection subcarriers. The entire process adopts a sample-by-sample streaming pipeline, eliminating the need for bulk data buffering and large-scale matrix transformations, and natively supports on-demand configuration of non-uniform, sparse, and asymmetric sub-bands. Thus, while ensuring spectrum synthesis quality, it achieves lightweight transmitter architecture and efficient utilization of spectrum resources. Attached Figure Description

[0020] Figure 1This is a block diagram of the pipeline architecture for the synthesis of 4096 sub-channels and the DUT in this invention;

[0021] Figure 2 This is the first-level block diagram of the tree-shaped binary channel synthesis of the present invention;

[0022] Figure 3 This is a second-level block diagram of the tree-shaped binary channel synthesis of the present invention;

[0023] Figure 4 This is a block diagram of the third to twelfth levels of the tree-shaped binary channel synthesis of the present invention;

[0024] Figure 5 This is a flowchart of the lightweight transmitter spectrum synthesis method based on tree-shaped subchannel synthesis according to the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0026] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a specific posture. If the specific posture changes, the directional indicators will also change accordingly.

[0027] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the use of "and / or" or "and / or" throughout the text includes three parallel solutions. For example, "A and / or B" includes solution A, solution B, or a solution where both A and B are satisfied simultaneously. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0028] Please see Figures 1-5This invention provides a technical solution: a lightweight transmitter spectrum synthesis method based on tree-shaped sub-channel synthesis. The method maintains an oversampling ratio throughout the synthesis process until the final synthesis output, thereby constructing a brick-wall-shaped spectrum. The method includes the following steps:

[0029] S1: Obtain the multiple baseband modulation signals to be synthesized, each baseband modulation signal corresponding to a QAM modulation symbol; receive the multiple baseband modulation signals in a sample-by-sample manner, so that the subsequent tree synthesis processing can run continuously in a streaming pipeline mechanism, which helps to avoid the storage overhead and processing delay caused by the whole data caching, and at the same time provides a regular input data structure for each level of recursive synthesis.

[0030] S2: Perform first-level tree synthesis processing on the multi-channel baseband modulation signals, use ZOH sample-and-hold double interpolation to interpolate each of the baseband modulation signals, and perform complex rotation mixing with ±π / 4 digital frequencies to generate multi-channel dual-frequency signals;

[0031] The operational expression for complex rotation mixing is: Where A and B are the QAM modulation values ​​at two frequency points, and n is the sample number; after the complex rotation mixing, useful information is modulated at the positions of the two frequency points ±π / 4.

[0032] By combining sample-and-hold interpolation with fixed-frequency complex rotation mixing, two modulation symbols can be moved to symmetrical frequency positions at the first stage with low computational complexity. This provides a regular dual-frequency input structure for subsequent multi-stage recursive synthesis, which helps reduce the reliance on complex rotation factors or high-order filters in traditional architectures.

[0033] S3: Perform multi-level recursive tree synthesis processing on the multi-channel dual-frequency signals. Each level of recursive processing sequentially performs double interpolation, HBF half-band filtering and ±π / 4 digital frequency mixing, so that the number of frequency points contained in each signal increases by one level and the number of channels decreases by one level.

[0034] Among them, double interpolation is zero-interpolation, that is, inserting zero-value samples between adjacent samples; HBF half-band filtering is used to filter and suppress the image components generated after interpolation; ±π / 4 digital frequency mixing is used to shift the filtered signal to the target frequency position.

[0035] Taking the synthesis of 4096 frequency points as an example, the multi-level recursive tree synthesis process includes twelve levels: the first level outputs 2048 dual-frequency point signals, the second level outputs 1024 four-frequency point signals, the third level outputs 512 eight-frequency point signals, the fourth level outputs 256 sixteen-frequency point signals, the fifth level outputs 128 thirteen-frequency point signals, the sixth level outputs 64 sixteen-frequency point signals, the seventh level outputs 32 thirteen-frequency point signals, the eighth level outputs 16 thirteen-frequency point signals, the ninth level outputs 8 fiveteen-frequency point signals, the tenth level outputs 4 tenteen-frequency point signals, the eleventh level outputs 2 twenty-four-frequency point signals, and the twelfth level outputs 1 fourteen-frequency point signals.

[0036] S4: In each stage of the multi-level recursive tree synthesis process, the digital relative frequency difference between the image frequency group and the signal frequency group is kept greater than π / 2, so that image suppression can be achieved by using only short-tap half-band filters at each stage; the number of taps of the short-tap half-band filters is 7, 15 or 31.

[0037] By maintaining sufficient digital frequency difference between the image and signal frequency groups at each stage, the half-band filter can obtain sufficient image rejection capability with a shorter number of taps. This helps to reduce the hardware resource consumption and computational overhead of the filter, while avoiding excessive requirements on the filter order due to the narrow transition band, thus providing a relatively relaxed design margin for the lightweight implementation of the overall architecture.

[0038] S5: After the multi-level recursive tree-structured synthesis process, the waveform is superimposed and synthesized level by level, resulting in a single-channel high-speed time-domain transmission waveform. Several levels of interpolation filtering are then added to increase the sample rate, thus meeting the DAC's spectral suppression requirements. This completes multi-subband spectral synthesis and rate enhancement. The tree-structured synthesis architecture can be a tree-division N-type or hybrid synthesis architecture.

[0039] By superimposing multiple levels of recursion, a single high-speed time-domain waveform is finally output. While completing spectrum synthesis and improving the sampling rate, the flexible configuration capability of the tree-based N-type or hybrid architecture can better adapt to the differentiated requirements of different sub-band numbers, bandwidth distributions and rate ratios, which helps to improve the system's compatibility with heterogeneous spectrum scenarios.

[0040] This invention overturns the inherent defects of traditional IFFT's high-risk full-frequency circle and fixed uniform grid at the underlying architecture level. It has overwhelming hardware and performance advantages in sparse spectrum and non-uniform heterogeneous spectrum scenarios, and solves the industry pain points of traditional architecture relying on high-order steep filtering, forced reserved blank protection carrier, poor spectrum adaptability, and serious redundant ineffective computing power.

[0041] This invention adopts a multi-segment tree-shaped rotating synthesis architecture, in which each sub-band signal occupies only 1 / N of the digital frequency circle. The sub-band spectrum is naturally located in the central region of the frequency circle and far from the spectrum boundary. The transition band is loose and the spectrum gradient is gentle. It can achieve excellent out-of-band attenuation and spectrum isolation without steep, high-order interpolation filters, thus eliminating the huge hardware overhead and power consumption burden caused by high-order filtering.

[0042] Meanwhile, the narrowband group synthesis mode of the present invention has extremely strong spectral convergence, fast out-of-band attenuation rate, extremely low adjacent channel spectral leakage, and no grid distortion problem. It does not require the 20% edge blank protection subcarrier required by the traditional architecture, and realizes full coverage utilization of the entire frequency band subband.

[0043] Most importantly, this architecture natively supports custom subband synthesis for non-uniform, sparse, and asymmetric spectrum. Branch parameters can be configured as needed based on actual service bandwidth, spectrum spacing, and sparse distribution characteristics, performing computation only on valid spectrum branches. This completely avoids the redundant computing power and hardware overhead caused by empty grids and invalid subbands in traditional FFT / IFFT. In the mainstream scenario of sparse spectrum and non-uniform heterogeneous spectrum, this invention can achieve a dramatic reduction in chip area and computational power consumption, while significantly improving spectrum utilization flexibility and resource efficiency.

[0044] Each baseband modulation signal at the transmitting end is connected to the terminal sub-band branch of the tree synthesis architecture. Through multi-level tree superposition, rate adaptation, and spectrum fusion, a single high-speed time-domain transmission waveform is finally output, directly completing multi-sub-band spectrum synthesis and rate enhancement, eliminating the redundant links of traditional DUC multi-level interpolation, filtering, and spectrum shifting.

[0045] This invention's multi-segment narrowband rotational synthesis architecture places subbands at the center of the frequency circle, provides ample transition band margin, and ensures smooth spectral changes. It eliminates the need for any high-order steep filtering units, thus cutting out the core hardware overhead of DUC high-order filtering. Simultaneously, relying on an on-demand adaptable tree-like branch structure, it activates only effective subband paths in sparse and non-uniform scenarios, eliminating redundant computation in empty grids. This results in a significant reduction in hardware gate count, chip area, instantaneous power consumption, and static power consumption. It is widely adaptable to the lightweight and low-power design requirements of 5G base stations, mobile terminals, IoT RF devices, and high-speed optical communications, aligning with the iterative trends of next-generation 6G and 7G communication hardware.

[0046] This invention employs a sample-by-sample streaming pipeline processing throughout, eliminating the waiting delay of FFT / IFFT block processing and the need to cache complete symbol data frames, resulting in faster signal synthesis response and stronger real-time performance. At the same time, the integrated architecture eliminates the redundant links of traditional DUC multi-level interpolation, filtering, and spectrum shifting, significantly reducing system processing latency and making it suitable for low-latency service scenarios in next-generation high-speed communication.

[0047] In one exemplary embodiment, a sample-by-sample pipeline can be used to process samples continuously, with delay registers configured only within each stage of the HBF half-band filter for filtering operations, without performing data caching operations on complete symbol blocks, and without calling large-capacity RAM to store symbol blocks to be processed.

[0048] The baseband modulation signal is continuously fed into the terminal branch of the tree synthesis architecture in units of individual samples. The system completes the first-level tree synthesis processing and subsequent multi-level recursive tree synthesis processing sample by sample. In the entire processing chain, only the delay register necessary for filtering operations is configured inside each level of the HBF half-band filter to complete the half-band filtering operation. The system does not collect or store complete symbol block data, so there is no need to configure a large-capacity RAM to cache complete symbol blocks. Unlike the traditional IFFT architecture, which needs to wait for the collection of complete symbol blocks before starting the block transformation operation, each sample flows in and immediately performs ZOH sample-and-hold double interpolation, HBF half-band filtering, and complex rotation mixing operations step by step along the tree synthesis architecture. After the samples flow step by step, the high-speed time-domain transmission waveform is directly output, eliminating the frame waiting delay caused by block processing. It is suitable for 5G / 6G communication and IoT real-time signal transmission scenarios. When implementing this embodiment using ASIC or FPGA, there is no need to configure the complex rotation factor ROM required by IFFT, reducing storage resource overhead.

[0049] In another exemplary embodiment, for service scenarios with sparse discrete spectrum and non-uniform bandwidth subbands, when there is an idle subband without service data, the end branch of the tree synthesis architecture corresponding to the idle subband is blocked, and the ZOH sample-and-hold double interpolation, HBF half-band filtering and complex rotation mixing operation of ±π / 4 digital frequency are stopped for the idle subband branch. Only the effective subband branch carrying service data is processed by tree synthesis at each level of S2-S5. The idle subband branch no longer participates in the signal superposition of subsequent levels.

[0050] For example, the tree-structured synthesis architecture is configured with 16 sub-band branches at the end. Services only use 3 of these as active sub-band branches, leaving the remaining 13 as idle sub-bands. Therefore, only the interpolation, filtering, and mixing operations of the 3 active sub-band branches are enabled, while the computation paths of the remaining 13 idle sub-band branches are disabled. In this mode, unlike traditional IFFT+DUC solutions, it is not necessary to perform operations on all grids, regardless of whether the grid carries services. This embodiment only consumes the computational resources corresponding to the active sub-band branches, reducing computational overhead in sparse scenarios. Furthermore, this tree-structured synthesis architecture allows each active sub-band branch to be configured with different bandwidths, enabling asymmetric and non-uniform sub-band spectrum synthesis, unconstrained by uniform frequency grids.

[0051] In another exemplary embodiment, relying on the constraint that the digital relative frequency difference between the mirror frequency group and the signal frequency group at each level is greater than π / 2, the sub-band signals of the synthesized output spectrum are distributed in the central region of the digital frequency circle. It is not necessary to set reserved blank protection subcarriers on both sides of the effective sub-bands, and all frequency points of the synthesized output spectrum have the ability to carry service data.

[0052] This embodiment demonstrates the technical advantages of omitting blank guard subcarriers. In each stage of the multi-level recursive tree synthesis process, the digital relative frequency difference between the image frequency group and the signal frequency group is continuously controlled to remain greater than π / 2. This ensures that the synthesized subband signals are distributed in the central region of the digital frequency circle, far from its spectral boundaries. Sufficient image suppression and out-of-band attenuation can be achieved using a short-tap HBF half-band filter; therefore, this embodiment eliminates the need to reserve blank guard subcarriers on both sides of the effective subband to suppress adjacent channel leakage. In contrast, the traditional IFFT+DUC scheme requires 10% blank guard subcarriers (20%) on both sides of the effective spectrum to suppress adjacent channel leakage, and these frequency points cannot transmit service data. The processing method in this embodiment eliminates this spectrum resource overhead, offering particularly significant benefits in high-speed optical communication and 6G dynamic spectrum scheduling scenarios.

[0053] As an exemplary numerical example, the total number of stages in the multi-level recursive tree synthesis process is configured according to the total number of output frequency points to adapt to the output requirements of different bandwidth transmit waveforms. When a synthesized signal with 4096 frequency points needs to be output, a twelve-stage recursive tree synthesis process is configured. The first-stage tree synthesis process outputs 2048 dual-frequency signals. Each subsequent stage of the recursive tree synthesis process sequentially doubles the number of frequency points and halves the number of channels. After a complete twelve-stage interpolation, HBF half-band filtering, complex rotation mixing, and signal superposition, a high-speed time-domain transmit waveform with 4096 frequency points is finally output. It can be understood that the number of recursive stages can be adjusted according to the required output frequency point scale of the system to form a tree-based N-type or hybrid synthesis architecture, and is not limited to a twelve-stage implementation.

[0054] The above description is merely an exemplary embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention specification and drawings under the technical concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis, characterized in that, Includes the following steps: S1: Obtain the multi-channel baseband modulation signals to be synthesized, where each of the baseband modulation signals corresponds to a QAM modulation symbol; S2: Perform first-level tree synthesis processing on the multi-channel baseband modulation signals, use ZOH sample-and-hold double interpolation to interpolate each of the baseband modulation signals, and perform complex rotation mixing with ±π / 4 digital frequencies to generate multi-channel dual-frequency signals; S3: Perform multi-level recursive tree synthesis processing on the multi-channel dual-frequency signals. Each level of recursive processing sequentially performs double interpolation, HBF half-band filtering and ±π / 4 digital frequency mixing, so that the number of frequency points contained in each signal increases by one level and the number of channels decreases by one level. S4: In each stage of the multi-level recursive tree synthesis process, the digital relative frequency difference between the image frequency group and the signal frequency group is kept greater than π / 2, so that image suppression can be achieved by using only short-tap half-band filters at each stage; S5: After the multi-level recursive tree synthesis process is superimposed and synthesized step by step, a single-channel high-speed time-domain transmission waveform is output, completing the multi-subband spectrum synthesis and rate enhancement.

2. The lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis according to claim 1, characterized in that, In S2, the operational expression for the complex rotation mixer is: Where A and B are the QAM modulation values ​​at two frequency points, and n is the sample number; after the complex rotation mixing, useful information is modulated at the positions of the two frequency points ±π / 4.

3. The lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis according to claim 1, characterized in that, In S3, the double interpolation is zero-interpolation, that is, inserting zero-value samples between adjacent samples; the HBF half-band filter is used to filter and suppress the image components generated after interpolation; the ±π / 4 digital frequency mixer is used to shift the filtered signal to the target frequency position.

4. The lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis according to claim 1, characterized in that, In S4, the number of taps in the short-tap half-band filter is 7, 15, or 31.

5. A lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis according to claim 1, characterized in that, The tree-based synthesis architecture is either a tree-based N-type or a hybrid synthesis architecture.

6. A lightweight transmitter spectrum synthesis method based on tree-structured subchannel synthesis according to claim 1, characterized in that, The method maintains an oversampling ratio throughout the synthesis process until the final synthesis output, thereby constructing a spectrum in the shape of a brick wall.

Images