Dual-band wireless network filtering and decimation, passive intermodulation (PIM) detection, and PIM nonlinearity calculation in PIM cancellation
By employing filtering, decimation, and a novel PIM detection method with pre-processing NCOs and DUC/DDC, the complexity and resource issues in dual-band PIM cancellation are addressed, achieving efficient PIM interference cancellation in wireless networks.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- MAVENIR SYST INC
- Filing Date
- 2025-01-27
- Publication Date
- 2026-07-30
AI Technical Summary
Current PIM detection and cancellation methods in dual-band wireless networks face challenges due to resource limitations in FPGA/ASIC solutions, complex signal modeling, and the need to handle combined nonlinear composite signals, especially with varying delays and carrier frequencies, which conventional systems struggle to address effectively.
Implement filtering and decimation to reduce complexity, use a novel PIM detection method with a two-dimensional grid search for delay detection, and introduce pre-processing NCOs and DUC/DDC to efficiently handle third and fifth-order intermodulation terms, reducing resource utilization.
This approach significantly reduces processing complexity and resource utilization in FPGA/ASIC solutions, enabling effective PIM cancellation in dual-band wireless networks by aligning delays and managing nonlinear composite signals.
Smart Images

Figure CN2025075473_30072026_PF_FP_ABST
Abstract
Description
Dual-band Wireless Network Filtering and Decimation, Passive Intermodulation (PIM) Detection, and PIM Nonlinearity Calculation in PIM CancellationBACKGROUND[XXXX]1. Field of the Disclosure
[0001] The present disclosure focuses on cancelling passive intermodulation (PIM) interference caused by the PIM product in a dual-band setup. More particularly, the disclosure relates to filtering & decimation, passive intermodulation detection (PIMD) , and PIM nonlinearity calculation in PIM cancellation in a dual-band wireless network.2. Description of Related Art
[0002] Passive intermodulation (PIM) is the non-linear products caused by non-linear characteristics of passive components, such as filters, connectors, and antenna elements. Many passive components show a non-linear transmission characteristic due to various reasons, for example, the non-linear feature of ferrite, which is widely used in RF components, and the non-linear product due to the junction of two different metals, which happens at the interaction of machinal components. RF circuits, particularly for the base stations, from within the RF feed sub-system after PA, such as duplexer through to antenna elements can create serious PIM product due to the high-power level in these stages of circuits. When the PIM products fall into one or more uplink (UL) bands, they become interference to the signal reception and must be handled properly.
[0003] PIM from within the RF feed system (duplexer through to antenna elements) can impact the Rx band of single, dual and triple-band RRHs, degrading Rx performance. PIM could be further catalogued into internal source PIM and external source PIM. Internal source PIM: PIM products (nonlinearity) generated in the passive Radio Frequency (RF) component in the cell including, for example, duplexer, cavity filter, antenna, and connector. External source PIM: PIM generated not in the cell but rather in objects in the surrounding environment, such as reflected PIM products, and PIM from other sources.
[0004] As technology has evolved into advanced B4G and 5G era, spectrum management also becomes more and more complicated. PIM cancellation has been applied to single transceiver (TRX) base stations. However, it is required that the modern base station be adapted to support two or more frequency bands. Dual-band design enhances the base station in both capability and flexibility. Unfortunately, PIM products in the dual-band base station are convoluted in terms of the frequency spectrum, and more likely become interference to the interested UL bands.
[0005] One problem with current systems relates to using receive (Rx) circuits as the PIM modeling feedback path. The received signal spectrum is comprised of many components. To precisely model the received signal and cancel it out, exceeds available Field Programmable Gate Array (FPGA) resources. The channel response from the Tx baseband all the way to Rx baseband varies drastically in magnitude, delay, and phase along the frequency axis. It goes through different channel responses for downlink (DL) bands, UL bands, and transient bands. These differences are due to the filtering characteristics of the duplexer. It should be noted that, even at the receiver, the power spectrum density (PSD) of the DL band might be still higher than that of the UL band, depending on the isolation provided by the duplexer. Consequently, the received signal spectrum is comprised of many components. To precisely model the received signal and cancel out PIM, exceeds the available FPGA resources. As such, the complexity of the signal as well as the intensive resource utilization are major hurdles for current systems.
[0006] Another problem current systems face is that the PIM interference signal or the nonlinear terms are the high-order and the combination of the two baseband signals of the dual-band transmission. This means, 1) the delays from each of the transmission base bands to the received signal are independent and different, depending on the independent transmission circuits and group delay variation for different DL bands, and 2) more importantly, in many cases, the PIM components available for detection are the combined nonlinear composite signals.
[0007] Conventional detection methods, such as those discussed in EP2880768B1 and US9768812B1, cannot process these combined composite signals as they are focused on single Tx base band as opposed to a dual band PIM cancellation solution.
[0008] Likewise, US9026064B2 has a passive intermodulation detection system to identify passive intermodulation at a base station site. The passive intermodulation detection system can generate a test signal in a first band that is transmitted by an antenna. Another antenna can receive a signal in another band, and the passive intermodulation detection system can analyze the received signal to determine whether an intermodulation product due to a nonlinearity is present. However, again this system is not designed to handle dual band nonlinearity.
[0009] Still another problem is that in the modeling of the PIM, there are multiple nonlinear terms (typically six or seven terms) involved even if we only consider the third (IM3) and fifth (IM5) order of the PIM interference. The challenge is that the carrier frequencies of these nonlinear terms are different to the carrier frequency of the Rx signal, of which the carrier frequency is the center frequency of the uplink band. One approach that may seem obvious would be to first calculate the IM3 and IM5 directly from the baseband transmission signals and then perform Digital Up-Conversion or Digital Down-Conversion (DUC / DDC) on the IM3 / IM5 terms. However, the problem with this approach is that we need to equip the DUC / DDC and the numerically controlled oscillator (NCO) for each of the nonlinearity terms, which over taxes FPGA resources.
[0010] Accordingly, there is a need for a method and system that overcomes, alleviates, and / or mitigates one or more of the aforementioned problems and other deleterious effects of existing PIM detection and cancellation methods.SUMMARY
[0011] Accordingly, it is desired to provide a system and method that makes the proposed dual-band PIM cancellation feasible, and in the meantime, reduces complexity of dual band signal and lowers resource utilization for cancellation of PIM in dual-band wireless networks.
[0012] It is further desired to provide a system and method that accounts for the independent nature of delays from each of the dual-band transmission base bands and for the combined nonlinear composite signals due to the PIM components.
[0013] It is still further desired to provide a system and method for cancelling PIM in a dual-band wireless network that deals with third (IM3) and fifth (IM5) order intermodulation where carrier frequencies of these nonlinear terms are different to the carrier frequency of the Rx signal.
[0014] The detailed technologies involved in this disclosure are: 1) filtering and decimation are proposed to reduce complexity and resource utilization, making it feasible in the major FPGA / ASIC solutions; 2) a novel PIM detection (PIMD) method is proposed for detecting the nonlinear terms, which is high-order and the combination of the two baseband signals in the dual band-band setup; and (3) a method to calculate the PIM nonlinear term by generating two pre-processing NCOs and applying DUC / DDC differently, reducing the FPGA resource utilization..
[0015] Since PIM interference will go through duplexer and Rx channel, the delay, which might not be constant with frequency, a memory polynomial is proposed instead of a memoryless model for PIM interference.
[0016] The Tx and Rx implicitly are configured with different carrier frequencies, and furthermore, nonlinear components in a dual-band case might have a carrier frequency differing from the PIM nonlinear terms. To solve the carrier frequency mismatch issue, DUC / DDC is applied in PIM signal generation module.
[0017] PIMC Filtering & Decimation. In one configuration, it is proposed to process the PIM cancellation with filtering and decimation. The proposed filtering and decimation approaches reduces the cancellation sampling rate to mitigate the processing complexity and the resource utilization. This are the revolutionary proposal, which makes the PIM cancellation feasible in the major FPGA / ASIC solutions.
[0018] The first step proposed is called PIMC filtering. The spectrum of the Rx signal is complicated, including the PIM interference, high power in the B3 DL to UL transition band, B3 DL leakage, and B1 DL leakage. A noise floor is underneath the interference. PIMC filtering filters out the unwanted spectrum, such as the transition band, and DL bands.
[0019] The second step proposed is called PIMC decimation. In this step, down-sampling is performed on the filtered signal to further reduce the complexity and resource utilization significantly.
[0020] Novel PIM Detection (PIMD) Method. The PIM interference signal or the nonlinear terms are the high-order and the combination of the two baseband signals of the dual-band transmission. The purpose of PIMD is to detect the delays from the two transmission baseband signals (dual-band DL signal) to the Rx signal (wanted UL signal) .
[0021] As stated previously, a major problem is that the delays from each of the transmission base bands to the received signal are independent and different as they depend partly on the independent delays caused by the two independent transmission circuits. The group delay at the cavity filter is different for different DL bands contributes to the delay difference for the PIM detection. The task for PIMD is to detect the two delays and align the two DL signals and UL signal in the further PIM cancellation steps. Additionally, the PIM components available for detection are often combined nonlinear composite signals. In this case, the signal is a two-dimensional function on both delays.
[0022] The proposed method is basically a two-dimensional grid search method for the two-dimensional cross correlation. The PIM nonlinear signal grids are reconstructed by applying the independent delays to the original transmission signal. The proposed method further includes two consecutive steps: 1) the coarse grid search, which could achieve the sample level delay detection; and 2) the fine grid search, which could achieve the sub-sample level delay detection. The current implementation can detect the delays to 0.1 sample.
[0023] PIM Nonlinearity Terms Calculation. In the modelling of the PIM, there are multiple nonlinear terms (typically six or seven terms) involved even if only the third and fifth order of the intermodulation (IM3 and IM5) are considered. The challenge is that the carrier frequencies of these nonlinear terms are different to the carrier frequency of the Rx signal, of which the carrier frequency is the center frequency of the UL band.
[0024] The new method introduces the pre-nonlinearity NCOs, which applies directly on the baseband transmission signals. The number of NCO and DDC / DUC modules are significantly reduced, and as a result, there is significant reduction in computational complexity and FPGA resource utilization.
[0025] An example is now provided to explain the new method, including the reduction of NCO and DDC / DUC. The example involves three nonlinear terms in Table 1. The nonlinear terms are g2 (x1, x2) =x1|x2|4; and It should be noted that x1and x2 are the baseband signals of the DL bands. The carrier frequency of the term in the RF domain are 2f1-f2, f1, and 3f1-2f2, respectively, where f1 and f2 are the carrier frequencies of the dual DL bands. Denote the UL band carrier frequency as fUL.
[0026] The inputs of the processing are: x1 and x2, f1, f2 and fUL. The outputs of the processing are and It should be noted that Δf1, Δf2, and Δf6 are given by: Δf1=2f2-f1-fUL; Δf2=f1-fUL; and Δf6=3f2-2f1-fULrespectively.
[0027] From the above, it should be clear that direct processing needs three NCOs and three DDC / DUCs.
[0028] The proposed method is introducing a pre-nonlinearity processing step. First two new terms are calculated: and where fz1=f1-fUL and fz2=f2-fUL. Then, calculate the nonlinear terms based on z1 and z2, which gives and The DDC / DUC (Δf1, Δf2, and Δf6) for the nonlinearity terms are automatically achieved, no need for the DDC / DUC processing in the post-nonlinearity stage.
[0029] The proposed processing needs two NCOs and two DDC / DUCs. It saves more NCOs and DDC / DUCs when more nonlinearity terms are involved.
[0030] For this application the following terms and definitions shall apply:
[0031] The term “down-sampling” as used herein means the process of reducing the sampling rate of a signal.
[0032] The term “decimation” as used herein means the process of reducing the sampling rate of a signal by a value and don’ t distort the interested spectrum of the signal by applying a decimation filter. Decimation works by merging every N sample into one. The decimation factor is an integer or a rational fraction greater than one.
[0033] The term “numerically controlled oscillator (NCO) ” as used herein is a digital signal generator that generates a synchronous (i.e., clocked) , discrete-time, discrete-valued representation of a waveform.
[0034] The term “data” as used herein means any indicia, signals, marks, symbols, domains, symbol sets, representations, and any other physical form or forms representing information, whether permanent or temporary, whether visible, audible, acoustic, electric, magnetic, electromagnetic or otherwise manifested. The term “data” as used to represent predetermined information in one physical form shall be deemed to encompass any and all representations of the same predetermined information in a different physical form or forms.
[0035] The term “passive intermodulation (PIM) ” as used herein is a form of distortion / signal interference that occurs when two or more signals mix in a passive, unpowered component, such as a cable, connector, fastener or the like.
[0036] The term “third-order intermodulation (IM3) distortion” as used herein is a measure of distortion in a nonlinear device when two signals with similar frequencies are present.
[0037] The term “fifth-order intermodulation (IM5) distortion” as used herein is a measure of distortion in a nonlinear device when two signals with similar frequencies are present.
[0038] The term “network” as used herein includes both networks and internetworks of all kinds, including the Internet, and is not limited to any particular type of network or inter-network.
[0039] The terms “first” and “second” are used to distinguish one element, set, data, object or thing from another, and are not used to designate relative position or arrangement in time.
[0040] The terms “coupled” , “coupled to” , “coupled with” , “connected” , “connected to” , and “connected with” as used herein each mean a relationship between or among two or more devices, apparatus, files, programs, applications, media, components, networks, systems, subsystems, and / or means, constituting any one or more of (a) a connection, whether direct or through one or more other devices, apparatus, files, programs, applications, media, components, networks, systems, subsystems, or means, (b) a communications relationship, whether direct or through one or more other devices, apparatus, files, programs, applications, media, components, networks, systems, subsystems, or means, and / or (c) a functional relationship in which the operation of any one or more devices, apparatus, files, programs, applications, media, components, networks, systems, subsystems, or means depends, in whole or in part, on the operation of any one or more others thereof.
[0041] The term "automatic" and variations thereof, as used herein, refers to any process or operation done without material human input when the process or operation is performed. However, a process or operation can be automatic, even though performance of the process or operation uses material or immaterial human input, if the input is received before performance of the process or operation. Human input is deemed to be material if such input influences how the process or operation will be performed. Human input that consents to the performance of the process or operation is not deemed to be "material. "
[0042] In one configuration, a method for passive intermodulation (PIM) cancellation in a dual-band wireless network is provided that comprises the steps of detecting non-linear products caused by non-linear characteristics of passive components in the dual-band wireless network, generating a first preprocessing numerically controlled oscillator (NCO) , and generating a second preprocessing NCO. The method further comprises the step of performing Digital Up-Conversion (DUC) or Digital Down-Conversion (DDC) on third and fifth order of intermodulation (IM3 / IM5) terms of x1 with the first preprocessing NCO, where x1 is a first baseband signal of a first DownLink (DL) band of the dual-band wireless network to generate z1. The method still further comprises the steps of performing DUC or DDC on IM3 / IM5 terms of x2 with the second preprocessing NCO, where x2 is a second baseband signal of a second DL band of the dual-band wireless network to generate z2, and calculating at least three first nonlinear terms of the carrier frequency of the first DL band based on z1. Also, the method comprises the steps of calculating second nonlinear terms of the carrier frequency of the second DL band based on z2, and cancelling the first and second nonlinear terms in an UpLink (UL) band of a Receive (Rx) signal.
[0043] The above-described and other features and advantages of the present disclosure will be appreciated and understood by those skilled in the art from the following detailed description, drawings, and appended claims.DESCRIPTION OF THE DRAWINGS
[0044] FIG. 1 depicts the Spectrum of the Rx signal, taking B66 Rx as an example.
[0045] FIG. 2A is a functional block diagram for PIM Identification (PIMI) .
[0046] FIG. 2B is a functional block diagram for PIM Cancellation (PIMC) .
[0047] FIG. 3 is an illustration of PIMC performance comparison for filtering and decimation.
[0048] FIG. 4 is a functional simplified block diagram showing different delays from the dual band transmission baseband to the Rx signal.
[0049] FIG. 5 is an example of a graph illustrating the delay response of a cavity filter.
[0050] FIG. 6 shows the IM3 and IM5 non-linearity terms in the separate dual-band PIM non-linearity modelling, taking B25 and B66 dual-band as an example.
[0051] FIG. 7 is a process flow chart of the proposed method for calculating the PIM nonlinear terms.DETAILED DESCRIPTION
[0052] Filtering and decimation in PIMI and PIMC. FIG. 1 shows the spectrum at B66 Rx input (right before the LNA) . The spectrum is complicated: we can see the PIM interference, high power in the DL to UL transition band, and B25 and B66 DL leakage. The noise floor is underneath the interference.
[0053] The spectrum of the Rx signal, taking B66 as an example, is shown. The center frequency of the Rx signal is supposed to be at the carrier frequency of the B66 UL band, which is 1745MHz. The plot also shows the processing band with and without decimation filtering. The full B66 UL band is falling into the processing band with decimation, while, in the meantime, the unwanted spectrum components, such as the DL UL transition band, the B25DL leakage band, are not included.
[0054] The digital domain signal at the original sampling rate (491.52MHz) without any decimation contains most of the above components. To precisely model the received signal and cancel out, it requires excessive resources. As such, a two-step process to mitigate the processing complexity and reduce the resource utilization is proposed.
[0055] First, the UL band modeling and cancellation are addressed by filtering out the unwanted spectrum, such as the transition band, and DL bands.
[0056] Second, down-sampling on the filtered signal is performed to further reduce complexity and resource utilization.
[0057] Turning now to FIGS. 2A and 2B, FIG. 2A shows a functional block diagram for PIMI, while FIG. 2B shows a functional block diagram for PIMC. A method is provided that applies to both PIM identification (PIMI) and PIM cancellation (PIMC) process. PIMI is implemented in software and on the captured data, which is the learning process of PIM cancellation. The PIMC processing is implemented in a Field Programmable Gate Array (FPGA) because it requires real time processing. The processing block diagrams for PIMI and PIMC are shown in FIGS. 2A and 2B respectively.
[0058] FIG. 3 is a graph the complexity and resource utilization reduction by filtering and decimation. The figure shows the PIMC in dB is a function on the memory depth of the MP modeling. The PIM modelling is using memory polynomial (MP) . The memory depth is the key parameter, proportional to the complexity and resource utilization. The circle is the typical working range of PIM cancellation, we could see that the memory depth is doubled from with filtering and with decimation to with filtering and no decimation. And the memory depth is further doubled from with filtering and no decimation to no filtering and no decimation. It is one fourth memory depth for with both filtering and decimation comparing to no filtering and no decimation, a significant resource reduction.
[0059] The PIM modelling is using memory polynomial (MP) . The memory depth is the key parameter reflecting the complexity and resource utilization. Typically, the required PIM cancelation is ~17dB (for -130dBc PIM) . The curves show that the memory depth for filtering & decimation, filtering & non-decimation, and non-filtering & non-decimation are 8, 15, and 33, respectively. M is proportional to the computational complexity and the resource utilization.
[0060] PIM Detection (PIMD) Method. Referring to FIGS. 4 –5, FIG. 4 is a simplified system block diagram showing different delays from the dual band transmission baseband to the Rx signal. FIG. 5 depicts an example of a delay response of a cavity filter.
[0061] In FIG. 4, PIM happens at point E. However, the delays from x1 and x2 to r are not equal. The delays need to be estimated in PIM detection. The delays are denoted as: τ1 and τ2. A and B are the starting points for PIM cancellation algorithm, C is the ending point for PIM cancelation algorithm, where the cancellation occurs. The signals are A, B, and C. The delays from A and B to C are different due to the signals going through different Tx circuits to E. The purpose of PIMD is to detect the delays of from A to C, and from B to C.
[0062] The delay response of a cavity filter has significant variation between pass and reject band. FIG. 6 shows an example of a delay response of a cavity filter. As a result, even when the delays of the DL signals could be detected, the detected delays are not usable for the DL signal as the delays for the UL signal, which is the interested target signal in PIM modeling.
[0063] PIM detection solution: The PIM detection is done in the interested UL bands. The corresponding Digital Up-Conversion or Digital Down-Conversion (DUC / DDC) and filtering are proposed.
[0064] In the UL bands, the terms used for PIMD are IM3 terms x1|x1|2, x1|x2|2, and PIM detection on these terms is more complicated than on x1 and x2.
[0065] PIMD for τ1. Filtered IM3 term x1|x1|2 is used for τ1 detection. The algorithm is straight forward by calculating the cross correlation. PIMD for τ2 . τ2 can be detected by doing coarse and fine time alignment on either the term of x1|x2|2 or The algorithm is more complicated than detecting τ1.
[0066] Denote v (t) is the signal of x1 (t-τ1) |x2 (t) |2 passing through a LP filter, τ1 is PIMD output for τ1. However, it should be noted that v (t-τ2) ≠x1 (t-τ1) |x2 (t-τ2) |2. Cross-correlation cannot be directly applied to find the coarse alignment. v (t-τ2) =x1 (t-τ1-τ2) |x2 (t-τ2) |2.
[0067] Build a group of signals v (t, τ2 (k) ) =x1 (t-τ1) |x2 (t-τ2 (k) ) |2, k=-K, …, K, where τ2 (k) =k / fs, fs is the sampling rate. Then, calculate ck=∑tr* (t) v (t, τ2 (k) ) . Search the maximum |ck|, then kmax / fs is the coarse time alignment (sample level) , denoted by τs.
[0068] Build a group of signals v (t, τ2 (k) ) =x1 (t-τ1) |x2 (t-τs-τ2 (k) ) |2, k=-K, …, K, where τ2 (k) =k / 2Kfs, fs is the sampling rate, 1 / 2Kfs is the fine alignment resolution.
[0069] Then, calculate ck=∑tr* (t) v (t, τ2 (k) ) . Search the maximum |ck|, kmax / 2Kfs is the fine time alignment (sample level) , denoted by τf.
[0070] The overall delay is therefore τ2=τs+τf.
[0071] PIM Nonlinearity Terms Calculation. A method is further proposed to calculate the PIM nonlinear term. The proposal is to generate two pre-processing numerically controlled oscillator (NCOs) and DUC / DDCs for the DL baseband signal instead of the conventional post processing NCOs and DUC / DDC.
[0072] The PIMC algorithm is based on the proposed separate dual-band PIM non-linearity modelling. With the dual band PIM modeling, the PIM non-linearity includes a few terms, each of which has comparably quite narrow BW. As such, the processing requirement for the modelling is lowered. With the PIM modeling, PIM nonlinear terms must be cancelled out for each UL band.
[0073] Taking B25 & B66 as an example, FIG. 6 shows the frequency range of the PIM nonlinear terms. The 3rd order terms (IM3) , and 5th order terms (IM5) are shown, respectively. The solid lines outline the UL range of B25 and B66. If a PIM nonlinear term overlaps the UL bands, it is a potential interference term and must to be cancelled out. The 3rd order nonlinearity terms are shown in the upper six boxes. The other boxes are the 5th order nonlinearity terms. The location of the DL and UL of B25 and B66 are shown in the dashed and solid lines. If a term overlaps with one of the UL bands, it is an interfering PIM term for that UL band. The mathematical form for each term is shown in the left column of Table 1. Table 1
[0074] Table 1 shows the IM3 and IM5 nonlinear terms need to be cancelled out for B66. The terms are denoted by g1 (x1, x2) to g7 (x1, x2) , where x1 and x2 are the baseband signals of the DL bands. The column of” fc” indicates the carrier frequency of the term in the RF domain. f1 and f2 are the carrier frequencies of the DL bands. In this example, x1 and f1 are for B25 DL, and x2 and f2 are for B66 DL.
[0075] FIG. 7 illustrates a process flow chart of the proposed method for calculating the PIM nonlinear terms. The steps for calculating the nonlinear terms are as follows:
[0076] Step O1, calculate the nonlinear kernel functions g1 (x1, x2) to g7 (x1, x2) in Table 1.
[0077] Step O2, generate the numerically controlled oscillator (NCO) for the nonlinear terms.
[0078] Step O3, perform digital up conversion or down conversion (DUC / DDC) on the output in step 1 by using the NCOs generated in step 2.
[0079] Steps 2 and 3 align the baseband signal of the nonlinear terms to the UL receive signal. The NCO values are calculated by fc-fUL, where fc is the values in the “fc” column, and fUL is carrier frequency of the targeting UL band.
[0080] [Rectified under Rule 91, 01.04.2025]Additionally, a process to mitigate the processing complexity and reduce the resource utilization is provided. The processing steps are also shown in FIG. 7.
[0081] Step P1, first generate the NCOs, which is applied to x1 and x2, instead of directly calculating the nonlinear kernel functions g1 (x1, x2) to g7 (x1, x2) .
[0082] Step P2, perform DUC / DDC on x1 and x2 by using the NCOs generated in step 1. Denote the output as z1 and z2.
[0083] Step P3, calculate the nonlinear kernel functions g1 (z1, z2) to g7 (z1, z2) .
[0084] Step P4, generate the post processing NCOs and perform the post processing DUC / DDC if needed. In this example, step 4 is not needed.
[0085] The complexity and resource utilization are reduced with the above-described method. The computational complexity for O1 and P3 are the same. The number of NCOs is 3 versus 2 for the old methods as opposed to the new methods. Additionally, the number of DUC / DDCs is 7 versus 2 for the old as opposed to the new methods.
[0086] The resource utilization of the FPGA shows that, if taking both B25 and B66 into account, the number of DSP slices is 69 versus 42 for the old methods as opposed to the new methods for one set of dual band transceivers. For the 4T / 4R dual band implementation, the total saving is 108 DSP slices, which is significant. It should be noted that NCOs are configured not using DSP slices.
[0087] The proposed methods could apply to tri-band scenarios without modification of the basic algorithm. The interference terms in the tri-band case must be identified. The proposed methods are dealing with the whole band instead of a single or multiple carriers within each band. It is adaptive to carrier configuration adjustment after the deployment. Because the whole band is considered, the PIM cancellation can support the PIM interference cancellation for NB-IoT in-band, NB-IoT guard band and Cat-M.
[0088] While the present disclosure has been described with reference to one or more exemplary embodiments, it will be understood by those skilled in the art that various changes can be made and equivalents can be substituted for elements thereof without departing from the scope of the present disclosure. In addition, many modifications can be made to adapt a particular situation or material to the teachings of the disclosure without departing from the scope thereof. Therefore, it is intended that the present disclosure not be limited to the particular embodiment (s) disclosed as the best mode contemplated, but that the disclosure will include all embodiments falling within the scope of the appended claims.
Claims
A method for passive intermodulation (PIM) cancellation in a dual-band wireless network comprising the steps of:detecting a non-linear product caused by non-linear characteristics of passive components in the dual-band wireless network;generating a first preprocessing numerically controlled oscillator (NCO) ;generating a second preprocessing NCO;performing Digital Up-Conversion (DUC) or Digital Down-Conversion (DDC) on third and fifth order of intermodulation (IM3 / IM5) terms of x1 with the first preprocessing NCO, where x1 is a first baseband signal of a first DownLink (DL) band of the dual-band wireless network to generate z1;performing DUC or DDC on IM3 / IM5 terms of x2 with the second preprocessing NCO, where x2 is a second baseband signal of a second DL band of the dual-band wireless network to generate z2;calculating at least three first nonlinear terms of the carrier frequency of the first DL band based on z1;calculating second nonlinear terms of the carrier frequency of the second DL band based on z2;cancelling the first and second nonlinear terms in an UpLink (UL) band of a Receive (Rx) signal.The method of claim 1,wherein the step of performing DUC / DDC on x1 includes generating z1 according to the following calculation:where fz1=f1-fUL and where f1 is the carrier frequency of the first DL band, and fUL is the carrier frequency of the UL band of the Rx signal.The method of claim 2,wherein the step of performing DUC / DDC on x2 includes generating z2 according to the following calculation:where fz2=f2-fUL and where f2 is the carrier frequency of the second DL band.The method of claim 3, wherein the step of calculating the nonlinear terms of the first and second DL bands comprises calculating three nonlinear terms as follows,andrespectively;where Δf1, Δf2, and Δf6 are given by:Δf1=2f2-f1-fUL;Δf2=f1-fUL; andΔf6=3f2-2f1-fUL.The method of claim 4, wherein the carrier frequency of the term in the RF domain for each of the three nonlinear terms are,2f1-f2, for the first term,f1, for the second term, and3f1-2f2, for the third term.The method of claim 5, wherein the passive components are selected from the group consisting of: filters, connectors, antenna elements, and any combinations thereof.The method of claim 6, further comprising the steps of:filtering out a DL to UL transition band of the Rx signal;filtering out first DL leakage from the first DL band; andfiltering out second DL leakage from the second DL band to generate a filtered signal.The method of claim 7, further comprising the step of:performing down-sampling on the filtered signal to further reduce the complexity and resource utilization.The method of claim 1, wherein delays from the first baseband signal and the second baseband signal are independent and different from each other.