Compression detection for IQ data

Abstract

A, preferably computer-implemented, method of processing in- phase and quadrature data, IQ data, the method comprising the steps of : determining a plurality of IQ samples from the IQ data, determining a constellation error of each one of the plurality of IQ samples, and identifying a (de-)compression method and / or IQ bit width of the IQ data based on the constellation error.

Description

[0001] DESCRIPTION

[0002] TITLE

[0003] Compression detection for IQ data

[0004] TECHNICAL FIELD

[0005] The presented disclosure relates to the technical field of radio communication and radio equipment. More particularly, the present disclosure relates to processing IQ (In-phase and Quadrature) data of a high speed transport link. The disclosure further relates to testing radio equipment such as a base station and its components.

[0006] BACKGROUND

[0007] In radio communication systems, the explosive growth of wireless data traffic requires more capacity of transport links such as those specified by the existing industry standards for modular designs, e.g., Common Public Radio Interface (CPRI) , Open Base Station Architecture Initiative, OBSAI, JESD204B, etc. CPRI refers to serial data links between a Remote Radio Head, RRH, which is also known as Radio Unit, RU, and a Baseband Unit, BBU, which is also known as Digital or Distributed Unit, DU, or Radio Element Controller, REC. JESD204B refers to a JEDEC Standard for serial interface for data conversion including, e.g., Analog-Digital Converter (ADC) or Digital-Analog Converter (DAC) . One solution is to increase the link speed, for example, from 10G CPRI to 40G or 100G CPRI. These protocols and techniques are used to transport in-phase and quadrature, IQ, data. The IQ data represents the radio signal sampled and digitized, e.g., of one carrier at one independent antenna element, the so-called antenna carrier (AxC) .

[0008] Such transport links cannot keep up with the fast growing trend of wireless data traffic and becomes a bottleneck of the radio communication system. Thus, due to the increased traffic and limited bandwidth it has also become known to compress IQ data transmitted. For example, from European pa- tent application EP3195507A1 it has become known to applying nonlinear companding in order to compress the IQ data .

[0009] However, fronthaul data packet analyzers , such as Siemens ' Veloce X-STEP, need to be able to correctly decode the data packets . A fronthaul data packet analyzers may capture data packets , for example when transmitted between a DU and an RU, and utili ze these data packets to test and or veri fy the functionality of the radio equipment concerned . Hence , the fronthaul data packet analyzer requires information about the format of the data packet . This is particularly relevant to IQ data, for example as defined in ORAN ( Open Radio Access Networks ) speci fications . An Ethernet and / or ( e ) CPRI packet comprising the IQ data may include a plurality of optional parameters relating to the IQ data . For example , such a parameter may indicate a compression method used for compressing the IQ data . Similarly, the length of the IQ data, i . e . , a parameter relating to the IQ bit width, may be set between 1- 16 bits for the in-phase and the quadrature data, respectively . Since these parameters are only optionally included in the data packets , they may not be known to the fronthaul data packet analyzer . However, determining the IQ bit width of the IQ data is important for veri fying a configuration of the distributed unit and / or the radio unit . Conventionally, relevant fronthaul communication parameters are provided manually to the fronthaul data packet analyzer, i . e . , by user input , and thus is subj ect to human error . Additionally, in some circumstances , this information may not be known to the user and has to be found out by trial and error . Accordingly, there is a need for a method and packet analyzer which automatically determines the relevant fronthaul communication parameters such as bit length of the IQ sample . To that end, a method of determining a bit length of an IQ sample has become known from European patent application EP 21723384 . The method serves for automatically detecting i f user data compression header is in IQ data frames (U-plane ) and / or i f an IQ data frame comprises user data compression parameter . The method further serves for determining the length of I and Q bits , i . e . the IQ bit width .

[0010] SUMMARY

[0011] There still is a need for automatically determining the presence and / or settings of one or more parameters relating to the IQ data transmitted over the fronthaul to be intelligible for the fronthaul data packet analyzer and / or a user .

[0012] The obj ect is achieved according to the following aspects .

[0013] According to a first aspect the obj ect is achieved by a, preferably computer-implemented, method of processing in- phase and quadrature data, IQ data . The method comprising the step determining a plurality of IQ samples from the IQ data . The method further comprising the step of determining a constellation error of each one of the plurality of IQ samples . The method further comprising the step of identi fying a ( de- ) compression method and / or IQ bit width of the IQ data based on the constellation error .

[0014] According to a second aspect the obj ect is achieved by an apparatus operative to perform the method steps of the first aspect .

[0015] According to a third aspect the obj ect is achieved by an apparatus comprising a processor and a memory, wherein the memory comprises program code that when executed by the processor performs the methods steps of the first aspect .

[0016] According to a fourth aspect the obj ect is achieved by a computer program, preferably stored on a non-transitory medium, comprising program code that when executed performs the method steps of the first aspect .

[0017] According to a fi fth aspect the obj ect is achieved by a system comprising a radio unit and / or a distributed unit and a computer program, e . g . , on an apparatus according to the sec- ond or third aspect , for performing the method steps according to the first aspect .

[0018] BRIEF DESCRIPTION OF THE DRAWINGS

[0019] FIG 1 shows an Ethernet packet with eCPRI payload .

[0020] FIG 2 shows a base station and fronthaul ( communication link) connecting radio unit and distributed unit .

[0021] FIG 3 shows a fronthaul data packet analyzer connected to a fronthaul communication link between a radio unit and a distributed unit .

[0022] FIG 4 shows a fronthaul data packet analyzer connected to a radio unit over a fronthaul communication link .

[0023] FIG 5 shows a fronthaul data packet analyzer connected to a distributed unit over a fronthaul communication link .

[0024] FIG 6 shows compression of IQ data from raw, uncompressed IQ data into compressed IQ data .

[0025] FIG 7 shows a constellation diagram for rectangular 16-QAM ( quadrature amplitude modulation) .

[0026] FIG 8 shows a constellation diagram for 8-PSK (phase-shi ft keying) .

[0027] FIG 9 shows an error vector ( Perror ) of an IQ sample for QPSK ( quadrature PSK)

[0028] FIG 10 shows a constellation diagram comprising point clouds of IQ s amp les .

[0029] FIG 11 to 24 show exemplary methods steps of processing IQ data .

[0030] FIG 25 shows an exemplary fronthaul data packet analyzer .

[0031] DETAILED DESCRIPTION

[0032] FIG 1 shows an Ethernet packet with eCPRI payload . The Enhanced Common Public Radio Interface , eCPRI , packet that is encapsulated in an Ethernet packet . The typical transport use of eCPRI is the Ethernet protocol which is packet-based . Dependent on the eCPRI Message Type the payload of an eCPRI message may comprise for example IQ data . eCPRI leverages the highspeed Ethernet available in the market such as l OGbps , 25Gbps , 40Gbps , l O OGbps and above . Thus , Ethernet in combination with eCPRI may be used as a fronthaul protocol for transporting IQ data to, from and / or between a radio unit and / or a distributed unit.

[0033] The payload size is specified in a common header in accordance with the eCPRI specifications. The eCPRI common header includes an eCPRI protocol revision parameter indicative of protocol version, reserved section which includes one or more reserved bits in accordance with the eCPRI specifications, an eCPRI messages concatenation indicator, an eCPRI message type parameter which indicates the type of services conveyed by the data frame and / or an eCPRI payload size parameter which indicates the size in bytes of the payload part corresponding to the data frame. Instead of eCPRI, CPRI or any other fronthaul protocol may be used. The payload may thus comprise a number of bits representing the IQ data. In the case of eCPRi the payload may comprise 0-65535 bytes. Now, the IQ data in the payload may or may not be compressed. That is the IQ data in the payload may be compressed IQ data. Furthermore, the uncompressed IQ data may possess a certain IQ bit width. Thus, herein the IQ bit width relates to the bitwidth of the uncompressed IQ data.

[0034] As shown in FIG 2, a fronthaul (network) is used to connect base stations, or the components thereof which are also referred to as distributed unit, DU, and radio unit, RU, but which also go by the name of (evolved) radio equipment control, (e)REC, and (evolved) radio equipment, (e)RE. Hence, a base station eNB, gNB may comprise one or more radio units and / or distributed units. The transfer of IQ data over the fronthaul (communication link) may then be provided using data packet based communication protocol such as Ethernet or IP. For user plane traffic over Ethernet, eCPRi messages are transmitted in Ethernet frames, also referred to as data packets. For user plane traffic over IP, eCPRi messages may be transmitted in UDP / IP packets.

[0035] As shown in FIG 3, in the uplink, one or more RU may compress IQ data for transmission to one or more DU. Thus, the RU may forward the compressed IQ data to a DU through a fronthaul communication link - or the other way round for the downlink . To achieve fronthaul compression, di f ferent compression methods have become known . When, for example , a low implementation complexity is preferred, the point-to-point compression and spatial filtering are recommended . When signals received at di f ferent RUs are closely correlated, the distributed source coding should be used . Considering the sparsity of the uplink C-RAN signals , compressed-sensing based compression is another alternative . At the DU, decompression and decoding is performed .

[0036] Now, for the veri fication and testing of 5G radio units , distributed units , and / or fronthaul network components a fronthaul data packet analyzer UM, such as Siemens ' Veloce X- STEP, may be used . Such fronthaul data packet analyzer, EM, may be looped in the communication link between the RU and DU, or may be connected to the RU and / or DU via a fronthaul communication link - as shown in FIG 3 .

[0037] Either way, a fronthaul data packet analyzer EM may be connected to a RU over a fronthaul communication link as shown in FIG 4 or may connected to a DU over a fronthaul communication link as shown in FIG 5 . Hence , the fronthaul data packet analyzer may perform the steps as j ust discussed in connection with the DU and / or RU, i . e . , may need to perform decompression and / or decoding of the ( eventually compressed) IQ data .

[0038] When monitoring traf fic over the fronthaul communication link, e . g . , using a fronthaul protocol such as ( e ) CPRI , the fronthaul data packet analyzer FM may extract the IQ data and process the IQ data into data points , i . e . IQ samples . These IQ samples may for example be plotted on a display of the fronthaul data packet analyzer or a display communicatively coupled to the fronthaul data packet analyzer . To that end, parameters of the fronthaul protocol are needed for the fronthaul data packet analyzer to process the IQ data correctly, e.g., to correctly (de- ) compress and / or (de-) code the IQ data. It should be noted that the configurations of the fronthaul protocol may be different for the uplink and the downlink. The IQ data in the uplink direction may represent radio signals from a user equipment. The IQ data in the downlink direction may represent IQ data transmitted from DU to RU. In any case, the IQ data is transmitted as a baseband signal, e.g., at a baseband frequency.

[0039] As discussed, IQ data representing radio signal are transported between the RU and DU over the fronthaul (communication link) . The fronthaul communication link usually uses a high data rate to meet the radio signal transport requirements. Many solutions have been proposed to decrease the fronthaul data rate, such as functional split and / or IQ data compression. For example, the "WG4 - The Open Fronthaul Interfaces Workgroup" defines standardization for fronthaul interfaces between BBU and radio equipment. The WG4 works on providing specifications and standards for a secure fronthaul, allowing multi-vendor interoperability.

[0040] O-RAN WG4 currently defines 7 different User Plane IQ compression methods, including the possibility of not applying any compression to the IQ data. As the case may be, the IQ data may be in the frequency domain. In addition, the resolution of IQ samples can theoretically range from 1 to 16 bits. In dynamic compression, U-Plane data packets carry the ud- CompHdr field that explicitly defines the compression method and / or IQ resolution applied for at least a section of IQ data. Herein, the IQ resolution is also referred to as IQ bit width. However, the udCompHdr is absent from U-Plane messages when the IQ resolution and / or compression method are statically configured, e.g., via the M-Plane upon O-RAN Radio Unit initialization. In this case, a device, such as a fronthaul data packet analyzer FM, which for example is monitoring and / or analyzing, e.g., O-RAN U-Plane, fronthaul traffic, would need to be able to automatically detect the (de- ) compression method and / or IQ resolution . It is thus necessary to provide such a fronthaul monitor, which is preferably able to determine U-Plane statistics and IQ analysis and / or display the results to a user .

[0041] For example , i f the RU supports static channel information compression through M-Plane configuration and static compression is configured, i . e . , the fronthaul communication link is statically configured, a fronthaul data packet analyzer may not be aware of the compression and / or IQ bit width of the IQ data, since the information is not comprised in the one or more data packets .

[0042] On the other hand, i f the RU supports dynamic channel information compression, then the DU may configure dynamic compression via M-Plane with or without providing a list of compression methods that will be used dynamically ( i . e . , a subset of methods supported by O-RU) . In that case , the compression method and / or IQ bit width may change and again the fronthaul data packet analyzer may not be aware of the updated compression and / or IQ bit width of the IQ data .

[0043] Thus , the udCompHdr parameter may be absent from U-Plane messages when the static IQ format and compression method is configured, e . g . , via the MPlane . In this way a single compression method and / or IQ bit width is provided (per UL and DI , per LTE and NR) . This has the advantage that no more overhead to U-Plane messages . However, again the fronthaul data packet analyzer may not be aware of the compression and / or IQ bit width of the IQ data . An overview of the possible parameters of the udCompHdr is provided in 8 . 3 . 3 . 13 f f . of O-RAN . WG4 . CUS . 0-R003-V12 . 00 . The IQ data and exponent packeti zation for di f ferent IQ bit widths is provided in Annex D of O-RAN . WG4 . CUS . 0-R003-V12 . 00 . Furthermore , some of the available compression methods such as Block floating point compression, Block scaling compression, p-Law compression, Beamspace compression, Modulation compression, and Se- lective RE sending are provided in Annex A of 0- RAN.WG4.CUS .0-R003-V12.00.

[0044] In FIG 6 compression of IQ data is shown, wherein, e.g., for each PRB, uncompressed In-phase (I) and Quadrature (Q) components are converted into a floating point format. Any other compression method may be used instead of the floating point compression shown. As a result, compressed IQ data is obtained that represents the (original or uncompressed) IQ data having an IQ bit width. The compression procedure receives 12 subcarriers with 24 uncompressed IQ data. The IQ data is subsequently compressed to a bit signed mantissa and unsigned exponent. The IQ data is thus represented as a compressed bit sign and a mantissa (e.g., 9-bit) and a shared exponent.

[0045] Fronthaul monitoring devices on the market require the compression and / or IQ resolution, i.e., IQ bit width, information to be input by the user. However, users don't always know the exact parameters and the analysis of U-Plane data becomes a trial and error with potentially several tens of compression-resolution combinations to be tried.

[0046] The IQ bit width, also referred to as resolution or sample width, is the bit width or the number of bits for the digitized uplink and / or downlink radio signals, i.e. the IQ data. The IQ bit width value is determined by the radio equipment manufacturers or operator. Selectable values are in the range between 1 to 16 bits for the In-phase (I) and Quadrature (Q) component, respectively. Given unlimited bit depth and sample rate, IQ data can perfectly recreate a radio signal from any period in time. In the real world, however, more bit depth and more samples come at a cost of more bits, and these bits can quickly consume gigabytes of storage and tens of gigabits per second of streaming bandwidth. Hence, usually an optimal compression and / or IQ bit width is employed.

[0047] U-Plane traffic comprises IQ data, i.e., digitally modulated signals, and each correctly decompressed and / or correctly in- terpreted IQ sample should be located very close to a constellation point on a constellation diagram. IQ data, i.e. each pair of I component and Q component, represents a complex number which may be mapped into a constellation diagram. The IQ data may thus comprise a sampling sequence of in-phase and quadrature components of a complex digital signal. This IQ data may thus be transmitted over the fronthaul communication link. For example, the (e)CPRI payload of a U-plane message can be used to transmit the IQ data, e.g., applying IQ data compression and / or comprising compression information in the udCompHdr . The compression information may comprise the applied compression scheme and / or the number of bits in the IQ sample before and / or after compression. The IQ data compression is performed using a compression parameter such as udCompParam. For example, when applying the block floating point compression method, the IQ data compression parameter and (original or uncompressed) IQ data are represented as an exponent and a mantissa in floating point form. For example, compression information, e.g., in the form of a parameter, such as udCompMeth, may indicate the (de- ) compression method. The compression information, e.g., the same parameter or a different parameter, such as udlqWidth, may indicate the bitwidth of each (un- ) compressed I and Q value, as for example given in Table 8-4 of ETSI TS 103859 V7.0.2 (2022-09) , i.e. O-RAN-WG4. CUS .0-v07.02. Hence, the compression information, such as udlqWidth, may indicate the number of bits after compression, i.e. of the compressed IQ data. The number of bits of the uncompressed IQ data that encode an IQ sample may thus be different, i.e. higher, than the number of bits of the compressed IQ data.

[0048] For example, due to various imperfections in the transmission, reception, and packetization of User Plane data, IQ samples carried in the fronthaul (in the form of IQ data) are not exactly in their ideal constellation points. In practice, there will be a level of error in the amplitude and / or phase of each IQ sample. The diagram for QPSK (quadrature PSK) in FIG 9 shows an IQ sample at an error vector (Perror) away from the ideal constellation point of "H" in binary value.

[0049] A constellation diagram is a representation of a signal modulated by a digital modulation scheme such as quadrature amplitude modulation or phase-shift keying. It displays the signal as a two-dimensional x-y-plane scatter diagram in the complex plane at sampling instants. In a digital modulation system, information is transmitted as a series of samples, each occupying a uniform time slot. During each sample, the carrier wave has a constant amplitude and phase, which is restricted to one of a finite number of values. Each sample encodes one of a finite number of "symbols", which in turn represent one or more binary digits (bits) of information. Each symbol is encoded as a different combination of amplitude and phase of the carrier, so each symbol is represented by a point on the constellation diagram, called a constellation point. The constellation diagram shows all the possible symbols that can be transmitted by the system, e.g., RU and / or DU, as a collection of points.

[0050] In FIG 7 a constellation diagram for rectangular 16-QAM (quadrature amplitude modulation) , with 16 constellation points around the origin, is shown. QAM is a combination of phase and amplitude modulation schemes, and 16-QAM carries log2 (16) = 4 bits of data per symbol. The black arrow represents the phase and amplitude of one IQ sample pointing at a constellation point with the binary value of "1001". In a frequency or phase modulated signal, the signal amplitude is constant, so the points lie on a circle around the origin.

[0051] In FIG 8 a constellation diagram for 8-PSK (phase-shift keying) , a form of phase modulation with 8 phase states around the origin is shown. Each valid symbol, i.e., a constellation point carries log2 (8) = 3 bits of information. The black arrow in the drawing represents the phase of one IQ sample pointing at a constellation point of "110" as its binary value . As mentioned herein, due to various imperfections in the transmission, reception, and packetization of User Plane data, IQ samples carried in the fronthaul are not exactly in their ideal constellation points. In practice, there will be a level of error in the amplitude and phase of each IQ sample .

[0052] The diagram in FIG 9 for QPSK (quadrature PSK) shows an IQ sample at an error vector (Perror) away from the ideal constellation point of "11" in binary value. Therein the constellation point vector is given as Pref. Thus, in FIG 9 a constellation error in the form of an error vector of an IQ sample is shown. The error vector has a direction and a length. Thus, a plurality of IQ samples result in a point cloud around the ideal constellation point, or are given as a plurality of error vectors.

[0053] In FIG 10, a constellation diagram of 16-QAM comprising such point clouds around the 16 ideal constellation points is shown. The average radius of the point clouds, i.e., the root mean square (RMS) average amplitude of the error vector, normalized to ideal signal amplitude reference, is referred to as the error vector magnitude (EVM) . EVM, sometimes also called relative constellation error or RCE, is a measure that may be used to quantify the performance of a digital radio transmitter or receiver. Informally, EVM is a measure of how far the points are from the ideal locations.

[0054] EVM may be used for automatically detecting compression and / or resolution parameters of the IQ data received, e.g., for User Plane data. When receiving O-RAN U-Plane data packets at an unknown compression and / or resolution, an evaluation of the IQ data and / or the data packets comprising the IQ data may be carried out, and IQ samples for each combination of compression methods and / or resolutions. That is, given 7 different compression methods and 16 different resolutions, 112 variants of the IQ samples may be obtained. Then, a EVM calculation is performed for each of the variants , and the compression-resolution combination with the lowest resulting EVM is identi fied, e . g . , as a result of the auto-detection . Naturally, the EVM value must be within acceptable range for the modulation scheme , so that the recovery of information from the constellation diagram is reliable and the point clouds are not at risk of getting mixed .

[0055] The proposed embodiments provided can be automated and / or can be implemented in software with existing ( de- ) compression algorithms and / or EVM calculation . The fronthaul data packet analyzer is not aware of a management plane request of compression details . The proposed embodiments do require no human intervention in terms of suggesting compression methods and / or IQ resolutions , or any manual trial and error in terms of analyzing visuali zed constellation diagrams to determine i f the parameters were correct or not . Thus , an automatic IQ compression and resolution detection mechanism is proposed . The detected compression and / or resolution may be displayed on a display of the fronthaul data packet analyzer .

[0056] Further exemplary embodiments comprising one or more method steps of processing, for example by a fronthaul data packet analyzer, in-phase and quadrature data, IQ data, are shown in FIG 11 to 25 .

[0057] In a first step S I , a plurality of IQ samples are determined from the IQ data . Hitherto , one or more data packets comprising the IQ data may be received . The data packets may comprise the I and Q components representing radio signals in one or more physical resource blocks , PRBs . In a step S2 , a constellation error of each one of the plurality of IQ samples is determined . The IQ samples are expected to be close to the symbols in a constellation diagram i f the correct ( de- ) compression method and / or IQ bit width is applied, i . e . , the IQ data is processed with the correct ( de- ) compression method and / or IQ bit width . In a step S3, a (de- ) compression method and / or IQ bit width of the IQ data is identified based on the constellation error. Thus, if the constellation error, e.g., the EVM of the plurality of IQ samples, is sufficiently close to the constellation points or symbols (of a constellation diagram) a (de- ) compression method and / or a bit width applied may be identified as the correct one. For example, if the constellation error exceeds a predetermined threshold a (de- ) compression method and / or IQ bit width may be identified, e.g., as correct or incorrect. In case the (de- ) compression method is identified as correct the (de- ) compression method can be chosen for configuring the fronthaul data packet analyzer or for verifying functionality of the radio unit and / or distributed or a device under test in general. In case the a (de- ) compression method and / or IQ bit width determined as incorrect, the a (de- ) compression method and / or IQ bit width may be discarded. In that case another a (de- ) compression method and / or IQ bit width may be chosen to determine a plurality IQ samples from the IQ data (received in said one or more data packets) .

[0058] Turning to FIG 12, in a step S4, a plurality of IQ samples may be determined from the IQ data by decompressing the IQ data according to a plurality of different (de- ) compression methods. In a step S5, the plurality of IQ samples are determined from the IQ data by using a plurality of different IQ bit widths for the (uncompressed) IQ data. It may be the case that at first IQ samples according to each one of the plurality of different (de- ) compression methods and / or IQ bit widths are determined. Alternatively, it may be the case that IQ samples are determined according to a first (de-

[0059] ) compression method and / or a first IQ bit width and that only after determining the constellation error of the said IQ samples (and eventually determining whether the constellation error is sufficiently close to the constellation points in the constellation diagram - and that thus the chosen first (de- ) compression method and a first IQ bit width are correct or incorrect) a second (de- ) compression method and / or a sec- ond IQ bit width are used to determine IQ samples from the IQ data .

[0060] Turning to FIG 13, in a step S6, firstly the IQ data is decompressed according one of the plurality of decompression methods, and secondly in a step S7 one or more of the plurality of different IQ bit widths are used for determining a plurality of IQ samples from the (uncompressed) IQ data. Thus, the IQ data received may be decompressed according to the plurality of different decompression methods and uncompressed IQ data may be obtained for each one of the different decompression methods. Then, the plurality of different IQ bit widths may be applied to each one of the uncompressed IQ data (obtained by the respective decompression method) . As a result, IQ samples according to each possible combination of (de- ) compression method and IQ bit width are obtained. For the IQ samples obtained by said combinations a constellation error may be determined and the combination with the lowest constellation error or constellation error exceeding, e.g., below, a predetermined threshold may be identified.

[0061] Turning to FIG 14, in a step S8, the IQ data is decompressed according to a subset or a prioritized order of the plurality of different (de- ) compression methods. The order in which the (de- ) compression methods are applied may be prioritized according to user preferences and / or the most common (de- ) compression methods used in the field. Similarly, in a step S9 the IQ bit widths are used in a prioritized order and / or a subset of the plurality of different IQ bit widths for the (uncompressed) IQ data. Again, the order in which the IQ bit widths are applied may be prioritized according to user preferences and / or the most common (de- ) compression methods used in the field. Thereby, the correct (de- ) compression method and / or IQ bit width can be identified quicker.

[0062] Turning to FIG 15, in a step S10, firstly setting the IQ bit width to one of the plurality of IQ bit widths. That is, the IQ bit width is set in the fronthaul data packet analyzer, e.g., by setting a parameter value (for processing the uncom- pressed IQ data) according to the IQ bit width, for example between 1 and 16 for the I and the Q component, respectively. For example, the bit resolution or IQ bit Width may be set to the most typical case, e.g., 9bits or 15bits. Then, the different decompression methods may be applied.

[0063] In a step Sil, the IQ data is decompressed according to the plurality of different (de- ) compression methods. The

[0064] (un) compressed IQ may then be processed according to the previously set IQ bit width. If there the constellation error of the resulting IQ samples is not sufficiently close to the one or more constellation points another IQ bit width may be set to a different value and the decompression of the IQ data according to the plurality of different (de- ) compression methods may be performed again, e.g., until the resulting IQ samples are sufficiently close to the constellation points.

[0065] Turning to FIG 16, in a step S12, one or more combinations of a (de- ) compression method and an IQ bit width may be excluded based on a size of a data packet comprising the IQ data. Hence, from the overall possible combinations of (decompression) methods and IQ bit widths certain combinations may be excluded. That is, one or more combinations may not make sense. For example, an IQ bit width of 1 bit (for the I and Q component, respectively) may not match the data packet size, since the size is a multiple of or derived from the bit resolution and / or the compression method. Thus, it is possible to rule out certain (impossible) combinations of (de- ) compression methods and bit resolution and obtain a remaining number or plurality of (de- ) compression methods and / or IQ bit widths, e.g., based on the data packet size.

[0066] Thus, in the following steps S13 and S14 may be performed as already described, wherein in a step S13 the IQ data is decompressed according to the remaining plurality of different (de- ) compression methods and the remaining plurality of IQ bit widths is used determine IQ samples. Turning to FIG 17, in a step S15, the fronthaul data packet analyzer may be configured based on the identified (de- ) compression method and / or IQ bit width for decoding the IQ data packets, e.g., transmitted over a fronthaul communication link. That is, the parameter values for (de- ) compressing the IQ data and / or for decoding the (uncompressed) IQ data may be set in accordance with the identified (de- ) compression method and / or the identified IQ bit width. Thus, the fronthaul data packet analyzer may be configured and / or enabled to determining U-Plane statistics and / or IQ data analysis of fronthaul traffic comprising one or more fronthaul data packets .

[0067] In a step S16, the identified (de- ) compression method and / or IQ bit width, for example the decompression method yielding the smallest constellation error, may be applied, e.g., by the fronthaul data packet analyzer, for decoding further data packets and / or processing further IQ data, e.g., received over the fronthaul communication link.

[0068] Turning to FIG 18, in a step S17 a constellation error of each one of the plurality of IQ samples is determined, e.g., as described herein. In a step S18 the constellation error may be assigned to the related decompression method and / or IQ bit width. In a step S19 the constellation errors of the plurality of different (de- ) compression methods and / or to a predetermined threshold are compared. Thus, the (de- ) compression method yielding the smallest constellation error may be determined or identified.

[0069] Turning to FIG 19, the case may appear that the IQ data is not compressed. In a step S20 a plurality of IQ samples is determined directly from the IQ data without decompressing the IQ data. In a step S21, a constellation error of the IQ samples is determined. In a step S23, the constellation errors of the IQ samples obtained by decompressing the IQ data are compared with the constellation error of the IQ samples obtained without decompressing the IQ data. Turning to FIG 20 , in a step S23 one or more data packets of a statically configured fronthaul communication link between a radio unit and a distributed unit are obtained . In such a case the data packets may not comprise a header information field, i . e . , come without a header information field ("ud- CompHdr" ) speci fying the ( de- ) compression method and / or the IQ bit width of the IQ data in the in the one or more data packets . The IQ data may represent user plane data transferred between radio unit and distributed unit , or vice versa .

[0070] The step S23 may comprise a step S24 of capturing one or more fronthaul data packets between a distributed unit and a radio unit by a fronthaul data packet analyzer, the one or more fronthaul data packets comprising the IQ data . As the case may be the fronthaul data packet analyzer may be directly connected to a radio unit and / or a distributed unit and may thus thereby obtain or capture the data packets transmitted over a fronthaul communication link .

[0071] Turning to FIG 21 , in a step S25 a configuration of a distributed unit and / or a radio unit is checked and / or identified based on the ( de- ) compression method and / or the IQ bit width identi fied . For checking a configuration relating to RU / DU the identi fied ( de- ) compression method and / or the IQ bit width may be compared with the one that the radio unit and / or the distributed unit are presumably configured with . For example , the case may appear that the config of RU or DU was done for an IQ bit width of 9 bits , but fronthaul transmission occurs with an IQ bit width of 8 bits . This may for example be due to errors in an automatic configuration tool for configuring the RU and / or DU, which have been abstracted away . Furthermore , a user may prefer the highest possible IQ bit width for quality reasons , but capacity-wise a low bit width is preferred of course . Hence , an optimi zation of the bandwidth of the fronthaul communication link may also be conducted . In a step S26, the IQ data based on the (de- ) compression method and / or the IQ bit width identified is compared with IQ data of a test pattern. Said test pattern may be stored in a memory of the fronthaul data packet analyzer or in a memory communicatively coupled to the fronthaul data packet analyzer. Thereby, the processing of the IQ data by the device under test, such as a radio unit and / or the distributed unit, may be checked or verified.

[0072] Turning to FIG 22, in a step S27 one or more of the plurality of (de- ) compression methods may comprise one or more of the following: Block Floating Point, Block Scaling, p-Law, Beamspace, Modulation, and / or Selective RE sending according to ETSI TS 103 859 V7.0.2 (2022-09) or "Q-RAN Fronthaul Control, User and Synchronization Plane Specification v07.02; (O-RAN-WG4. CUS .0-v07.02 ) " . These (de) compression methods may be ordered and / or selected in order to be executed by the fronthaul data packet analyzer.

[0073] The plurality of IQ bit widths may comprise an IQ bit width wherein the I and Q component are each between 1 and 16 bits wide, respectively. Here as well the order of utilization by the fronthaul data packet analyzer may be determined.

[0074] Turning to FIG 23, in a step S29 at least one constellation diagram with at least one constellation point is determined. In a step S30 for each IQ sample a distance to a selected one of the constellation points is determined. Here, a distance measure for determining the distance between the IQ sample and the (ideal) constellation points may be used. In a step S31 a constellation error, e.g., an error vector magnitude, EVM, based on the distance (s) may be determined. Hence, the distances of the individual IQ samples are combined and the constellation error, such as EVM, of all of the IQ samples from the IQ data may be combined to determine the constellation error. Subsequently, in a step S32 the constellation error, e.g., a constellation error value, and / or a constella- tion error diagram may be displayed. Said constellation error may also be used as described herein to identify a (de- ) compression method and / or an IQ bit width.

[0075] Turning to FIG 24, in a step S24, the steps as described herein may be repeated for detecting an updated, e.g., changed, IQ data compression and / or IQ bit width. For example, a new set of IQ data may be obtained or captured via the fronthaul communication link. The new set of IQ data may be processed in connection with the various embodiments herein.

[0076] Turning to FIG 25, an exemplary fronthaul data packet analyzer FM is shown. The fronthaul data packet analyzer FM may comprise one or more interfaces II, 12 for connecting to a fronthaul communication link, e.g., in order to be connected to a radio unit and / or a distributed unit, or may be looped in the fronthaul communication link. The one or more interfaces II, 12 are coupled to a processor P for processing the data packets received over the fronthaul (communication link) . The fronthaul data packet analyzer FM may thus comprise a processor P and / or a memory M. The memory M may store program code that is executed by the processor. The processor may thus serve for executing the steps described herein, which are implemented in software and stored in said memory M. Furthermore, the fronthaul data packet analyzer FM may comprise a display D or be coupled to a display D, e.g., for displaying the IQ samples in a constellation diagram and / or for informing a user about the identified (de- ) compression method and / or IQ bit width.

[0077] It is thus proposed to apply a plurality of or even all possible (de- ) compression methods and / or IQ data bit widths, e.g., for one or more, preferably all, bytes of one or more data packets. By measuring the error vector magnitude, EVM, it should be possible to determine, i.e., detect, one or more sections, e.g., given by a number of bits, preferably an octet of bits, i.e. a byte, of a data packet. It may not be necessary to run through all data packets of a plurality of data packets captured, e.g., since the same (de- ) compression method and / or IQ bit width are normally used through the capture of data packets. It may take a lot of effort to process one data packet since there may be a large amount of different (de- ) compression methods and / or IQ bit width combinations. For example, the processing needs to be started anew for every section, e.g., byte, of a data packet. The one or more data packets may comprise one or more sections, preferably given by a number of bits, e.g., bytes, and / or may be structured according to an IQ data frame format, as e.g. shown in Table 8-2 of ETSI TS 103859 V7.0.2 (2022-09) , wherein a section corresponds to one or more "Octets". Thus, the length of a data packet may affect the performance. A data packet may contain several sections of different (de-

[0078] ) compression methods and / or IQ bit width. Thus, the proposed compression detection must determine and / or report all the sections with used (de- ) compression method and / or IQ bit width. The process may be optimized so that a section of the data packets that are already discovered, i.e. for which the proper (de- ) compression method and / IQ bit width has already been determined, are not processed any more in further data packets, e.g., of the same captured and / or for or from the same antenna carrier, AxC .

[0079] Furthermore, there needs to be proper IQ data in the one or more sections of a data packet in order to allow compression detection, e.g., based on the EVM as described herein. It may thus be necessary to perform compression detection for every section, and even byte, of a data packet if the start of a section within a data packet is not available. One data packet may contain several sections and in addition the number of sections in a data packet or the length of sections in the data packet may not be known or available.

[0080] The following (de- ) compression methods may need to be tested with an IQ bit width from 1 bit to 16 bits:

[0081] • No compressoin

[0082] • Block floating point • Block scaling

[0083] • p-law

[0084] • Modulation compression

[0085] • BFP+selective RE sending

[0086] • Mod-compr + selective RE sending

[0087] Furthermore, the modulations may need to be tested for every combination of (de- ) compression method and IQ bit width. The following modulations may need to be tested:

[0088] • BPSK

[0089] • QPSK

[0090] • 16-QAM

[0091] • 64-QAM

[0092] • 256-QAM

[0093] Hence, given 7 different compression methods and 16 different resolutions, 112 variants of the IQ samples may be obtained from the IQ data. These 112 variants may then be tried with the different modulations, e.g. BPSK, QPSK, 16-QAM, 64-QAM, 256-QAM.

[0094] Still further, in case modulation compression is used, there can be shifting in constellation points, e.g., QPSK may be shifted by -1 / 2, wherein as a result I value is represented by -1,0 or 0,0 instead of -1 / 2, 1 / 2 or 5, 5 . Hence, all shifting values should be tried when modulation compression is used. Another possibility would be to investigate offset between values to detect modulation and shifting.

[0095] A constellation error, such as an EVM, may be calculated and / or a threshold value may be defined, in order to define and / or determine sections in a data packet. For example, if 1 or 2 bytes for which no (de- ) compression method and / or IQ bit width could be determined, these 1 or 2 bytes may be identified as and / or correspond to the udCompParam field, e.g., between or at the beginning of the IQ data in the (adjacent) sections, cf. again Table 8-2 of ETSI TS 103859 V7.0.2 (2022— 09) . An IQ data detection window may be defined. The IQ data detection window may comprise one or more, e.g., a plurality of consecutive bytes of the IQ data in a data packet. The data in the (IQ data) detection window may be processed as described herein, e.g., the IQ data therein may be processed according to the one or more (de- ) compression methods and / or the IQ bit width. A shift of the IQ detection window may be performed. The shift may comprise one or more, or a plurality of, bytes in a data packet. The shift of the detection window may be performed in order to avoid using one or more data fields, for example bytes, which comprise data fields, such as, e.g., one or more of the following sectionlD, start Prbu, numPrbu, upCompHdr, udCompLen, and / or udCompparam. The shift of the detection window may be performed in order to avoid using part of the data packet for the (de- ) compression and / or IQ bit width calculation. Thus, when / if one or more attempts of determining the (de- ) compression method and / or IQ bit width, e.g., of data in a detection window, fails, those one or more bytes (in the detection window) may be skipped. The processing of data may then be again applied for a new set of bytes in the detection window, e.g., one or more successive bytes in the data packet.

Claims

PATENT CLAIMS1. A, preferably computer-implemented, method of processing in-phase and quadrature data, IQ data, the method comprising the steps of: determining a plurality of IQ samples from the IQ data, determining a constellation error of each one of the plurality of IQ samples, and identifying a (de- ) compression method and / or IQ bit width of the IQ data based on the constellation error.

2. The method according to the preceding claim, determining a plurality of IQ samples from the IQ data by decompressing the IQ data according to a plurality of different (de- ) compression methods, and / or determining the plurality of IQ samples from the IQ data by using a plurality of different IQ bit widths for the (uncompressed) IQ data.

3. The method according to any one of the preceding claims, firstly decompressing the IQ data according one of the plurality of decompression methods, and secondly using one or more of the plurality of different IQ bit widths for determining a plurality of IQ samples from the (uncompressed) IQ data .

4. The method according to any one of the preceding claims, decompressing the IQ data according to a subset or a prioritized order of the plurality of different (de- ) compression methods, and / or using the IQ bit widths in a prioritized order and / or a subset of the plurality of different IQ bit widths for the (uncompressed) IQ data.

5. The method according to any one of the preceding claims, firstly setting the IQ bit width to one of the plurality of IQ bit widths,decompressing the IQ data according to the plurality of di fferent ( de- ) compression methods .6 . The method according to any one of the preceding claims , excluding one or more combinations of a ( de- ) compression method and an IQ bit width based on a si ze of a data packet comprising the IQ data .7 . The method according to any one of the preceding claims , configuring a fronthaul data packet analyzer based on the identi fied ( de- ) compression method and / or IQ bit width for decoding the IQ data packets transmitted over a fronthaul communication link, in particular for determining U-Plane statistics and / or IQ data analysis of fronthaul traf fic comprising one or more fronthaul data packets .8 . The method according to any one of the preceding claims , applying the identi fied ( de- ) compression method and / or IQ bit width, for example the decompression method yielding the smallest constellation error, for decoding further data packets and / or processing further IQ data .9 . The method according to any one of the preceding claims , assigning the constellation error to the related decompression method and / or IQ bit width, comparing the constellation errors of the plurality of di fferent ( de- ) compression methods and / or to a predetermined threshold .10 . The method according to any one of the preceding claims , determining a plurality of IQ samples directly from the IQ data without decompressing the IQ data, and determining a constellation error of the IQ samples , and comparing the constellation errors of the IQ samples obtained by decompressing the IQ data with the constellation error of the IQ samples obtained without decompressing the IQ data .

11. The method according to any one of the preceding claims, obtaining one or more data packets of a statically configured fronthaul communication link between a radio unit and a distributed unit, i.e. without a header information field ("ud- CompHdr") specifying the (de- ) compression method and / or the IQ bit width of the IQ data in the in the one or more data packets, wherein preferably the IQ data represent user plane data transferred between radio unit and distributed unit, or vice versa.

12. The method according to any one of the preceding claims, capturing one or more fronthaul data packets between a distributed unit and a radio unit by a fronthaul data packet analyzer, the one or more fronthaul data packets comprising the IQ data.

13. The method according to any one of the preceding claims, identifying and / or checking a configuration of a distributed unit and / or a radio unit based on the (de- ) compression method and / or the IQ bit width identified, and / or comparing the IQ data based on the (de- ) compression method and / or the IQ bit width identified with IQ data of a test pattern .

14. The method according to any one of the preceding claims, wherein the plurality of (de- ) compression methods comprise one or more of the following: Block Floating Point, Block Scaling, p-Law, Beamspace, Modulation, and / or Selective RE sending according to ETSI TS 103 859 V7.0.2 (2022-09) or "0- RAN Fronthaul Control, User and Synchronization Plane Specification v07.02; (Q-RAN-WG4.CUS.0-v07.02) ".

15. The method according to any one of the preceding claims, wherein the plurality of IQ bit widths comprising an IQ bit widths wherein the I and Q component are each between 1 and 16 bits wide.

16. The method according to any one of the preceding claims, determining at least one constellation diagram with at least one constellation point, and determining for each IQ sample a distance to a selected one of the constellation points, and determining a constellation error, e.g., an error vector magnitude, EVM, based on the distance (s) , and preferably displaying the constellation error, e.g., a constellation error value, and / or a constellation error diagram.

17. The method according to any one of the preceding claims, repeating the steps of any one of the preceding claims for detecting an updated, e.g., changed, IQ data compression and / or IQ bit width.

18. An apparatus operative to perform the method steps of any one of the preceding claims.

19. An apparatus comprising a processor and a memory, wherein the memory comprises program code that when executed by the processor performs the methods steps of any one of the preceding claims 1-17.

20. A computer program, preferably stored on a non-transitory medium, comprising program code that when executed performs the method steps of any one of the preceding claims 1-17.

21. A system comprising a radio unit and / or a distributed unit and a computer program, e.g., on an apparatus according to claims 18 and / or 19, for performing the method steps according to any one of the preceding claims 1-17.

Images