Method of compensating for clock skew between devices during UWB-MMS ranging

Abstract

A low complexity clock drift estimation scheme for UWB-MMS ranging is presented that assumes that a sampling frequency offset (SFO) and a carrier frequency offset (CFO) are driven from the same clock source. The SFO is first estimated with a plurality of CIR segments, and the SFO is used to compensate for the CFO of the CIR segments. A fine CFO estimate is then obtained from the compensated CIR. Combining the coarse SFO and fine CFO estimates to resample and phase rotate the original CIR can significantly improve the performance of the CIR combination, thereby improving the performance of ranging.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a method of compensating for clock offsets between devices in a UWB MMS ranging procedure. The present disclosure additionally relates to a UWB device configured to perform the proposed method. The present disclosure further relates to a computer implemented method. BACKGROUND

[0002] To overcome regulatory power limitations and extend the coverage range of Ultra-Wideband (UWB), Multi-Millisecond Ranging was proposed in the recent IEEE 802.15.4ab standard. However, due to the fact that ranging segments are separated by milliseconds, the Channel Impulse Response (CIR) combination for increasing the Signal-to-Noise Ratio (SNR) in the Multi-Millisecond scheme faces the challenge of impairment from clock offsets between two devices. On one hand, a Sampling Frequency Offset (SFO) will cause the CIR of the ranging segments separated by 1 ms to be smeared. On the other hand, a Carrier Frequency Offset (CFO) will cause the phase of the segments to be wrapped, which needs to be compensated for to integrate the inter-segment CIR coherently.

[0003] US 2022 / 0137177 Al discloses techniques for extending operating ranging distances with a hybrid of Ultra-Wideband (UWB) and Narrowband (NB) signaling.

[0004] US 8,085,876 B2 discloses apparatuses and methods for sampling frequency offset and carrier frequency offset estimation and correction in a communication system. SUMMARY

[0005] A first aspect of the present disclosure relates to a method of compensating for clock offsets between devices in a UWB MMS ranging procedure, comprising the steps of:

[0006] - determining a CIR segment as a result of a ranging segment of the MMS ranging procedure;

[0007] - applying a specified interpolation to the CIR segment;

[0008] - determining a sampling frequency offset SFO between a first UWB device and a second UWB device from the interpolated CIR segment;

[0009] - compensating for a carrier frequency offset CFO by means of the determined sampling frequency offset SFO;

[0010] - determining a residual estimate of the carrier frequency offset CFO;

[0011] - determining a clock offset by means of the estimated sampling frequency offset SFO and the carrier frequency offset CFO; and

[0012] - compensating for the clock offset to obtain a fine CIR segment; and

[0013] combining the fine CIR segments to a channel impulse response, CIR.

[0014] In this way, a low complexity clock drift compensation scheme for UWB-MMS ranging procedures is provided. It is assumed that the sampling frequency offset (SFO) and the carrier frequency offset (CFO) are driven from the same clock source. The SFO is first estimated with the multiple CIR segments and the SFO is used to compensate the CFO of the CIR segments. Then, a fine CFO estimate is obtained from the compensated CIR. Combining the coarse SFO and the fine CFO estimate to resample and phase rotate the original CIR segments can significantly improve the performance of the CIR combination. The proposed method is suitable for MMS ranging procedures, where clock differences of the devices involved in these MMS ranging procedures can be compensated. Thus, an advanced ranging performance can be provided.

[0015] A second aspect of the present disclosure relates to a UWB device configured to perform the proposed method.

[0016] A third aspect of the present disclosure relates to a computer implemented method comprising executable instructions which, when executed by a UWB device, cause the UWB device to carry out the proposed method.

[0017] In one or more embodiments, in step b), the sampling frequency offset SFO is estimated using a linear relationship between the CIR peak positions.

[0018] In one or more embodiments, the sampling frequency offset SFO is related to the peak of the channel impulse response, CIR.

[0019] In one or more embodiments, the sampling frequency offset is determined from an up-sampled channel impulse response, CIR. For example, this can be done by up-sampling from 1 GHz to 32 GHz.

[0020] In one or more embodiments, the estimated sampling frequency offset SFO is obtained by the following equation:

[0021] f {SFOest,ppm} = slope(CIR peak positions) / L

[0022] where L is the interpolation factor (e.g. L = 32 when up-sampling from 1 GHz to 32 GHz).

[0023] In one or more embodiments, the slope of the channel impulse response, CIR peak positions, is determined by a linear regression or a differentiation method.

[0024] In one or more embodiments, in case one peak position of a CIR segment is designated far away from the peak position of another CIR segment, it is removed from the slope calculation. In this optional step, an outlier remover is used to increase the accuracy.

[0025] In one or more embodiments, in step e), the compensated CIR peaks are utilized to estimate the residual carrier frequency offset CFO, wherein a phase unwrapping and differentiation method is used. Phase unwrapping is necessary because phase wrapping can occur across N>2 segments.

[0026] In one or more embodiments, in step f), the total clock offset is determined as the sum of the estimated sampling frequency offset SFO and the estimated carrier frequency offset CFO. BRIEF DESCRIPTION OF DRAWINGS

[0027] The above discussion / summary does not intend to describe each and every embodiment or implementation of the present disclosure. The figures and detailed description that follow also exemplify various embodiments. The aspects defined above and further aspects of the present disclosure will become apparent from the examples of embodiments described hereinafter, reference being made to the accompanying drawings.

[0028] All illustrations of the drawings are schematic. Like or similar elements or features are referred to with the same reference signs or with reference signs differing only in a first number which is identical among the different figures. In order to avoid unnecessary repetition, elements or features already explained with respect to a previously described embodiment are no longer explained with respect to later described embodiments.

[0029] Various example embodiments can be more fully understood from the following detailed description, taken in connection with the accompanying drawings, in which:

[0030] Figure 1 shows non-MMS ranging frames and MMS ranging frames in a UWB ranging procedure;

[0031] Figure 2 shows a MMS ranging frame illustration;

[0032] Figure 3 is an illustration showing CIR tailing on CIR peak positions due to clock drift between two UWB devices;

[0033] Figure 4 is an illustration showing combined CIR segments without compensation;

[0034] Figure 5 is a block diagram of a conventional multi-branch search algorithm for compensating clock drift between two UWB devices;

[0035] Figure 6is a flowchart of the proposed method;

[0036] Figure 7 is a graphical illustration of the phase of the CIR peak before stage 1 compensation;

[0037] Figure 8 is a graphical illustration of the phase of the CIR peak after stage 1 compensation;

[0038] Figure 9 is a block diagram of the proposed two-stage clock drift compensation process;

[0039] Figure 10 is a graphical illustration of individual CIR segments after compensation; and

[0040] Figure 11 is a graphical illustration of combined CIR segments after compensation.

[0041] While various embodiments discussed herein allow for modifications and alternative forms, there are shown in the drawings and will hereinafter be described in detail a number of aspects of the application. However, it should be understood that the application is not limited to the particular embodiments described, but is intended to cover all modifications, equivalents, and alternatives falling within the scope of the application including aspects defined in the claims. Additionally, the term "example" as used throughout this application is only used to illustrate, not to limit. DETAILED DESCRIPTION

[0042] Aspects of the present disclosure are believed to be applicable to a variety of different types of apparatuses, devices, systems and methods. Various aspects can be appreciated, for example, via the following discussion in connection with the illustrative contexts of a non-limiting example.

[0043] In the following description, various specific details are set forth to describe the specific examples presented herein. However, it should be understood that one or more other examples and / or variations of the examples presented can be practiced without using all the specific details given below. In other instances, well-known features have not been described in detail so as not to obscure the description of the examples presented herein. For ease of illustration, the same reference numbers can be used in different illustrations and / or drawings to indicate the same or other like elements. Also, while aspects and features can be described or depicted in certain instances throughout the drawings, it is to be understood that such depictions are not intended to be exhaustive or limiting of the various aspects and features presented herein and that one or more aspects and features other than those described and / or depicted can be employed in accordance with aspects of the present disclosure.

[0044] The regulatory limit for the average power of a UWB device is -41.3 dBm / MHz, measured over 1 ms. Multi-millisecond ranging (MMS ranging) overcomes this limit by spreading the transmission over multiple milliseconds, as primarily described in Figure 1As shown in the image.

[0045] exist Figure 1 A single (non-MMS) ranging frame RP is seen in the upper part, and... Figure 1 The lower part shows several MMS ranging segments RF1…RF4. These MMS ranging segments RF1…RF4 are much shorter than the ranging frame RP, and therefore can be transmitted at a higher transmission power than the non-MMS ranging frame RP. Thus, the central idea behind MMS ranging is to break a single ranging frame RP into multiple smaller and shorter ranging segments, which can be transmitted at higher transmission power, improving ranging performance. Each of the MMS ranging segments can use an average power of -41.3 dBm / MHz.

[0046] This disclosure uses timing frequency estimation to obtain improved clock offset. It assumes that the carrier and symbol clocks come from the same reference, making the following equations applicable:

[0047] SFO ppm =CFO ppm (1)

[0048] SFO ppm Sampling frequency offset between the two UWB devices (in ppm)

[0049] CFO ppm Carrier frequency offset between two UWB devices (in ppm)

[0050] This is commonly used in practical system implementations. Compared to conventional multi-branch brute-force search, the proposed timing frequency estimation algorithm has very low complexity and is quite robust under very low SNR conditions. The basic procedure of the proposed algorithm is as follows:

[0051] 1) Interpolate the CIR for each CIR segment, then estimate the SFO from the oversampled CIR peak. An outlier remover can optionally be used to reduce estimation inaccuracies and increase the robustness of the algorithm.

[0052] 2) Since CFO = SFO, compensate the CFO of the CIR segment based on the estimated SFO value.

[0053] 3) Estimate residual CFO from the peak value of compensated CIR segments

[0054] 4) The total clock offset difference is the sum of the coarse SFO estimate from step 1 and the fine CFO estimate from step 3.

[0055] 5) Combine CIR segments with timing (resampling) and frequency compensation (phase rotation).

[0056] Using a fine duty cycle approach, the transmission power of each MMS ranging fragment RF1...RFn can be boosted, increasing the SNR and extending the ranging distance between the UWB devices 10, 20. In Figure 2 In the specific ranging session context, the MMS ranging fragments RF1...RF4 are split between the first UWB device 10 acting as the initiator and the second UWB device 20 acting as the responder in the two-way ranging process in order to determine the distance between the two UWB devices 10, 20. We see that the first UWB device 10 transmits the MMS ranging fragments RF1, RF2 and the second UWB device transmits the MMS ranging fragments RF3, RF4.

[0057] Figure 2 A diagram of the multi-millisecond method with MMS ranging fragments RF1...RFn is shown in Fig. 1. The ranging fragments RF1...RFn are sent by each UWB device 10, 20 at 1 ms intervals so that power transmission can be increased, as mainly in Figure 1 The MMS ranging fragments RF1...RFn can be implemented as e.g. ranging sequence fragments (RSF) and / or ranging integrity fragments (RIF). The second UWB device 20 acting as the responder performs a cross-correlation with a known sequence to obtain one channel impulse response (CIR) fragment, which is shown in Fig. 2 as a single curve. The context with one-way ranging is not shown in the diagram. Figure 3

[0058] The MMS ranging frame RSF contains a synchronization symbol, while the ranging frame RIF contains a scrambling timestamp sequence (STS) that increases the integrity of the ranging measurement.

[0059] However, since the ranging fragments RF1...RFn are separated by 1 ms, even a clock offset between the two communicating UWB 10, 20 devices can pose a challenge for the CIR combination. Specifically, two impairments can be observed:

[0060] 1. CIR tailing, where the peak of the CIR from a ranging fragment is shifted to one side of the peak of the CIR from the previous ranging fragment with an unknown offset

[0061] 2. Phase wrapping, where the phase at each CIR peak from each CIR fragment will be a wrapped version of the actual phase

[0062] For example, assume a clock offset of 0.5 ppm between the first UWB device 10 acting as the initiator and the second UWB device 20 acting as the responder. Figure 3 ​The diagram illustrates the result of a cross-correlation performed by the second UWB device 20 with eight ranging segments RF1…RF8, where one CIR is generated by one ranging segment RF1…RF8, and the second UWB device 20, acting as the responder, acquires eight ranging segments. In this case, in Figure 3 As can be seen, the peak value of CIR segment #2 is shifted relative to the peak value of CIR segment #1. Similarly, the peak value of CIR segment #3 is shifted relative to the peak value of CIR segment #2, and so on. Therefore, the original combination... Figure 3 The CIR will cause such Figure 4 The distance measurement shown is inaccurate, and there is no clearly defined peak value, which can lead to poor distance measurement efficiency.

[0063] Furthermore, the phase shift between CIRs caused by the CFO will also degrade the performance of the CIR combination. Due to the 1 ms MMS ranging segment separation, a very small CFO is crucial to avoid phase entanglement between CIR segments. Phase entanglement is a result of a modulus function between zero degrees and 2π. For example, the CFO should be:

[0064]

[0065] Equation (2) shows that a CFO of ≤500Hz would be tolerable. At a carrier frequency of 8GHz (which is the carrier frequency at UWB channel 9 as defined in IEEE 802.15.4), this corresponds to a CFO of less than 0.0625ppm. However, this is difficult to achieve in practical system implementations, thus requiring post-processing to estimate and compensate for the clock skew.

[0066] The conventional approach for estimating clock skew would be to use a multi-branch brute-force search to find the optimal SFO / CFO values ​​to achieve the CIR combination, as primarily in... Figure 5 The block diagram is shown. It can be seen that the estimation accuracy depends on the number of branches B1…Bn. Increasing the number of branches B1…Bn can significantly increase the complexity of the entire system, which is undesirable in system implementation. In summary, Figure 5 The conventional block diagram implements a rather complex algorithm with multiple branches B1…Bn, where the accuracy of the conventional algorithm implemented in this way depends on the number of branches B1…Bn. Different SFO and CFO are tried to see which gives the best result regarding compensation. If more knowledge is desired, more branches B1…Bn need to be created. Each branch B1…Bn obtains specific values ​​for SFO and CFO, and it is checked which branch B1…Bn gives the best result. In this way, the conventional method implements a maximum likelihood procedure.

[0067] In contrast, the present disclosure proposes a low-complexity two-stage timing frequency estimation method to estimate the clock offset, which can significantly reduce implementation complexity while maintaining good performance. A flowchart of the proposed method is shown in Figure 6

[0068] At a first step 100, a windowed channel impulse response (CIR) is obtained from each ranging fragment RFl... RFn. In step 101, the CIR is interpolated and as a result of each ranging fragment, the peak value of each CIR fragment is obtained. It is desired to estimate the SFO from the up-sampled CIR (e.g. from 1 GHz to 32 GHz). Since ideally, the CIR peak value of each fragment is shifted according to the SFO value, the SFO can be estimated from the peak position of the CIR. In step 103, the SFO estimate can thus be obtained by the following equation:

[0069] f {SFOest,ppm} = slope(CIR peak position) / L (3)

[0070] where L is the interpolation factor. The slope of the CIR peak position can be computed by linear regression or by a differentiation method. Alternatively, in optional step 102, an outlier remover is used to increase the precision of the slope computation. In this context, intuitively, if one CIR peak position is far away from the other CIR peak positions, it can be removed from the slope computation. This exploits the a priori information that the clock offset is constant during the MMS transmission.

[0071] Next, in step 104, since CFO = SFO is assumed, the CIR of each fragment is phase compensated by means of the estimated SFO value, so that the residual CFO range is reduced and the phase difference between two fragments RFl... RFn caused by the residual CFO will be smaller than π. The phase of the CIR peak before and after CFO compensation in step 104 is shown in Figure 7 and 8

[0072] The steps 100 to 104 explained above represent the first stage ST1 of the proposed method. The remaining steps 105 to 107 represent the second stage ST2 of the proposed method.

[0073] The second stage ST2 starts in step 105, in which the compensated CIR peak can be utilized to estimate the fine CFO. In this context, for example, the phase unwrapping and differentiation method can be used. Phase unwrapping is necessary because phase wrapping can occur across N > 2 fragments. With phase unwrapping, the phase differentiation can be utilized to estimate the fine CFO value in the following way:

[0074] ​​

[0075] D delay interval between the nth segment and the n+D segment

[0076] ph n unfolded phase of the nth CIR peak after compensation

[0077] N number of segments

[0078] f c carrier frequency

[0079] where the delay D is configurable and can be optimized according to a trade-off between SNR and estimation accuracy. The delay D is determined by residual phase wrapping.

[0080] Then, in step 106, the total clock offset (CO) estimate is determined as the sum of the coarse estimate and the fine CFO estimate:

[0081] f {co,ppm} = f {SFOest,ppm} + f {CFOest,ppm} (5)

[0082] By this, the CIR segments can be resampled and phase-synchronized. Finally, in step 107, the fine CIR segments can be coherently combined in order to start the ranging process on the UWB devices 10, 20 with the clock offset eliminated. In this context, it should be noted that the proposed method also works with a one-way ranging process between the UWB devices 10, 20.

[0083] Figure 7 The phase of the CIR peak before CFO compensation in the first stage ST1 is shown. We see that the curve of the phase at each CIR segment index is not straight due to clock drift. The apparent phase wrapping can be seen as the effect of splitting the ranging segments RFl... Rfn at 1 ms intervals.

[0084] Figure 8 The essentially linear curve of the phase of the CIR peak after CFO compensation in the first stage ST1 is shown. The linear curve is the result of residual CFO accumulation. We see that the phase wrapping has essentially been attenuated due to compensation of the clock drift between the UWB devices 10, 20.

[0085] Figure 9 is a block diagram of the two-step timing frequency estimation algorithm of the flow Figure 6 We see that the first stage ST1 and the second stage ST2 mentioned above have been explained in more detail above with respect to the flowchart of Figure 6 The terms "resampling" and "phase rotation" of the second stage ST2 can be seen as compensation processes.

[0086] The resulting CIR is shown in Figure 10 and the resulting combined CIR is shown in Figure 11 In contrast to Figure 4 one can see a well-defined single peak between indices 3.28 and 3.29. The indices are related to the order of the output as a result of the cross-correlation.

[0087] The proposed method has been exemplarily illustrated as a two-way MMS ranging procedure between UWB devices. It goes without saying that the proposed method can also be implemented in the context of a one-way ranging or other MMS ranging procedures between UWB devices. Moreover, all numerical values given above are to be understood as exemplary.

[0088] As another example, where a description can refer to a "first" type of structure, a "second" type of structure, etc., where the adjectives "first" and "second" are not used to connote any substantive meaning or provide any substantive significance to the structures; rather, such adjectives are merely used in the English language antecedents to distinguish one such similarly named structure from another similarly named structure.

[0089] Based on the above discussion and illustration, those skilled in the art will readily appreciate that changes and modifications in the various embodiments can be made without departing from the example embodiments and applications described herein. For example, the methods illustrated in the figures can involve steps performed in different orders, where maintaining one or more aspects of the embodiments herein, or can involve fewer or more steps.

[0090] It should be noted that the term "comprising" does not exclude other elements or steps and "a" or "an" does not exclude a plurality. Also, elements described in association with different embodiments can be combined. It should also be noted that reference signs in the claims should not be construed as limiting the scope of the claims.

[0091] The systems and methods described herein can be embodied, at least in part, by one computer program or multiple computer programs, which can exist in a single computer system or across multiple computer systems in active and inactive states, in a variety of forms. For example, they can exist as software programs of instructions, embodied either fixed or transitory in source code, object code or executable code, or other formats, for execution by a processor. Any of the above can be embodied on a computer readable medium, which can include storage devices and signals, in compressed or uncompressed form. The computer readable medium can be a machine-readable storage device, a machine-readable storage substrate, a memory device, or a combination of one or more of them.

[0092] As used herein, the term "computer" refers to any electronic device comprising a processor, such as a general central processing unit (CPU), a specialized processor, or a microcontroller. A computer is capable of receiving data (input), of performing a sequence of predetermined operations on the data, and of producing a result in the form of information or signals (output). Depending on the context, the term "computer" will mean, in particular, a processor or, more generally, a processor associated with a combination of related elements contained within a single chassis or housing.

[0093] The term "processor" or "processing unit" refers to a data processing circuitry, which can be a microprocessor, a co-processor, a microcontroller, a microcomputer, a central processing unit, a field-programmable gate array (FPGA), a programmable logic circuit, or any circuitry manipulating signals (analog or digital) based on operational instructions stored in a memory. The term "memory" refers to a storage circuitry or storage circuitries, such as a read-only memory, a random access memory, a volatile memory, a non-volatile memory, a static memory, a dynamic memory, a flash memory, a cache memory, or any circuitry storing digital information.

[0094] As used herein, a "computer-readable medium" or "storage medium" can be any means that can contain, store, communicate, propagate or transport the program for use by or in connection with the instruction execution system, apparatus or device. The computer-readable medium can be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device or propagation medium. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CDROM), an digital versatile disc (DVD), a Blu-ray disc (BD), and a memory stick.

[0095] It is noted that the above-described embodiments have been described with reference to different subjects. In particular, some embodiments can have been described with reference to method claims whereas other embodiments can have been described with reference to device claims. However, a person of skill in the art will gather from the above description that, unless otherwise indicated, any aspect or feature described as a part of a certain subject-matter also is, more generally, a part of a related subject-matter. For example, any combination of features along a method claim, and any combination of features along a device claim are also, more generally, considered to be within the scope of the present application.

[0096] Reference Signs:

[0097] 10 first UWB device

[0098] 20 2nd UWB device

[0099] 100..107 method steps

[0100] B1...Bn branch

[0101] CO clock offset

[0102] CFO carrier frequency offset

[0103] SFO sampling frequency offset

[0104] RF1...RFn MMS ranging fragment

[0105] RP non-MMS ranging frame

[0106] ST1, ST2 stage.

Claims

1. A method for compensating for clock offset CO between UWB devices (10, 20) during MMS ranging, characterized in that, Includes the following steps: a) Determine the CIR segment as the result of the ranging segment (RF1……RFn) in the MMS ranging process; b) Apply the specified interpolation to the CIR segment; c) Determine the sampling frequency offset SFO between the first UWB device (10) and the second UWB device (20) from the interpolated CIR segment; d) Compensate for carrier frequency offset CFO by means of the determined sampling frequency offset SFO; e) Determine the residual estimate of the carrier frequency offset CFO; f) The clock offset CO is determined by means of the estimated sampling frequency offset SFO and the carrier frequency offset CFO; g) Compensate for the clock offset to obtain a fine CIR segment; as well as h) Combine the CIR segments into the channel impulse response CIR.

2. The method according to claim 1, characterized in that, In step b), the linear relationship between CIR peak locations is used to estimate the sampling frequency offset SFO.

3. The method according to claim 1 or 2, characterized in that, The sampling frequency offset SFO is related to the peak value of the channel impulse response CIR.

4. The method according to any one of the preceding claims, characterized in that, The sampling frequency offset (SFO) is determined based on the impulse response (CIR) of the sampling channel.

5. The method according to claim 4, characterized in that, The estimated sampling frequency offset (SFO) is obtained through the following equation: f {SFOest,ppm} = Slope (CIR peak position) / L Where L is the interpolation factor.

6. The method according to claim 5, characterized in that, The slope of the CIR peak position of the channel impulse response is determined by linear regression or differential methods.

7. The method according to claim 5 or 6, characterized in that, If a peak location of a CIR segment is specified as being far from the peak location of another CIR segment, it is removed from the slope calculation.

8. The method according to any one of the preceding claims, characterized in that, In step e), the residual carrier frequency offset CFO is estimated using the compensated CIR peak, where phase expansion and differentiation methods are used.

9. A UWB device (10, 20), characterized in that, It is configured to perform the method according to any of the preceding claims.

10. A computer-implemented method, characterized in that, It includes executable instructions that, when executed by the UWB device (10, 20), cause the UWB device (10, 20) to perform the method according to claims 1 to 8.

Images