Avoiding gaussian filter in GFSK devices with upsampled correlative coding of baseband signal
Upsampling techniques like duobinary and polybinary coding address the demodulation mismatch in RFID tags by generating compatible GFSK signals, improving BER by 5-8 dB and reducing interference, thereby enhancing BLE-backscatter system performance.
Patent Information
- Application Number
- PCT/IB2025/057652
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2024-07-30
- Filing Date
- 2025-07-28
- Publication Date
- 2026-02-05
AI Technical Summary
Existing RFID tags using FSK transmission face demodulation mismatches with GFSK receivers, leading to suboptimal performance and increased interference, particularly in BLE-backscatter systems, due to the mismatch between FSK signals and GFSK receivers, which are optimized for Gaussian-filtered signals.
Implementing upsampling techniques such as duobinary and polybinary coding to generate waveforms with two-step or three-step transitions, eliminating the need for complex Gaussian filtering, and ensuring compatibility with GFSK receivers by producing a frequency-modulated signal that compacts spectrum and improves Bit Error Rate (BER) performance.
The proposed upsampling methods improve BER by approximately 5-8 dB compared to classical FSK, reducing interference and power consumption, while maintaining compatibility with GFSK receivers, thus enhancing communication efficiency in BLE-backscatter systems.
Smart Images

Figure IB2025057652_05022026_PF_FP_ABST
Abstract
Description
Attorney Docket: SDSUP002WOAVOIDING GAUSSIAN FILTER IN GFSK DEVICES WITH UPSAMPLED CORRELATIVE CODING OF BASEBAND SIGNALRELATED APPLICATION
[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 677,242, filed on July 30, 2024, and entitled “FSK AND GFSK DEVICES WITH UPSAMPLED CORRELATIVE CODING OF BASEBAND SIGNAL” which is hereby incorporated by reference in its entirety.BACKGROUND
[0002] Introduction
[0003] There is a need to significantly improve the performance of communication links between Radio Frequency Identification (RFID) tags and Bluetooth devices. An RFID tag is a small integrated circuit device attached to or implanted in an object. These tags contain information that can be read via radio waves at short range. Such tags are increasingly being researched for their ability to provide the benefits of RFID without the need for expensive stand-alone readers. Strictly speaking, these tags do not generate Bluetooth standard-specified waveforms but are nevertheless compatible with Bluetooth standard-compliant receivers. Gaussian Frequency Shift Keying (GFSK) transmission generally is replaced in these tags by simpler FSK transmission from the tag. Unfortunately, there often is a mismatch when FSK signals are demodulated by GFSK receivers.
[0004] Backscatter is a nonstandard way of transmitting information wherein data is communicated by just modulating the reflection coefficient of the antenna load. This way, complex and power-hungry transmitter components such as radio frequency (RF) oscillators, mixers and filters can be avoided. Furthermore, integrating RF power harvesting circuitry in such devices can allow them to operate in a battery- free mode. The idea of transmitting information by reflecting an incident RF signal was first proposed in [1] H. Stockman, “Communication by means of reflected power,” in Proc. IRE, Oct 1948, pp. 1196-1204. Radio Frequency Identification (RFID - See [2] EPC Global US, Class 1 Generation 2 UHF RFID protocol for operation at 860MHz - 960MHz. http: / / www.epcglobalus.org.) is the first major application of this technology. More recently, other applications have come up in the domains of sensors for remote monitoring applications. See, [3] N. Cho, S.-J. Song, S. Kim, and H.-J. Yoo, “A 5.1 microwatt UHF RFID tag chip integrated with sensors for wireless environmental monitoring,” in Proc. 31st Eur. Solid State Circuits Conf., 2005, pp. 279-282;Attorney Docket: SDSUP002WO[4] R. Muller, “A minimally invasive 64-channel wireless micro ECoG implant,” IEEE J. Solid-State Circuits, pp. 344-359, Jan. 2015; and [5] D. Piech, J. Kay, B. Boser, and M. Maharbiz, “Rodent wearable ultrasound system for wireless neural recording,” in Proc. IEEE Engineering in Medicine and Biology Society, 2017, pp. 1-5.
[0005] Backscatter tags have received much importance in academic research over the last decade. While tags were initially used primarily for identification, they are now being used for a variety of sensor communication applications as well. [3-5], A shortcoming of early RFID technology is that dedicated readers were necessary to 30 receive the data transmitted from tags. Over the last few years, researchers have developed tag hardware that can transmit signals decodable by widely available commodity receivers such as WiFi, [6] O.-H. P. Wang, C. Zhang, H. Yang, M. Dunna, D. Bharadia, and P. P. Mercier, “A low-power backscatter modulation system communicating across tens of meters with standards-compliant Wi-Fi transceivers,” IEEE J. Solid-State Circuits, pp. 2959-2969, Nov 2020 or Bluetooth Low Energy (BLE) [7] P. Zhang, C. Josephson, D. Bharadia, and S. Katti, “FreeRider: Backscatter communication using commodity radios,” in Proc. ACM CoNEXT ’ 17, 2017, pp. 389-401;[8] L. Yuan, Q. Wang, J. Zhao, and W. Gong, “Multiprotocol backscatter with commodity radios for personal loT sensors,” IEEE / ACM Trans. Networking, pp. 1132-1144, June 2023;[9] S. Chen, M. Zhang, J. Zhao, W. Gong, and J. Liu, “Reliable and practical Bluetooth backscatter with commodity devices,” IEEE / ACM Trans. Networking, pp. 1717-1729, Aug. 2021;
[0010] M. Katanbaf, A. Safari, and J. R. Smith, “Receiver selectivity limits on bistatic backscatter range,” in Proc. IEEE Internation Conf. On RFID, 2020, pp. 1-8;
[0011] S.-K. Who, M. Donna, D. Bharadia, and P. P. Mercier, “A WiFi and Bluetooth Low Energy backscatter combo chip with beam steering capabilities,” IEEE Open Journal of the Solid-State Circuits Society, pp. 239-248, Sep. 2023. In BLE-B ackscatter (BLE-B), the tag transmits on one of three specifically designated advertisement channels. Any BLE-compliant receiver periodically monitors the advertisement channels and thereby identifies the tag. BLE connectivity has another advantage. The tag can readily connect to the Cloud infrastructure through BLE- equipped cell phones that are ubiquitous today.
[0006] Researchers of BLE-B tags have proposed a variety of designs. While some designs require sine wave incident signals
[0010] , other designs can work with ambient Bluetooth signals [7], Tags that require sine wave incident signals can work with either standalone carrier generators or with carriers generated by reverse- whitening the data in a second Bluetooth transmitter [9], To the best of the inventor’s knowledge, tags in the literature generate classical Frequency Shift Keying (FSK) waveforms instead ofAttorney Docket: SDSUP002WOGaussian-FSK (GFSK) that is specified in the Bluetooth standard. Nevertheless, the signals can be demodulated by an unmodified Bluetooth receiver. Tags that work with ambient Bluetooth signals are efficient in that they use available productive signals to embed their data. Furthermore, their signals are fully Bluetooth compliant since they rescatter an incident Bluetooth signal at the appropriate frequency
[0011] , However, there are other implementation complexities. The receiver may need to know the data in the incident Bluetooth signal in order to separate the tag’s data.
[0007] It has been observed that that the error performance of BLE-B is worse than expected given the signal strength [9], Works such as [9] and
[0010] achieve acceptable performance by increasing the frequency deviation from 250 kHz (BLE specified) to 500 kHz. It has been reported that this approach improves performance on the given channel by 6 dB due to Frequency Modulation (FM) gain,
[0012] S. Nagaraj and R. Yaqo, “A Frequency Modulation technique for SNR improvement in backscatter radios,” IEEE Commun. Lett., pp. 3956-3959, Dec 2021, but increases interference to neighboring channels due to widening of the main lobe to 3 MHz (versus 2 MHz of BLE).
[0008] BLE and BLE-B ackscatter
[0009] BLE (Bluetooth Low Energy): The BLE standard employs frequency hopping over a set of 40 available physical channels,
[0013] M. Woolley, Bluetooth core specification version 5.0 feature enhancements. Bluetooth, 2021. The channels are between 2.4-2.48 GHz and are spaced 2 MHz apart. Some of the channels are set aside for transmitting advertising packets only, while the rest can carry data packets. A data stream at the rate of R = 1 Mbps is GFSK modulated. The first step in the modulator is the filtering of the digital pulses with a Gaussian filter of 3 dB bandwidth 0.5 MHz. This makes the data signal smooth and also limits its spectral side lobes. This signal is input to a frequency modulator with peak frequency deviation of Af = 250 kHz on either side of the carrier frequency. The modulation index, P, of the frequency deviation of Af = 250 kHz on either side of the carrier frequency. The modulation index, P, of the transmission is defined as fi = Af / B = 0.5. From Carson’s rule,
[0014] B. P. Lathi and Z. Ding, Modern Digital and Analog Communication, 5th ed. Oxford University Press, 2018, the bandwidth of the GFSK modulated signal isBbie= 2B( + 1) (1)= 1.5 MHz, (2)Attorney Docket: SDSUP002WO where B represents baseband and Bble is RF baseband, necessitating the 2 MHz channel spacing while also allowing for a guard band between neighboring channels.
[0010] BLE-B (BLE-B ackscatter): Existing works on BLE-backscatter generate classical FSK waveforms instead of GFSK. A data bit 1 is backscattered as a signal of frequency Af above the center frequency while a data bit 0 is transmitted at the frequency zl / 'below. The use of FSK avoids the intensive Gaussian filtering in the tag, which can require significant energy consumption, which is to be avoided in low power applications. The rectangular instantaneous frequency profile is created by switching the antenna-load impedance at the required frequency. A small amount of harmonic loss is inherent in all backscatter communication systems since the switched signal is always rectangular, instead of sinusoidal. The intended frequency of transmission is at the fundamental frequency of the rectangular signal. The loss due to components at other harmonics is typically < 10 %.
[0011] The transmitted FSK signal from a BLE-B tag can be described by the equation: s(t) = Asin[27 ( c+ ^i(~ l) i frect(t~ iT) t + 0i)], (3) where, A is the amplitude of the signal, bi E {-1, +1 } is the i-th tag data bit and 9i is the angle that ensures phase-continuity of the signal between bits. Also, T is the tag data bit duration and fc is the center frequency of the backscattered signal. Lastly, the ‘rect’ function indicates that the frequency in the i-th bit duration is constant at fc+ (-l)biAf.
[0012] FSK - GFSK Mismatch
[0013] BLE vendors typically can use a few different GFSK receiver technologies with generally similar performances. The basic principle in most of these receivers is FM discrimination, wherein, frequency variations are converted to amplitude variations by differentiation
[0014] , For analysis, we use the simple design shown in Figure 1 which shows a system model of a GFSK receiver based on FM discrimination. An FM discriminator 104 converts frequency variations in an FM modulated signal 102 to amplitude variations in an AM modulated signal 106. A low pass filter 108 filters the AM modulated signal and provides an output signal 110 to a detector (not shown). The Amplitude Modulated (AM) signal at the output of the differentiator can be demodulated using any AM demodulation technique such as envelope detection or quadrature (EQ) correlation.Attorney Docket: SDSUP002WO
[0014] Figure 2 is an illustrative graph showing Signal to Interference and Noise Ratio (SINR) versus receiver bandwidth when an GFSK signal is demodulated 202 or an FSK signal is demodulated 204 is demodulated by an FM discrimination technique.
[0015] Suppose that the receiver operates in the presence of additive white Gaussian noise n(t) of Power Spectral Density (PSD) No / 2 Watts / Hz. The Carrier-to-Noise Ratio (CNR) at the receiver input is:
[0016] The differentiator converts the instantaneous frequency of the FSK signal in (3) into an amplitude, which can be demodulated using any AM technique such as envelope detection. At this point, the baseband signal can be represented as:rect(t -iT) + n'(t), (5) and is simply a Non-Retum-to-Zero (NRZ) line code. At the same time, the differentiator alters the noise PSD to a parabola. The effect of an envelope detector on the noise can be ignored at practical CNR values
[0014] , The two-sided noise PSD is:Noise PSD = NO \j2nf\2 (6)= 2TT2 2NO. (7)
[0017] The parabolic noise PSD above necessitates a judicious choice for the bandwidth of the Low Pass Filter (LPF) that follows. The LPF bandwidth is chosen to maximize the SINR at the detector input. This filter plays a critical role as the noise power at its output varies as the cube of its bandwidth due to the parabolic PSD in (7). Figure 2 shows the variation of the SINR at the output of this LPF as a function of its bandwidth for both GFSK and FSK signals. The signal and interference powers are analytically calculated from the power spectrum of the signal and the noise power is calculated from (7)
[0015] A. J. Viterbi and J. K. Omura, Principles of Digital Communication and Coding. McGraw- Hill, 1979; and
[0016] S. Benedetto and E. Biglieri, Principles of Digital transmission: with wireless applications. Kluwer Academic / Plenum Publishers, New York, 1999.
[0018] In commercially available BLE devices, this LPF is optimized assuming a GFSK modulated signal at the receiver input. From Figure 2, the optimal bandwidth for an ideal brickwall LPF is about 570 kHz. This is consistent with empirical studies by other authors, based on BER minimization, which have shown an optimum 3-dB bandwidth ofAttorney Docket: SDSUP002WO about 600 kHz,
[0017] R. Gonce, Software defined Bluetooth Low Energy transceiver for reconfigurable loT smart sensors. Ecole polytechnique de Louvain, 2020. This optimum filter bandwidth depends on the input CNR. Commercial BLE receivers are optimized for a CNR of 14 dB that accounts for the required detector performance (BER = 0.1%) and also for the noise figures of the analog / RF components in the front end. Since front-end noise figure is not considered in this work, we have assumed a CNR of 11 dB.SUMMARY
[0019] In one aspect, a wireless communication system includes a baseband signal processing circuit, a modulator circuit and an antenna. The baseband signal processing circuit is configured to produce a baseband signal that provides a sequence of symbols at a baseband frequency, wherein each symbol can have either one of a first symbol value and a second symbol value, and wherein transitions between symbol values within the baseband signal occur in a single voltage level step. The baseband signal processing circuit is configured to sample the baseband signal at twice the baseband frequency to produce an upsampled signal that provides a sequence of samples of the baseband signal at twice the baseband frequency. The based band signal processing circuit is configured to combine each sample within the upsampled signal with a preceding sample within the upsampled signal to produce a modified baseband signal that provides a sequence of modified symbols at twice the baseband frequency. Each symbol can have either one of the first symbol value or the second symbol value, and each transition between symbol values within the modified baseband signal includes an intermediate symbol value. The modulator circuit is coupled to modulate frequency of a carrier signal based upon symbol values of symbols within the modified baseband signal to produce a frequency modulated carrier signal. The antenna is coupled to transmit the modulated carrier signal.
[0020] In one aspect, a wireless communication system includes a baseband signal processing circuit, a modulator circuit and an antenna. The baseband signal processing circuit is configured to produce a baseband signal that provides a sequence of symbols at a baseband frequency, wherein each symbol can have either one of a first symbol value and a second symbol value and wherein transitions between symbol values within the baseband signal occur in a single voltage level step. The baseband signal processing circuit is configured to sample the baseband signal at N (where N>2) times the baseband frequency to produce an upsampled signal that provides a sequence of samples of theAttorney Docket: SDSUP002WO baseband signal at N times the baseband frequency. The baseband signal processing circuit is configured to combine each sample within the upsampled signal with immediately preceding N-l samples within the upsampled signal to produce a modified baseband signal that provides a sequence of modified symbols at N times the baseband frequency, wherein each symbol can have either one of the first symbol value and the second symbol value, and wherein each transition between symbol value levels within the modified baseband signal includes N voltage level change steps. The modulator circuit is coupled to modulate frequency of a carrier signal waveform based upon symbol value levels of symbols within the modified baseband signal to produce a frequency modulated carrier signal. The antenna is coupled to transmit the modulated carrier signal waveform.BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In the drawings, which are not necessarily drawn to scale, like numerals may describe similar components in different views. Like numerals having different letter suffixes may represent different instances of similar components. The drawings illustrate generally, by way of example, but not by way of limitation, various embodiments discussed in the present document.
[0022] Figure 1 is an illustrative block diagram representing a system model of a GFSK receiver based on FM discrimination.
[0023] Figure 2 is an illustrative graph showing SINR versus receiver bandwidth when GFSK or FSK is demodulated by an FM discrimination technique.
[0024] Figure 3A is an illustrative example NRZ waveform for six (6) symbols of an example data sequence.
[0025] Figure 3B is an example corresponding upsampled duobinary (U-D) waveform for the six (6) symbols of Figure 3A.
[0026] Figure 4 is an illustrative block diagram of a first example upsampled duobinary (U-D) signal generator circuit in accordance with some embodiments.
[0027] Figure 5A is an illustrative graph showing spectra of conventional FSK based on NRZ with subcarrier frequency fd = 1 MHz.Attorney Docket: SDSUP002WO
[0028] Figure 5B is an illustrative graph showing spectra of FSK based on U-D with subcarrier frequency fd = 3 MHz in accordance with the generator circuit of Figure 4.
[0029] Figure 6 is an illustrative example graph showing comparison of baseband spectra of NRZ, U-D, and Gaussian-filtered NRZ.
[0030] Figure 7 is an illustrative example graph showing BER comparison of classical FSK signal BER and U-D FSK signal BER when demodulated by a GFSK receiver.
[0031] Figure 8 is an illustrative block diagram of an example upsampled polybinary (U- D) signal generator circuit that in accordance with some embodiments.
[0032] Figure 9 is an illustrative graph that shows an FSK signal generated when an upsampled polybinary sequence is frequency modulated and transmitted.
[0033] Figure 10 is an illustrative graph that shows the comparison of baseband spectra of classical NRZ, upsampled polybinary NRZ, and baseband spectra NRZ.
[0034] Figure 11 is an illustrative example block diagram showing an upsampling duobinary processing block coupled to provide U-D output signals as input to a Gaussian filter block.
[0035] Figure 12 is an illustrative block diagram of a backscatter wireless tag device in accordance with some embodiments.DETAILED DESCRIPTION
[0036] Upsampled Duobinary (U-D) NRZ
[0037] The mismatch between classical FSK modulated signals and GFSK receivers is a problem to resolve since the prior works [7]-[l 1] identify a need to enable an unmodified BLE device to receive a tag signal. This disclosure describes a new tag signaling circuit and method that is better matched to GFSK receivers. The disclosed circuit and method produce a waveform based on a two-step transition from data bit 0 to 1 and vice-versa. No complex filtering, which can be energy intensive, is necessary in the tag. Low energy consumption is an important parameter in many RFID applications. As explained below, simulations show that about 5 dB improvement in the Bit Error Rate (BER) can be achieved by this technique.Attorney Docket: SDSUP002WO
[0038] More particularly, an example new waveform is described that better compacts the spectrum in the sub-600 kHz band without the need for complex waveform generation (such as Gaussian filtering). The waveform and its generation are described below.
[0039] Instantaneous Frequency: Dependence of the instantaneous frequency co(t) on data bits is very similar to bipolar NRZ, except that the single-step transitions from -1 — 1 and 1 — -1 of NRZ are replaced by two-step transitions (through 0). The two steps for a 2x U-D signal are from -1 to 0 to +1 and from +1 to 0 to -1, where the +1 and -1 are symbol values and 0 is an intermediate symbol value. Mathematically, it can be described as follows. Suppose that either bi = -1 and bi+ = 1 or bi = 1 and bi+1 = -1. Then aft) = 0 for (i - 0.5)T < t < iT. (8)
[0040] When there is no bit transition at iT, the instantaneous frequency will follow classical bipolar NRZ. (where 7'is the tag bit data duration). A sample bit sequence and the corresponding instantaneous frequency profile are shown in Figures 3A-3B. More particularly, Figure 3A is an illustrative example NRZ waveform for six (6) symbols of an example data sequence bi = +1, -1, -1, +1, +1, -1, -1, and Figure 3B is an example corresponding upsampled duobinary (U-D) waveform for the six (6) symbols of Figure 3A. Duobinary coding is a form of correlative coding. Different symbol values are represented by different voltage levels. Note that the waveform in Figure 3A is binary with two values E {+1, -1}. The waveform in Figure 3B is ternary with three values E {0, +1, -1}. Values +1 and -1 are symbol values and value 0 is an intermediate symbol value. Symbol levels are, in hardware, represented by voltages that are proportional to the corresponding symbol values. In the waveform of Figure 3A, the symbol value +1 may be represented by a first voltage level (e.g., 2 volts) and the symbol value -1 may be represented by a second voltage level (e.g., 0 volts). In the waveform of Figure 3B, the symbol value +1 may be represented by a first voltage level (e.g., 2 volt), the intermediate symbol value 0 may be represented by a second voltage level (e.g., 1 volts), and the symbol value -1 may be represented by a third voltage (e.g., 0 volts).
[0041] Figure 4 is an illustrative block diagram of an example upsampled duobinary (U-D) signal generator circuit 400 in accordance with some embodiments. The U-DAttorney Docket: SDSUP002WO generator 400 includes an upsampling block 402 that samples bi values of an NRZ signal at twice the NRZ frequency to produce an upsampled signal that provides a sequence of sample values ci of the NRZ signal values at twice the NRZ frequency. Each upsample value is provided to a delay circuit, which delays the upsample value ci by a time delay £>, to produce a delayed upsample value. Each upsample value and each corresponding delayed upsampled value is provided to adder 406, which sums them and provides a corresponding upsample output value di. In an example embodiment, D is a 0.57 time delay.
[0042] Figure 5A is an illustrative graph showing spectra of conventional FSK based on NRZ with subcarrier frequency fd= 3 MHz. Figure 5B is an illustrative graph showing spectra of FSK based on U-D with subcarrier frequency fd= 3 MHz in accordance with the generator circuit of Figure 4.
[0043] Hardware Implementation: Referring to the generator circuit 400 of Figure 4, let ci be the data sequence obtained by upsampling (by a factor of 2) the original data sequence bi. The Duobinary function,
[0018] J. G. Proakis, Digital Communications, 4th ed. McGraw-Hill, 2001, is applied on the sequence ci to get the U-D bit stream di as follows: di = ci + ci- (9) which can be easily implemented on passive tags. This is the only extra processing needed over that of BLE-B in the literature to implement U-D transmission. The duobinary output will be ternary {0, +2,-2}.
[0044] Backscatter Waveform: The tag signal is first generated as the Frequency Modulation of the baseband signal co(t) at the subcarrier frequency of fd. This signal can be generated by switching between three frequencies fd + fd and fd - Af depending on the value of aft), which can be +2, 0 or -2, respectively, at the duobinary output.
[0045] Lastly, the above generated tag signal controls an RF switch connecting the antenna to the load impedances. If the interrogator’s signal is a sine wave at frequency fi, the backscattered signal will be at fc =fi + fd. This design is common to other BLE-B tags in the literature. This switching results in the tag backscattering a GFSK-compatible signal at the center frequency fc =fi + fd.Attorney Docket: SDSUP002WO
[0046] Figure 6 is an illustrative example graph showing comparison of baseband spectra of NRZ 602, U-D 604 and Gaussian-filtered NRZ 606.
[0047] Figure 7 is an illustrative example graph showing BER comparison of classical FSK signal BER 704 and U-D FSK signal BER 702 when demodulated by a GFSK receiver.
[0048] Simulation Results
[0049] Important advantages of the U-D waveform above are in regards to (i) spectrum of the baseband signal and (ii) compatibility with GFSK receivers. The inventor compared, using analysis and simulations, classical FSK based on based on rectangular NRZ baseband and FSK based on U-D waveforms. It is noted that U-D waveforms are also rectangular.
[0050] In Figures 5A-5B, there is shown a comparison of spectra of classical FSK and U-D FSK described in Section IV. It will be appreciated from Figures 5A-5B, that both have roughly the same width of the main lobe - about 1.5 MHz. However, the first side lobe of U-D FSK is about 10 dB below that of classical FSK. This is significant from the perspective of minimizing interference to adjacent channels. For this simulation, chosen 3 MHz was chosen as the subcarrier frequency fd.
[0051] Baseband spectra also were compared for the two techniques. For classical FSK, the baseband signal is just an NRZ signal and so has the sine-shaped spectrum. On the other hand, the U-D baseband signal spectrum is impacted by the duobinary transfer function which also has a low-pass characteristic. For comparison, the baseband spectrum of GFSK (which is Gaussian filtered NRZ) is also shown in Figure 6. It can be seen that both GFSK and U-D FSK have more energy in the sub- 600 kHz band than classical FSK. From the discussion above, this directly translates to a higher SINR at the detection stage.
[0052] Lastly, the BERs of classical FSK and U-D FSK are compared in Figure 7. In this simulation, both classical FSK and U-D FSK are demodulated by a conventional quadrature GFSK demodulator, which makes decisions on the transmitted bit sequence by sampling the sign of the phase derivative IQ' - QI' at appropriate time instants. An improvement of over 5 dB is observed with U-D signaling over classical FSK. Uncoded data bits were considered for this simulation.
[0053] Upsampled PolybinaryAttorney Docket: SDSUP002WO
[0054] The main idea of the previous section can be extended to polybinary signaling with appropriate upsampling rates. For example, the NRZ sequence bi can be upsampled by three (3) to get the sequence ci at thrice the symbol rate. Polybinary Function can then be applied to the sequence as shown below to get the sequence di as: di = ci + ci-1 + ci-2. (10)
[0055] The sequence di E {-3, 1, +1, +3} has four possible values and is at triple the symbol rate of bi. Sequence di is Frequency Modulated and transmitted. Frequency Modulation can be achieved by mapping di values {-3, 1, +1, +3} to frequencies {fd — Af fd — Af / 3, fd +Af{ . respectively. No complex filtering is necessary. Polybinary coding is a form of correlative coding.
[0056] Dependence of the instantaneous frequency co(t) on data bits is similar to bipolar NRZ, except that the single-step transitions from -1 — 1 and 1 — -1 of NRZ are replaced by three-step transitions. The three steps for a 3x U-D signal are from -3 to -1 to +1 to +3 and from +3 to +1 to -1 to -3, where +3 and -3 are symbol values and +1 and -1 are intermediate symbol values. Symbol levels are, in hardware, represented by voltages that are proportional to the corresponding symbol values. A first symbol value +3 may be represented as a first voltage level. A second symbol value -3 may be represented as a second voltage level that is less than the first voltage level. A first intermediate symbol value +1 may be represented as a third voltage level midway between the first voltage level and a fourth voltage level. A second intermediate symbol value -1 may be represented as the fourth voltage level midway between the third voltage level and the second voltage level.
[0057] Figure 8 is an illustrative block diagram of an example upsampled polybinary (U-D) signal generator circuit 800 that in accordance with some embodiments. The U-D generator 800 includes an upsampling block 802 samples bi values of an NRZ signal at thrice the NRZ frequency to produce an upsampled signal that provides a sequence of sample values ci of the NRZ signal values at thrice the NRZ frequency. Each upsample value is provided to delay circuit chain 804 that includes a first delay circuit 804-1 and a second delay circuit 804-2 that each delays the upsample value ci by a time delay D. The first delay circuit 804-1 and the second delay circuit 804-2 are coupled to delay each upsample value ci by successive time delay D increments. TheAttorney Docket: SDSUP002WO upsample circuit 802 and the first and second delay circuits 804-1, 804-2 are coupled such that a current non-delayed output of the upsample circuit 802 and the current contents of each of the first and second delay circuits 804-1, 804-2 are provided to an adder 406, which sums them and provides a corresponding upsample output value di. In an example embodiment, D is a 0.5T time delay. Stated differently, the upsample circuit 802 and the first and second delay circuits 804-1, 804-2 are coupled such that that for each upsample value cz, the first delay circuit provides previous upsample value ci-1 to the adder 406, and second delay circuit provides previous upsample value ci-2 to the adder 406.
[0058] It is contemplated that a polybinary U-D signal generator with upsampling by N can be provided that includes a delay chain of length N-l and with N-l steps required to transition between symbol values.
[0059] Figure 9 is an illustrative graph that shows the FSK signal generated when the upsampled poly binary sequence di is frequency modulated and transmitted. The subcarrier frequency in this example is also set to 3 MHz. The side lobes are even smaller than those of U-D signals.
[0060] Figure 10 is an illustrative graph that shows the comparison of baseband spectra of classical NRZ 1002, upsampled polybinary NRZ 1004, and baseband spectra NRZ 1006. The spectra of Gaussian-filtered NRZ and upsampled polybinary NRZ are nearly indistinguishable in the sub-600 kHz band. Based on this, it is expected their BER performance to be nearly identical in BLE receivers.
[0061] Upsampled polybinary can be used in active BLE transmitters as well owing to the nearly indistinguishable performance (when compared with GFSK) and much simpler waveforms. Upsampled polybinary does not need Gaussian filters. Furthermore, instantaneous frequency generation is also simplified. Only four instantaneous frequencies need to be generated as opposed to the continuum of frequencies that GFSK requires. This fact can be used in the design of BLE transmitters to lower the power consumption. This is particularly suited for BLE devices that require batteries to last years.
[0062] U-D GFSK
[0063] The U-D technique can be applied even to GFSK modulation to improve the performance by a further 3 dB. In this technique, data bits bi are first U-D encoded asAttorney Docket: SDSUP002WO in Figure 4 or as in Figure 8 to get the signal di. This signal can then be Gaussian filtered and frequency modulated as in a traditional GFSK modulator and transmitted. Figure 11 is an illustrative block diagram showing a transmission signal encoding circuit 1100 that includes an upsampling duobinary processing block 1102 that is coupled to provide a U-D signal as input to a modulation circuit 1105 that includes a Gaussian filter block 1104 that is coupled to provide an output signal to an FSK modulation block 1106. In some embodiments, a U-D signal is converted from a baseband frequency to 2.45 GHz in BLE in two steps, first to a mid-range (intermediate) frequency (e.g., under 50 MHz) digitally and then to 2.45 using an analog RF circuit. The RF transmission signal is provided to an antenna 1108 for transmission. The resulting signal, which is U-D processed, Gaussian filtered, and frequency modulated, can also be demodulated by a traditional BLE device and the resulting BER is at least 3 dB better than GFSK as transmitted from current BLE transmitters. This is due to the fact that U-D GFSK better compacts the baseband energy in the sub-600 kHz band.
[0064] Tag Device with Backscatter Transmission
[0065] Figure 12 is an illustrative block diagram of a backscatter wireless tag device 1200 in accordance with some embodiments. The backscatter tag device includes a baseband processing circuit 1202, a frequency modulator 1204, an RF switch 1206, an antenna 1208, an energy harvesting and storage and logic block 1210, a receive path with a wakeup algorithm logic block 1212, and controller 1214. Persons skilled in the art will appreciate that backscatter techniques are used to produce an RF transmission signals through reflection of an RF receive signal. Persons skilled in the art will appreciate that backscatter techniques can obviate the requirement for a power source and active power-intensive radio frequency (RF) components on a wireless tag device to generate an RF transmission signal.
[0066] The baseband processor 1202 is configured to be operable to convert data into a signal used by the frequency modulator 1204 to modulate a carrier frequency for transmission over the antenna. An example baseband processor is configured to implement the U-D generator 400 of Figure 4. As such, the baseband processor 1202 provides a sequence of symbols U-D NRZ symbols di at a baseband frequency. An example baseband processor 1202 configured to interoperate with Bluetooth has aAttorney Docket: SDSUP002WO baseband signal in a frequency range under 2 MHz or 4 MHz depending on BLE mode (IM or 2M).
[0067] In operation, the antenna 1208 receives an RF interrogator signal 1216. The energy harvesting and storage and logic block 1210 harvests and stores energy imparted with the RF interrogator signal 1216 through a rectifier to convert RF signal to DC and charge a capacitor for example. The energy harvesting and storage and logic block 1210 is coupled to use the harvested energy to power the device 1200. The receive path with a wakeup algorithm logic block 1212 configured to detect a signature in the incident signal 1216 as a sign to trigger wake up and start transmitting. The controller circuit 1214 is configured with logic to determine NRZ data input to provide the baseband processor 1202 for U-D NRZ modulation based upon input received from the wakeup algorithm logic block 1212.
[0068] During RF transmission, the frequency modulator 1204 modulates the ternary U-D signal so as to map the ternary U-D NRZ signal to ternary FSK modulated signal with frequencies,^ - Affd,fd + Af\o produce a frequency modulated carrier signal. A portion of the modulated signal at a frequency fd~ Af corresponds to a -1 value portion of the U-D signal. A portion of the modulated signal at a frequency fd corresponds to a 0 value portion of the U-D signal. A portion of the modulated signal at a frequency fd + Af corresponds to a +1 value portion of the U-D signal.
[0069] The RF switch 1206 reflects a received interrogator signal that is incident upon the antenna 1208 to produce a backscatter RF transmission signal 1218. More particularly, the RF switch 1204 receives the U-D signal from the baseband modulator 1202 and causes the antenna 1208 to radiate the RF transmission signal 1218 that is generated by modulator 1204. In some embodiments, an RF switch 1204 is configured to adjust a load impedance coupled to the antenna in accordance with a voltage of the frequency modulated (with or without Gaussian filtering) carrier signal to be backscattered, which results in an adjusted frequency modulated RF signal 1218 being radiated by the antenna 1208. A rectangular instantaneous frequency profile is created by switching the antenna-load impedance at the required frequency. For example, fortheU-D generator 400, the antenna load impedance is modulated at frequency fa. For the polybinary (U-D) signal generator circuit 800, the antenna load impedance is modulated by mapping di values {-3, 1, +1, +3} to frequencies {fd~ Affd- Af / 3,fd + Af / 3,fd + A }, respectively.Attorney Docket: SDSUP002WO
[0070] An alternative embodiment baseband processor can be implemented to produce a baseband signal based upon by N-times upsampling where N>2 (e.g., 3x, 4x, etc.) and corresponding correlative coding as described with reference to Figure 8. Moreover, an alternative embodiment modulator can be implemented that includes a Gaussian filter as explained with reference to Figure 11.
[0071] The above description is presented to enable any person skilled in the art to create and use a wireless communication system that generates an upsampled correlatively coded baseband signal. Various modifications to the embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the invention. In the preceding description, numerous details are set forth for the purpose of explanation. However, one of ordinary skill in the art will realize that the embodiments in the disclosure might be practiced without the use of these specific details. In other instances, well-known processes are shown in block diagram form in order not to obscure the description of the invention with unnecessary detail. Identical reference numerals may be used to represent different views of the same or similar item in different drawings. Thus, the foregoing description and drawings of examples in accordance with the present invention are merely illustrative of the principles of the invention. Therefore, it will be understood that various modifications can be made to the embodiments by those skilled in the art without departing from the spirit and scope of the invention, which is defined in the appended claims.
Claims
Attorney Docket: SDSUP002WOCLAIMS1. A wireless communication system comprising: a baseband signal processing circuit configured to, produce a baseband signal that provides a sequence of symbols at a baseband frequency, wherein each symbol can have either one of a first symbol value and a second symbol value, and wherein transitions between symbol values within the baseband signal occur in a single voltage level step; sample the baseband signal at twice the baseband frequency to produce an upsampled signal that provides a sequence of samples of the baseband signal at twice the baseband frequency; and combine each sample within the upsampled signal with a preceding sample within the upsampled signal to produce a modified baseband signal that provides a sequence of modified symbols at twice the baseband frequency, wherein each symbol can have either one of the first symbol value or the second symbol value, and wherein each transition between symbol values within the modified baseband signal includes an intermediate symbol value; a modulator circuit coupled to modulate frequency of a carrier signal based upon symbol values of symbols within the modified baseband signal to produce a frequency modulated carrier signal; and an antenna coupled to transmit the modulated carrier signal.
2. The wireless communication system of claim 1, wherein the modulator is coupled to use a Gaussian filter to modulate the modified baseband signal.
3. The wireless communication system of claim 1, wherein the modulator is configured to use frequency shift keying to modulate the modified baseband signal.Attorney Docket: SDSUP002WO4. The wireless communication system of claim 1, wherein the modulator is configured to use a Gaussian filter and frequency shift keying to modulate the modified baseband signal.
5. The wireless communication system of claim 1, wherein the baseband signal comprises a Non-Retum-to-Zero (NRZ) signal.
6. The wireless communication system of claim 1, wherein combining includes combining each sample within the upsampled signal with an immediately preceding sample delayed within the upsampled signal by one- half of symbol duration before upsampling.
7. The wireless communication system of claim 1, wherein the baseband signal comprises an NRZ signal; wherein the modified baseband signal includes three voltage levels: a first voltage level that represents a first symbol value, a second voltage level that represents a second symbol value, and a third voltage level between the first voltage level and the second voltage level that represents a transitional step between the first symbol value and the second symbol value; and wherein the carrier signal includes three frequencies: a first frequency fd + Af that corresponds to the first symbol value, a second frequency fd - Af that corresponds to the second symbol value, and a frequency fd that corresponds to the transitional step.
8. The wireless communication system of claim 7, wherein the modulator is coupled to use a Gaussian filter to produce a GFSK modulated signal based upon the modified baseband signal; andAttorney Docket: SDSUP002WO wherein the GFSK modulated signal has a peak frequency deviation of Af such that voltages at the output of the Gaussian filter correspond to frequencies between fd + Af and fd - Af.
9. The wireless communication system of claim 1, wherein the baseband signal frequency is in a frequency range, less than 2 MHz; and wherein the carrier frequency has a frequency range 2 MHz and 40 MHz.
10. The wireless communication system of claim 1 further including: an energy harvesting and storage circuit coupled to harvest energy from an interrogator signal and to power to the baseband signal processing circuit.
11. The wireless communication system of claim 1 further including: a controller circuit configured with logic to determine data to provide the baseband processor based upon the interrogator signal; and an energy harvesting and storage circuit coupled to harvest energy from an interrogator signal and to power to the baseband signal processing circuit and the controller circuit.
12. The wireless communication system of claim 1 further including: an RF switch configured to produce a backscatter RF signal based upon the frequency modulated carrier signal.
13. The wireless communication system of claim 1, wherein the preceding sample within the upsampled signal includes an immediately preceding sample within the upsampled signal.
14. A wireless communication system comprising:Attorney Docket: SDSUP002WO a baseband signal processing circuit configured to, produce a baseband signal that provides a sequence of symbols at a baseband frequency, wherein each symbol can have either one of a first symbol value and a second symbol value, and wherein transitions between symbol values within the baseband signal occur in a single voltage level step; sample the baseband signal at N (where N>2) times the baseband frequency to produce an upsampled signal that provides a sequence of samples of the baseband signal at N times the baseband frequency; and combine each sample within the upsampled signal with immediately preceding N-l samples within the upsampled signal to produce a modified baseband signal that provides a sequence of modified symbols at N times the baseband frequency, wherein each symbol can have either one of the first symbol value and the second symbol value, and wherein each transition between symbol value levels within the modified baseband signal includes N voltage level change steps; a modulator circuit coupled to modulate frequency of a carrier signal waveform based upon symbol value levels of symbols within the modified baseband signal to produce a frequency modulated carrier signal; and an antenna coupled to transmit the modulated carrier signal waveform.
15. The wireless communication system of claim 14, wherein the modulator is coupled to use a Gaussian filter to modulate the modified baseband signal.
16. The wireless communication system of claim 14, wherein the modulator is configured to use frequency shift keying to modulate the modified baseband signal.
17. The wireless communication system of claim 14,Attorney Docket: SDSUP002WO wherein the modulator is configured to use a Gaussian filter and frequency shift keying to modulate the modified baseband signal.
18. The wireless communication system of claim 14, wherein N=3; wherein the modified baseband signal includes four voltage levels: a first voltage level that represents a first symbol value, a second voltage level that represents a second symbol value, and a third voltage level between the first voltage level and a fourth voltage level that represents a first transitional step between the first symbol value and the second symbol value, and the fourth voltage level between the second voltage level and the third voltage level that represents a second transitional step between the first symbol value and the second symbol value; and wherein the carrier signal includes four frequencies: a first frequency fd + Af that corresponds to the first symbol value, a second frequency fd - Af that corresponds to the second symbol value, a frequency fd + Af / 3 that corresponds to the first transitional step, and fd - Af / 3 that corresponds to the second transitional step.
19. The wireless communication system of claim 18, wherein the modulator is coupled to use a Gaussian filter to produce a GFSK modulated signal based upon the modified baseband signal; and wherein GFSK modulated signal has a peak frequency deviation of Af so that voltages at the output of the Gaussian filter correspond to frequencies between fd + Af and fd - Af.
20. The wireless communication system of claim 14, wherein the preceding N-l samples within the upsampled signal includes an immediately preceding N-l samples within the upsampled signal.
Citation Information
Patent Citations
Multi-antenna transmission and receive apparatus
KR101579898B1
Circuit and method for binary modulation
US20050157816A1
Digital Signal Transmission for Wireless Communication
US20100124290A1
Systems and methods for delta-sigma digitization
US20190273518A1
Multi-antenna reader channel state information acquisition
US20240250852A1