WO2019089557A1 - Communication rfid échelonnable utilisant une excitation multifréquence - Google Patents

Communication rfid échelonnable utilisant une excitation multifréquence Download PDF

Info

Publication number
WO2019089557A1
WO2019089557A1 PCT/US2018/058167 US2018058167W WO2019089557A1 WO 2019089557 A1 WO2019089557 A1 WO 2019089557A1 US 2018058167 W US2018058167 W US 2018058167W WO 2019089557 A1 WO2019089557 A1 WO 2019089557A1
Authority
WO
WIPO (PCT)
Prior art keywords
frequency
samples
rfid tags
reader
messages
Prior art date
Application number
PCT/US2018/058167
Other languages
English (en)
Inventor
Prasun Sinha
Tanmoy DAS
Gopi TUMMALA
Original Assignee
Ohio State Innovation Foundation
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ohio State Innovation Foundation filed Critical Ohio State Innovation Foundation
Publication of WO2019089557A1 publication Critical patent/WO2019089557A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K7/00—Methods or arrangements for sensing record carriers, e.g. for reading patterns
    • G06K7/10—Methods or arrangements for sensing record carriers, e.g. for reading patterns by electromagnetic radiation, e.g. optical sensing; by corpuscular radiation
    • G06K7/10009—Methods or arrangements for sensing record carriers, e.g. for reading patterns by electromagnetic radiation, e.g. optical sensing; by corpuscular radiation sensing by radiation using wavelengths larger than 0.1 mm, e.g. radio-waves or microwaves
    • G06K7/10019—Methods or arrangements for sensing record carriers, e.g. for reading patterns by electromagnetic radiation, e.g. optical sensing; by corpuscular radiation sensing by radiation using wavelengths larger than 0.1 mm, e.g. radio-waves or microwaves resolving collision on the communication channels between simultaneously or concurrently interrogated record carriers.
    • G06K7/10069—Methods or arrangements for sensing record carriers, e.g. for reading patterns by electromagnetic radiation, e.g. optical sensing; by corpuscular radiation sensing by radiation using wavelengths larger than 0.1 mm, e.g. radio-waves or microwaves resolving collision on the communication channels between simultaneously or concurrently interrogated record carriers. the collision being resolved in the frequency domain, e.g. by hopping from one frequency to the other
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K7/00—Methods or arrangements for sensing record carriers, e.g. for reading patterns
    • G06K7/0008—General problems related to the reading of electronic memory record carriers, independent of its reading method, e.g. power transfer

Definitions

  • This invention generally relates to Radio-Frequency Identification (RFID) and, in particular, to systems, methods, and computer program products for reading RFID tags.
  • RFID Radio-Frequency Identification
  • An RFID tag includes an antenna and a circuit that adjusts the electrical load coupled to the antenna.
  • the load can be adjusted so that it has an impedance that matches the antenna or that is mismatched to the antenna.
  • Energy of the radio signal incident on the antenna from a tag reader is either absorbed or reflected by the RFID tag depending on whether the impedance is matched or mismatched.
  • RFID tags modulate the signal backscattered by the antenna by changing the impedance of the load. The change in impedance is typically achieved by selectively activating a transistor that alters the impedance of the load. Modulating the load impedance causes a modulated backscattered signal to be reflected from the RFID tag. This modulated signal can be used to convey information to the tag reader.
  • RFID tags are popular for several reasons.
  • the low cost of RFID tags enables large numbers of tags to be deployed in a single location, such as a store shelf.
  • RFID tags also do not require an internal power source, and can be designed to harvest energy from signals received from a tag reader. This implies a zero-maintenance cost after deployment.
  • RFID tags can communicate with the tag reader in non-line-of-sight scenarios. These features make RFID tags attractive for object tracking and monitoring.
  • the low energy design of RFID tags also results in several limitations. For example, the Media Access Control (MAC) protocols implemented in RFID tags must be kept simple to conserve energy.
  • MAC Media Access Control
  • RFID tags In addition, the low Radio Frequency (RF) power output of the RFID tags results in a low Signal-to-Noise Ratio (SNR) at the reader and limits messages to a few hundred bits of data.
  • SNR Signal-to-Noise Ratio
  • RFID tags may be attached to a large number of items in a store or warehouse, resulting in an extremely dense deployment.
  • FIG. 1 depicts a scenario in which a tag reader could receive responses from hundreds RFID tags associated with books on a bookshelf 2.
  • the exemplary bookshelf 2 includes seven shelves, has a width w ⁇ of three feet and a height hi of seven foot. For a fully stocked bookshelf 2 having approximately 30 books on each shelf, and assuming each book has one RFID tag, approximately 210 tags can be expected to respond to a tag reader 4 having a typical communication range d ⁇ of 18 feet.
  • Hardware non-uniformities include varying RFID capacitor charging times and clocks having varying accuracy.
  • the varying RFID capacitor charging time causes message transmissions from tags to be delayed by a random amount of time known as a starting delay.
  • Variability in the accuracy of clocks in RFID tags results in clock drift, which causes the bit duration to vary between RFID tags.
  • Collision avoidance protocols require the implementation of complex logic in the RFID tag to avoid collisions of the messages transmitted from the tags. Collision avoidance protocols also require a large amount of time that grows exponentially with the number of tags to read multiple RFID tags. Collision embracing MAC protocols allow multiple RFID tags to transmit simultaneously. Simultaneous transmission tends to reduce communication time in a dense deployment. But the performance of collision embracing protocols typically suffers due to the above described hardware heterogeneity issues, small message size, and low SNR of the RFID tags.
  • MAC protocols can also be categorized based on the underlying assumptions used to define the protocols.
  • One category of protocols assumes that there is no starting delay or clock drift in the RFID tags being interrogated.
  • the collision resistant RFID protocol known as "Buzz" fits into this category because it requires symbol level synchronization. Buzz first estimates the channels from the tags. The time requirement for channel estimation grows polynomially with the number of tags. Due to small message size, this channel estimation overhead can become large.
  • the tag reader estimates the number of RFID tags within the communication range and assigns a unique identity to each tag.
  • the reader estimates the channels from the RFID tags using a set of preambles that are generated from the identities assigned to the RFID tags.
  • RFID tags use their unique identities to determine in which slot to transmit their data. Because the reader knows the channels from the RFID tags and the set of RFID tags that are transmitting in any slot, the reader can decode the colliding messages from the RFID tags by combining the information over multiple slots.
  • FIG. 2 depicts a graph 6 illustrating the performance of the Buzz protocol with respect to decoding colliding messages.
  • the graph 6 includes a vertical axis 8 corresponding to the number of slots in which data can be transmitted, a horizontal axis 10 corresponding the number of colliding RFID tags, and a plot 12 of the number slots verses number of colliding RFID tags.
  • LFB Laissez-Faire Backscatter
  • BST Backscatter Spike Train
  • LFB uses an external crystal oscillator-based clock that consumes 10 ⁇ of power. This power consumption is an order of magnitude greater than the typical power consumption of an entire RFID tag, which is typically about 1.05 ⁇ .
  • typical RFID tags have integrated clocks made of resistors, capacitors, and transistors to reduce both the power consumption and the size of the tags.
  • the LFB protocol is described in more detail in "Laissez-Faire: Fully Asymmetric Backscatter Communication", Pan Hu et al., Proceedings of the ACM SIGCOMM Conference, London 2015, the disclosure of which is incorporated herein by reference in its entirety.
  • the BST protocol is described in more detail in "Leveraging Interleaved Signal Edges for Concurrent Backscatter", P. Zhang et al., Proceedings of the 1 st ACM Workshop on Hot Topics in Wireless, Maui 2014, the disclosure of which is incorporated herein by reference in its entirety.
  • BiGroup is a protocol that can decode collided messages even in the presence of clock drifts.
  • BiGroup relies on an In-phase/Quadrature-phase (I-Q) clustering method to decode data from the RFID tags that performs poorly with large numbers of tags. For example, if there are C RFID tags, and an RFID tag is transmitting either a 0 or a 1 (e.g., binary keying), there could be up to 2 C different clusters. Thus, for large numbers of RFID tags, there may not be enough data to distinguish all the clusters due to the small message size. In addition, because of the low SNR inherent with RFID tags, the distance between two clusters can be small. This may result in clusters being grouped into a single cluster by the clustering algorithm, thereby causing failure in data decoding. Typically, BiGroup is incapable of decoding messages from more than 3 colliding RFID tags.
  • a tag reader includes a processor and a memory coupled to the processor.
  • the memory contains program code configured to, when executed by the processor, cause the tag reader to transmit an excitation signal including a first frequency and a second frequency separated by a distance such that the first frequency provides a first communication channel and the second frequency provides a second communication channel in a response signal containing modulated versions of the first frequency and the second frequency.
  • the program code is further configured to cause the reader to receive the response signal and decode a message from a tag using the modulated versions of the first frequency and the second frequency.
  • the message is one of a plurality of colliding messages
  • the program code is further configured to cause the reader to filter and decode the colliding messages.
  • the message is one of a plurality of independent messages embedded in a received signal
  • the program code is configured to decode the message from the tag by sampling the received signal to generate respective sets of samples of modulated versions of the first frequency and the second frequency, defining an input matrix including a plurality of rows, each row containing one of the respective sets of samples, and using independent component analysis to determine a number of the independent messages based on at least a portion of the input matrix, and decode each of the independent messages from the sets of samples.
  • each of the modulated versions of the first frequency and the second frequency include an in-phase channel and a quadrature channel
  • the sets of samples include an in-phase set of samples of the in-phase channel and a quadrature set of samples of the quadrature channel for each of the first frequency and the second frequency.
  • the reader further comprises a plurality of receive antennas, and sampling the received signal to generate respective sets of samples of modulated versions of the first frequency and the second frequency includes sampling the received signal at each of the receive antennas, and generating the respective sets of samples for each of the receive antennas.
  • the program code is further configured to case the reader to apply a window to the input matrix that defines the portion of the input matrix on which the determined number of independent messages is based.
  • the input matrix includes a plurality of columns, and the window defines the portion of the input matrix by selecting a sub- set of the columns of the input matrix.
  • the program code is further configured to cause the reader to, in response to a failure to decode the message, adjust a size of the window.
  • the program code is further configured to cause the reader to separate the modulated versions of the first frequency and the second frequency from the response signal using a fast Fourier transform.
  • the reader further comprises a transmit antenna, a receive antenna, and the program code is further configured to cause the reader to cancel a direct signal received by the receive antenna from the transmit antenna.
  • a method of decoding a message from a tag comprises transmitting the excitation signal including the first frequency and the second frequency separated by the distance such that the first frequency provides the first communication channel and the second frequency provides the second communication channel in the response signal containing modulated versions of the first frequency and the second frequency.
  • the method further includes receiving the response signal and decoding the message from the tag using the modulated versions of the first frequency and the second frequency.
  • the message is one of the plurality of independent messages embedded in the received signal, and decoding the message includes sampling the received signal to generate respective sets of samples of modulated versions of the first frequency and the second frequency, defining the input matrix including the plurality of rows, each row containing one of the respective sets of samples, using independent component analysis to determine the number of the independent messages based on at least the portion of the input matrix, and decoding each of the independent messages from the sets of samples.
  • each of the modulated versions of the first frequency and the second frequency include the in-phase channel and the quadrature channel
  • the sets of samples include the in-phase set of samples of the in-phase channel and the quadrature set of samples of the quadrature channel for each of the first frequency and the second frequency.
  • sampling the received signal to generate respective sets of samples of modulated versions of the first frequency and the second frequency includes sampling the received signal at each of the plurality of receive antennas, and generating the respective sets of samples for each of the receive antennas.
  • the method further comprises applying the window to the input matrix that defines the portion of the input matrix on which the determined number of independent messages is based.
  • the input matrix includes the plurality of columns, and the window defines the portion of the input matrix by selecting the sub-set of the columns of the input matrix.
  • the method further comprises, in response to a failure to decode the message, adjusting the size of the window.
  • the method further comprises separating the modulated versions of the first frequency and the second frequency from the response signal using the fast Fourier transform.
  • a computer program product for decoding a message from an RFID tag.
  • the computer program product comprises a non- transitory computer-readable storage medium and program code stored on the non-transitory computer-readable storage medium.
  • the program code When executed by one or more processors of the tag reader, the program code causes the reader to transmit the excitation signal including the first frequency and the second frequency separated by the distance such that the first frequency provides the first communication channel and the second frequency provides the second communication channel in the response signal containing modulated versions of the first frequency and the second frequency.
  • FIG. 1 is an isometric view of a tag reader interrogating a plurality of tags attached to books in a bookshelf.
  • FIG. 2 is a graphical view illustrating a number of slots verses number of colliding tags for the Buzz protocol.
  • FIG. 3 is a graphical view illustrating the effects of hardware heterogeneity, orientation, distance, and receive antenna selection on the linear independence of channels through which the tags transmit data to the reader.
  • FIG. 4 is a graphical view illustrating the effects of electromagnetic coupling between tags and the distance between adjacent carrier frequencies on the linear independence of channels through which the tags transmit data to the tag reader.
  • FIG. 5 is a diagrammatic view of a tag reader transmitting an excitation signal to, and receiving a response signal from, a plurality of tags.
  • FIG. 6 is a diagrammatic view of a processing system that processes the response signal of FIG. 5.
  • FIG. 7 is a graphical view illustrating samples of messages transmitted from the tags of FIG. 5
  • FIG. 8 is a graphical view illustrating run times of the processing system of FIG. 6 using different filter types and sample window sizes.
  • FIG. 9 is a graphical view illustrating real and imaginary components of a sampled frequency channel of the processing system of FIG. 6.
  • FIG. 10 is a graphical view illustrating of the output of an Independent Component Analysis (ICA) module of the processing system of FIG. 6.
  • ICA Independent Component Analysis
  • FIG. 11 is a textual view of exemplary algorithm that may be executed by the ICA module to process samples received from a cancellation module of the processing system of FIG. 6.
  • FIG. 12 is a graphical view of the number of decoded tags for various numbers of colliding tags, sampling rates, data rates, decoding protocols, and number of receive antennas for the processing system of FIG. 6.
  • FIG. 13 is a graphical view of the number of tags that can be decoded.
  • FIG. 14 is a graphical view of the number of tags that can be decoded.
  • FIG. 15 is a graphical view of the decoding performance and mutual information among samples with different amounts of clock drift and starting delay for the processing system of FIG. 6.
  • FIG. 16 is a graphical view of the decoding performance and mutual information among samples with different amounts of clock drift and starting delay for the BiGroup protocol.
  • FIG. 17 is a graphical view of groupings of colliding messages showing clustering errors.
  • FIG. 18 is a diagrammatic view of a computer that may be used to implement the processing system of FIG. 6.
  • Embodiments of the invention may use this frequency independence and/or the independence of the messages from different tags to decode colliding messages. Diversity across different frequencies may thereby be exploited to simultaneously decode colliding messages transmitted from multiple RFID tags.
  • Decoding may be formulated as a statistical signal processing problem which is effectively solved using Independent Component Analysis (ICA).
  • ICA Independent Component Analysis
  • Challenges in using ICA addressed by embodiments of the invention may include collecting messages from the tags at different frequencies. Simple frequency hopping can require a large amount of time to collect the samples that grows linearly with the number of frequencies.
  • An excitation signal including multiple carrier frequencies and Fast Fourier Transform (FFT) based filtering of the received signals may be used to overcome this problem and enable collection of samples at multiple frequencies simultaneously.
  • FFT Fast Fourier Transform
  • Another challenge is that the direct application of ICA on the received samples may require a large amount of time.
  • embodiments of the invention may use a window-based algorithm.
  • the channel of an RFID tag may be also be independent across multiple antennas. Embodiments of the invention may exploit this spatial diversity to decode colliding messages from a large number of tags. For example, suppose a tag requires a bandwidth of B for message transmission and the available bandwidth is W. Because the data rate of the RFID tags is low (e.g., on the order of several kbps), the message bandwidth B is normally much less than the available bandwidth W.
  • the RFID tag sends a message to the reader by selectively reflecting or otherwise generating a modulated version of the excitation signal transmitted by the reader. Modulation of the excitation signal by the RFID tag may result (for amplitude modulation) in the modulated signal having a bandwidth of 2B .
  • the excitation signal consists of a single carrier frequency. It has been determined that dividing the available bandwidth W into W/2B portions and collecting data using an excitation signal containing W/2B different carrier frequencies provides significant advantages.
  • ICA may enable embodiments of the invention to decode colliding messages without any channel information by making one or more assumptions. For example, one assumption may be that messages generated by different tags should be statistically independent. This assumption may hold because the data transmitted by the RFID tags is normally random. Another assumption may be that the statistical independence among messages increases as the hardware heterogeneity among RFID tags increases. Thus, the channels through which the tag reader receives messages from different RFID tags can be assumed to be independent of each other across multiple carrier frequencies. These assumptions have been determined to be valid based on the experimental results presented below.
  • the samples collected from a sub-channel may be treated as complex numbers that can be divided into real and imaginary parts. In this way, a sub-channel can be converted into two sets of samples.
  • 250 different sets of samples may be obtained from 125 sub-channels.
  • Each one of the sets of samples may contain a different linear combination of the colliding messages from the RFID tags because the channels are independent across frequencies. Using 250 different linear combinations, it may be possible to decode messages from 250 colliding RFID tags.
  • the number of colliding RFID tags that can be decoded increases linearly with both the bandwidth and the number of receive antennas. This may be close to a theoretical maximum number of RFID tags that can be decoded with an available bandwidth W and K antennas. Because the degrees of freedom in the system is KW/B, the upper limit for decoding colliding RFID tags may also be
  • FIR Finite Impulse Response
  • IIR Infinite Impulse Response
  • the output of the FFT filter at different frequencies may be used as an input for the ICA module.
  • direct application of ICA on received samples may be inefficient for multiple reasons. For example, as messages become misaligned, the statistical independence among the message may decrease. This decrease in statistical independence may adversely affect the performance of ICA, and can therefore increase the runtime of the ICA module. However, the statistical independence within a subset of received samples defined by a window may be higher than the statistical independence of all the samples within the message. Embodiments of the invention may leverage this dependence between window size and statistical independence to effectively apply ICA on an optimized window of samples. Embodiments of the invention may also include methods for correcting errors in the output of the ICA module.
  • the disclosed protocol is scalable and can decode messages from increasing numbers of RFID tags as the available bandwidth W and the number of receive antennas ⁇ increases.
  • This protocol is believed to be the first protocol capable of decoding 37 colliding messages using a single receive antenna without synchronization across RFID tags or embedding highly accurate clocks in the RFID tags.
  • Experimental results show that embodiments of the invention implemented using USRP defined radios and trace-driven simulations can decode 12 times as many RFID tags as known protocols.
  • a statistical measure called the absolute correlation coefficient p may be used. For two different channel vectors C t and C;, the absolute correlation coefficient p may be determined using Equation 4:
  • the results described below are based on several assumptions. These assumptions may include an assumption that the value of the absolute correlation coefficient p is in the range of 0 to 0.5 most of the time, and that there is therefore a strong linear independence between communication channels.
  • the communication channel values of the RFID tags are measured from messages that are sent by the RFID tags, and are measured for different frequencies. Unless stated otherwise, the exemplary data described below is for adjacent frequencies 1 MHz apart and tag readers having two antennas (e.g., one for transmitting excitation signals to the RFID tags, and another for receiving response signals emitted by the RFID tags). The distance between the RFID tags and the antennas of the tag reader used to collect the data was set to 18 inches, which produced a round-trip path length between the tag reader and the RFID tags of 36 inches.
  • FIG. 3 depicts graphs 14- 17 including plots 22-29 of a Cumulative Distribution Function (CDF) verses absolute correlation coefficients p of the channel for hardware heterogeneity (graph 14), orientation (graph 15), distance (graph 16), and different receive antennas (graph 17).
  • CDF Cumulative Distribution Function
  • graph 14 illustrates the effect of hardware heterogeneity on the linear independence of the communication channels.
  • Graph 14 was generated by placing different RFID tags at the same position in front of the antennas of the tag reader, and determining the resulting values of the absolute correlation coefficient p for the same frequency across different tags (plot 22) and different frequencies from the same tag (plot 23).
  • the linear independence due to hardware heterogeneity is relatively low as only about 40% of the p values fall within the range [0, 0.5] .
  • Graph 15 illustrates the effect of RFID tag orientation on the linear independence of the communication channels.
  • Graph 15 was generated by placing a single RFID tag in front of the antennas of the tag reader. In this configuration, the RFID tag and the antennas of the tag reader were generally parallel to each other. The orientation of the RFID tag was changed by rotating the RFID tag and determining the p values at multiple angles, e.g., 0, 45, 90, 135 and 180 degrees. As can be seen from the plot 24, the p values for channels having the same frequency but different orientations is relatively low, as demonstrated by the CDF for p G [0,0.5] being approximately 66%.
  • Graph 16 illustrates linear independence due to distance.
  • an RFID tag was placed in front of the antennas of the tag reader randomly within a circle having a diameter of about 8 inches.
  • the values of the absolute correlation coefficient p for the same frequency at different distances show that p G [0,0.5] 78% of the time.
  • the cumulative distribution of the absolute correlation coefficient p for different frequencies at the same distance shows that the absolute correlation coefficient p G [0,0.5] 43% of the time. This may indicate that two RFID tags placed at different distances from the antennas of the tag reader will have communication channels with a relatively strong linear independence as compared to the communication channels of a single RFID tag having different frequencies.
  • Graph 17 illustrates the linear independence of channels using different receive antennas.
  • a single RFID tag was placed in front of two receive antennas and the channels characterized for each antenna.
  • the cumulative distribution of the absolute correlation coefficient p for channels having the same frequency and different antennas show that p G [0,0.5] 65% of the time.
  • the absolute correlation coefficient p G [0,0.5] 43% of the time for channels having different frequencies and the same antenna is shown.
  • FIG. 4 depicts graphs 38 and 44.
  • Graph 38 includes plots 40, 42 of the cumulative distribution function of absolute correlation coefficient p for the same frequency/different tag (plot 40) and different frequency/same tag (plot 42). Plots 40, 42 show the effect of
  • Graph 44 includes plots 46-49 of the cumulative distribution function of the absolute correlation coefficient p for multiple RFID tags across different frequencies in a dense deployment.
  • a dense deployment of RFID tags e.g., six RFID tags, was placed in front of the antennas of the tag reader within a rectangular box 12 inches wide and 18 inches long.
  • the communication channels between the RFID tags and the tag reader were then characterized for various carrier frequency spacing. Varying the separation between adjacent carrier frequencies allowed the independence of the communication channels to be measured across a wide range of frequency spacings.
  • the data represented by plot 46 was collected using a frequency spacing of 250 KHz, the data represented by plot 47 was collected using a frequency spacing of 500 KHz, the data represented by plot 48 was collected using a frequency spacing of 1 MHz, and the data represented by plot 49 was collected using a frequency spacing of 2 MHz.
  • the values of the absolute correlation coefficient p are between 0 and 0.5. This indicates that there is a strong linear independence among
  • the tag reader 58 may include a transmit antenna 60 that transmits an excitation signal 62 comprising one or more carrier frequencies 64.
  • the excitation signal may propagate toward each of a plurality of RFID tags 66 through an uplink channel 67 provided by the electromagnetic path between the transmit antenna 60 and the RFID tag 66.
  • the excitation signal 62 may also propagate toward a receive antenna 72 of tag reader 58 through a direct channel 73 provided by the electromagnetic path between the transmit antenna 60 and the receive antenna 72.
  • Each RFID tag 66 receiving the excitation signal 62 may respond by emitting a response signal 68 comprising one or more messages 70.
  • Each response signal 68 may propagate toward a receive antenna 72 of tag reader 58 through a downlink channel 69 provided by the electromagnetic path between the RFID tag 66 in question and the receive antenna 72.
  • Each message 70 of response signal 68 may comprise a corresponding carrier frequency 64 of the excitation signal 62 modulated with one or more symbols by an RFID tag 66.
  • the response signals 68 may be received by a receive antenna 72 of tag reader 58, and may be superimposed over each other as well as the direct signal received through the direct channel 73.
  • FIG. 6 depicts a received signal 74 being provided to a processing system 80 from the receive antenna 72.
  • the received signal 74 may comprise the direct signal and each response signal 68 received by the receive antenna 72.
  • the processing system 80 may include an FFT module 82 that processes the received signal 74 into frequency channels 90, a cancellation module 83 in communication with the FFT module 82, an ICA module 84 in communication with the cancellation module 83, and a decoding module 85 in communication with the ICA module 84.
  • the tag reader 58 may include a processor and memory storing program code that, when executed by the processor, implements the various modules and processes used by the tag reader 58 to generate the excitation signal 62 as well as receive and process the response signals 68.
  • the tag reader 58 may further include a transmitter (not shown) configured to transmit the excitation signal 62 and a front-end receiver (not shown) configured to receive the signal from the receive antenna 72 and provide the received signal 74 to the processing system 80.
  • the front-end receiver of tag reader 58 may, for example, amplify, filter, and/or down-convert the signals received by the receive antenna 72 into a baseband version of the received signal 74 suitable for processing by the FFT module 82.
  • the FFT module 82 may divide the received signal 74 into multiple frequency channels 90 and provide the frequency channels 90 to the cancellation module 83.
  • the carrier frequencies 64 contained in the received signal 74 may be stronger than the portions of the response signals 68 carrying the messages 70.
  • the cancellation module 83 may cancel these carrier frequencies 64, thereby removing them from the frequency channels 90 received from the FFT module 82.
  • the resulting filtered frequency channels 92 may be provided to the ICA module 84.
  • the ICA module 84 may process the filtered frequency channels 92 to separate the colliding messages 70 received from the RFID tags 66 and provide recovered messages 94 to the decoding module 85.
  • the decoding module 85 may in turn decode each recovered message 94 to extract the data contained therein.
  • the tag reader 58 may have a sampling rate that is significantly higher than the data rate of the RFID tags 66.
  • the sampling rate may enable the tag reader 58 to transform each bit of data in the messages 70 received from the RFID tags 66 into multiple samples at the tag reader 58. That is, the tag reader 58 may over-sample the received signal 74.
  • Embodiments of the invention may leverage the encoding used by commercial RFID tags (e.g., "FM0" encoding) to generate the messages 70. For example, in FM0 encoding, a T bit is mapped to either a ⁇ or '00' symbol, and a '0' bit is mapped to either a '01 ' or '10' symbol.
  • FM0 encoding ensures that there is a transition (change from 0 to 1 or from 1 to 0) at the start of each bit, and that there is an additional transition during transmission of a zero bit.
  • FM0 encoding may thereby improve the bit error rate in the decoded messages as compared to unencoded bits.
  • FIG. 7 depicts graphs 100 and 102 illustrating exemplary message sampling for a message encoded using FM0 encoding.
  • Graph 100 includes a plot 104 showing sampling of a message transmitted from one RFID tag, and graph 102 includes a plot 106 showing sampling of a message transmitted from another RFID tag. Only the first 2000 samples of each message are shown. The messages each have a different starting delay and clock drift, so the number of bits in the first 2000 samples are different for each message.
  • the tag reader may use QAM modulation to transmit an excitation signal including an in-phase channel (I-channel) and a quadrature channel (Q-channel).
  • the I and Q channels may include a summation of F frequencies, as shown by equations 5 and 6:
  • n is the sample number
  • T is the sampling interval
  • a k and fc are the amplitude and phase of the k th frequency, respectively.
  • the RFID tag may multiply or otherwise modulate the received excitation signal 62 with a signal generated internally.
  • the phase and amplitude of the excitation signal 62 may also modified as it travels to and from the RFID tag due to the various factors discussed above.
  • the received signals for the I and Q channels i? / [n] and R Q [n] are provided by: kW
  • each carrier frequency 64 may create two mirror images of bandwidth B around each carrier frequency 64 of excitation signal 62, as illustrated by the response signal 68 of FIG. 5.
  • the received samples from the I and Q channels may be combined to create complex samples R [n], which are provided by:
  • the response signal 68 also contains two copies of the message 70, one centered on each of the carrier frequencies 64.
  • Each copy of the message 70 transmitted from a single RFID tag 66 may contain the same data.
  • the different carrier frequencies 64 may add a measure of frequency diversity to the response signal 68.
  • FFT or other suitable frequency based filtering may be used to separate the copies of the message 70 after they are received by the tag reader 58.
  • the transmitter of the excitation signal 62 and the receiver of the response signal 68 may be provided by the same device using a common clock. Using a common clock signal eliminates any frequency offset between the transmitter and the receiver so that the carrier frequencies 64 are equally separated. Equal separation of the carrier frequencies 64 of excitation signal 62 may facilitate separation of the messages 70 by the FFT module 82.
  • the FFT may be performed on a window of F consecutive samples, referred to herein as the FFT window. It should also be noted that after FFT based filtering, the number of samples for a single carrier frequency 64 may be reduced by a factor of F.
  • the i th output of the FFT is f [i] for 1 ⁇ i ⁇ F, and may be provided by:
  • the FFT operation may be performed on samples R[0] and R [1] .
  • the output of the FFT can then be calculated using Equations 9 and 11 to produce:
  • the required runtime for applying the FFT on F samples is F X log(F) .
  • S samples in the received signal can be divided into SIF FFT windows, and the FFT is applied to each window, in which case the total runtime is 0 X (5 X log(F)).
  • the runtime for a single IIR filter is Ox(SxL), where L is the number of taps in the IIR filter.
  • the total runtime is 0 X (F X S X L).
  • FFT based filtering is faster than IIR based filtering by a factor of:
  • a further advantage of FFT based filtering is that the output has no instability.
  • FIG. 8 depicts graphs 108 and 110.
  • Graph 108 includes plots 112, 114 illustrating an asymptotic runtime comparison of FFT based filtering with FIR and IIR based filtering using 50,000 samples.
  • Plot 112 shows the improvement ratio verses number of frequencies for a FIR filter
  • plot 114 shows the improvement ratio verses number of frequencies for an IIR filter.
  • the data shown is for an elliptic IIR filter with an optimal number of taps, e.g., 7 taps.
  • Graph 110 includes a plot 116 illustrating a performance comparison for different initial window sizes w_init for 40 colliding RFID tags. As can be seen from plot 116, as the initial window size w_init increases, the number of decoded RFID tags using ICA decreases, falling off rapidly for w_init > 4,000. This decrease in the number of decoded RFID tags may be due to poor statistical independence.
  • the received signal 74 may be a sum of the direct signal and the response signals 74 received by the receive antenna 72.
  • the direct signal received by the receive antenna 72 may include the portion of excitation signal 62 received directly from the transmit antenna 60 as well as reflections of the excitation signal 62 that are reflected from passive objects in the
  • the response signals 68 received from the RFID tags 66 are typically much weaker (e.g., 15-20 dB) than the direct signal.
  • Conventional RFID communication uses an excitation signal 62 that consists of a single carrier frequency and a narrow receive sampling bandwidth, e.g., 1 MHz or less. This may allow the communication channel for the direct signal to be modeled as a single-tap channel, which can be estimated by taking an average of the received samples when no RFID tags 66 are emitting response signals 68.
  • embodiments of the invention may use a receive sampling bandwidth that is much larger than in a single frequency system. This may allow the direct signal channel to be modeled as a multi-tap channel rather than a single tap channel. A multi-tap channel can be estimated by sending a known preamble from the transmit antenna 60. However, this may require additional
  • Embodiments of the invention may avoid using a preamble by estimating the direct signal channel after FFT based filtering and canceling out the direct signal.
  • the output signal after application of the FFT may be segmented into multiple frequency channels 90 each at a specific frequency.
  • Each frequency channel 90 may contain a single carrier frequency 64 and have smaller bandwidth than the received signal 74. This may allow the communication channel corresponding to the frequency of each frequency channel 90 to be modeled as a single-tap channel, which may be estimated by taking the average of received samples when the RFID tags 66 are not transmitting.
  • FIG. 9 depicts a graph 118 including a plot 120 of the real part of an exemplary frequency channel 90, a graph 122 including a plot 124 of the imaginary part of the frequency channel 90 of graph 120, a graph 126 including a plot 128 of the real part of another frequency channel 90 at a different frequency than the frequency channel 90 of graphs 118, 122, and a graph 130 including a plot 132 of the imaginary part of the frequency channel 90 of graph 126.
  • These combinations may include response signals 68 from multiple (e.g., two) RFID tags 66 each of which has been excited by an excitation signal 62 including the two carrier frequencies 64.
  • the received samples and the outputs of the FFT module 82 may be complex numbers that can be divided into real and imaginary parts.
  • the ICA module 84 may divide the samples corresponding to each carrier frequency into two different sets of samples to define IF sets of samples from F frequencies.
  • the received signal can be modeled as:
  • Y is a 2F X S matrix of received samples with S being the number of received samples.
  • Each row of the matrix Y contains the colliding samples from P RFID tags received at a specific frequency
  • A is the 2F X P channel matrix
  • X is the P X S matrix where the j th row contains the samples received from the j th RFID tag
  • N is the 2F X S noise matrix.
  • ICA estimates the number of independent messages in the matrix Y using a Principle Component Analysis (PCA) algorithm.
  • PCA Principle Component Analysis
  • ICA does not require the number of colliding RFID tags as an input, which may facilitate operation in an environment where the number of responding RFIDs is not known prior to decoding of the received signal 74.
  • the matrix W is closely related to matrix A, and in general, WA « I x .
  • the number of frequencies F required to decode P RFID tags may be determined as
  • the rank of a matrix cannot be greater than the number of rows or columns of the matrix, which mean P ⁇ 2F.
  • at least P/2 carrier frequencies 64 and/or frequency channels 90 may be required to decode P RFID tags 66.
  • Equation 14 can be rewritten as
  • the algorithm used to determine the colliding messages may identify the colliding messages one by one using a method called projection pursuit. To decode the message from a single RFID tag, the required runtime using this method is / X R, where / is the number of iterations and R is runtime required for a single iteration. / may depend on the rate of
  • the runtime R required for each iteration may be proportional to the size of the input matrix (Y), i.e., 0(SF) .
  • the total runtime required to decode P RFID tags is 0(SF/P).
  • the ICA module 84 may also have four different sets of samples, two of which contain data from the RFID tags, and the rest of which are noise.
  • FIG. 10 depicts a graph 134 including a plot 136 of one output of the ICA module 84 and a graph 138 including a plot 140 of another output of the ICA module 84.
  • Plot 136 matches the data illustrated by plot 102 of FIG. 7 for a staring delay of 1 ⁇ and a bit duration of 4.12 ⁇ 8.
  • Plot 140 matches the data illustrated by plot 106 of FIG. 7 for a staring delay of 9 ⁇ 8 and a bit duration of 5.3 ⁇ 8.
  • Plots 136, 140 illustrate the separated signals after processing by the ICA module 84. Reducing the sampling rate by a factor of two reduces the number of samples by a factor of two in the messages.
  • the starting delay and the bit duration (which is determined by the clock drift) of the output signals are the same as in the actual signals.
  • the bit durations of the messages have been reduced by a factor of two because there are two carrier frequencies 64 in the excitation signal 62 and the size of the FFT window is two.
  • the output of the ICA module 84 may include ambiguities that can flip the sign of the received samples.
  • RFID tags using FM0 encoding may encode the data by changing the power level of the received samples. This encoding may be independent of the sign of the received samples because the sign of a sample does not change the power of that sample.
  • the data message can be correctly decoded if all the transitions are correctly detected.
  • the positions of all the transitions in the output samples are determined by comparing the power levels of consecutive samples.
  • due to noise there could be some erroneous transitions.
  • bit duration may be used to filter out the erroneous transitions. For an -bit message, there may be at most 2M transitions.
  • bit duration is b
  • bit duration may be different for different RFID tags due to clock drift.
  • a preamble of three consecutive Ts may be appended before each message. This preamble may be used to correctly estimate the bit duration b.
  • the start and end times of the messages are not necessarily all the same.
  • the samples received from an RFID tag Before the start of a message and after the end of a message, the samples received from an RFID tag may be all zeros. But when received samples from two RFID tags are zero for a long period of time, statistical independence among the samples from those two RFID tags may decrease. This reduced statistical independence may cause certain problems. For example, the performance of ICA may be adversely affected by low statistical independence. In particular, ICA may fail to properly separate out the samples from different RFID tags, leading to failures in message decoding. In addition, the number of iterations for the projection pursuit algorithm used by ICA may increase by a large number (e.g., 10 to 15 times), resulting in a substantial increase in the processing time by the ICA module 84.
  • ICA may be used on a window of received samples.
  • the initial window size may be determined by the number of samples in a message without any distortions due to hardware heterogeneity. For a data rate D and a message length of M bits, the number of samples within a message may be MIDT in the absence of clock drift and starting delay where T is the sampling interval. Due to FFT based filtering, the number of samples may be reduced by a factor of F, which is the number of frequencies in the excitation signal 62.
  • the initial window size w_init is provided by:
  • FIG. 11 depicts an exemplary algorithm that describes processing of the received samples by the ICA module 84 in an embodiment of the invention.
  • the depicted algorithm controls the size of the window of samples.
  • the ICA technique may be applied on an input matrix Y of dimension 2F X S.
  • ICA is applied on the samples contained within the window of size w, i.e., a sub-set of columns comprising the first w columns of the matrix Y (see line 7 of the algorithm).
  • the output of the ICA module is processed as described above (see line 10 of the algorithm). If a message is decoded, the actual samples corresponding to that message may be created from the estimated channel value, starting delay, and clock drift. These reconstructed samples may be canceled out from the current window (see line 12 of the algorithm).
  • the ICA module 84 may determine if the average of the sample values in a column is smaller than the noise level. If the average is smaller than the noise level, the starting index (s) of the window may be increased to exclude those columns (see lines 17-21 of the algorithm). When no messages are decoded from the current window, it may be because the current window does not contain a full message from any RFID tags 66. In this case, it may be better to increase the window size to include all the samples from a message (see lines 23-28 of the algorithm).
  • WISP Identification and Sensing Platform 5 programmable RFID tags, which operate within a bandwidth of 26 MHz (902-928 MHz). By way of comparison, commercial RFID tags typically operate within a bandwidth of 100 MHz (860-960 MHz).
  • FIG. 12 depicts bar graphs 142-145 that illustrate the performance of a prototype system for different w init values.
  • Bar graph 142 includes a vertical axis 150 corresponding to the number of decoded RFID tags, and a horizontal axis 152 corresponding the number of colliding RFID tags.
  • Bars 154-158 show the number of RFID tags decoded by the prototype system using an excitation signal 62 having four carrier frequencies 64, a data rate of 100 Kbps, and a sampling rate of 20 Million Samples Per Second (MSPS).
  • MSPS Million Samples Per Second
  • Bar graph 143 includes a vertical axis 170 corresponding to the number of decoded RFID tags, and a horizontal axis 172 corresponding the sampling rate. Bars 174-178 show the number of RFID tags decoded by the prototype system using an excitation signal having four carrier frequencies, six colliding RFID tags, a data rate of 100 Kbps, and a sampling rates of 1 MSPS, 2 MSPS, 4 MSPS, 8 MSPS, and 20 MSPS, respectively. BiGroup was unable to decode any messages with this number of RFID tags, so no results for BiGroup are included in bar graph 143. At sampling rates below 4 MSPS, the number of samples in the colliding messages were too small for the ICA module to decode data from six RFID tags.
  • Bar graph 144 includes a vertical axis 184 corresponding to the number of decoded RFID tags, and a horizontal axis 186 corresponding the data rate. Bars 188- 190 show the number of RFID tags decoded by the prototype system using an excitation signal having four carrier frequencies, six colliding RFID tags, and a sampling rate of 20 MSPS at data rates of 100 kbps, 200 kbps, and 350 kbps, respectively.
  • the performance of the prototype system slowly decreases. This decrease in performance may be attributed to at least two factors. First, as the data rate increases, the number of samples of each colliding message decreases due to the shorter duration of the messages. Second, as the data rate increases, the bit duration decreases, but the number of samples during a single transition remains unchanged. Thus, the total number of samples during transitions increases in the colliding messages. The samples collected during transitions may have adversely affected the performance of the prototype system.
  • Bar graph 145 includes a vertical axis 194 corresponding to the number of decoded RFID tags, and a horizontal axis 196 corresponding the number of colliding tags. Bars 198-201 show the number of RFID tags decoded using a single receive antenna, and bars 206-211 show the number of RFID tags decoded using two receive antennas.
  • the excitation signal had only one carrier frequency, there were six colliding RFID tags, and the data rate was 100 MSPS.
  • the channel values collected during the experiments were used to model the communication channels between the RFID tags and the reader.
  • the SNRs of the messages collected during experiments were used to model the SNRs of the received messages.
  • the starting delays of different messages were measured during the experiments, and the measured values used to model the starting delay in the simulation.
  • the clock drifts of the messages were measured during experiments, and the measured values used to model the clock drift in the simulation.
  • FIG. 13 depicts graphs 218 and 220 illustrating a predicted number of RFID tags that can be decoded, and a corresponding bit error rate for the prototype system, respectively, as the number of RFID tags increases.
  • Graph 218 includes a vertical axis 222 corresponding to the number of decoded RFID tags, a horizontal axis 224 corresponding to the number of colliding RFID tags, and a plot 226.
  • Plot 226 shows the number of decoded RFID tags verses the number of RFID tags being interrogated by the prototype tag reader.
  • Graph 220 includes a vertical axis 228 corresponding to a bit error rate for the output of the decoding module 85, a horizontal axis 230 corresponding to the number of colliding RFID tags, and a plot 232.
  • Plot 232 shows the bit error rate verses the number of RFID tags being interrogated by the prototype tag reader. As the number of RFID tags was increased from 5 to 70, the prototype system was predicted to be able to decode colliding messages from up to 40 RFID tags. This is indicated by the plot 226, which shows the performance of the system improving in a nearly linear fashion as the number of RFID tags is increased from 5 to 40. Beyond 40 RFID tags, the performance of the system decreases. Plot 232 shows a corresponding increase in the bit error rate beyond 40 RFID tags.
  • FIG. 14 depicts graphs 234 and 236 showing the predicted decoding and bit error rate performance, respectively, of the prototype system as the number of carrier frequencies in the excitation signal is increased.
  • Graph 234 includes a vertical axis 238 corresponding to the number of decoded RFID tags, and a horizontal axis 240 corresponding to the number of colliding RFID tags.
  • Plots 242-246 show the number of decoded RFID tags verses the number of RFID tags being interrogated by the prototype tag reader for an excitation signal having five carrier frequencies, 10 carrier frequencies, 15 carrier frequencies, 20 carrier frequencies, and 25 carrier frequencies, respectively.
  • Graph 236 includes a vertical axis 252 corresponding to the bit error rate for the output of the decoding module and a horizontal axis 240 corresponding to the number of colliding RFID tags.
  • Plots 256-260 show the bit error rate verses number of RFID tags being interrogated by the prototype tag reader for an excitation signal having five carrier frequencies, 10 carrier frequencies, 15 carrier frequencies, 20 carrier frequencies, and 25 carrier frequencies, respectively.
  • FIG. 15 depicts graphs 266-269 illustrating the predicted performance of the prototype system for different amounts of clock drift and starting delays.
  • Graph 266 includes a vertical axis 274 corresponding to the number of decoded RFID tags, and a horizontal axis 276 corresponding the percentage of clock drift.
  • Plot 278 shows the number of RFID tags decoded verses clock rate error for the prototype system for zero starting delay.
  • Graph 267 includes a vertical axis 280 corresponding to the number of decoded RFID tags, and a horizontal axis 282 corresponding the starting delay in ⁇ 8.
  • Plots 284 and 286 show the number of RFID tags decoded verses starting delay for the prototype system and for a system using the LFB protocol, respectively, without any clock drift.
  • Graph 268 includes a vertical axis 288 corresponding to the amount of mutual information among the window of samples, and a horizontal axis 290 corresponding the percentage of clock drift.
  • Plot 292 shows the amount of mutual information between random samples verses clock rate error for the prototype system.
  • Graph 269 includes a vertical axis 294 corresponding to the amount of mutual information among the window of samples, and a horizontal axis 296 corresponding the starting delay.
  • Plot 298 shows the amount of mutual information between random samples verses staring delay.
  • Clock drift may cause the messages from the RFID tags to be misaligned even if all the RFID tags begin generating messages at the same time. This misalignment may increase the statistical independence among the samples received from the RFID tags.
  • the statistical independence among the samples may be characterized by the amount of mutual information between the samples.
  • the effect on the number of decoded RFID tags verses clock drift percentage is shown by graph 266. When samples from two RFID tags are completely independent, the mutual information between them is zero. When the mutual information is less than 0.0645, and the clock drift is greater than 2.5%, the prototype system is predicted to decode 37 out of 40 RFID tags. The performance of the prototype system is predicted to saturate for mutual information levels less than 0.0645.
  • the maximum number of decoded RFID tags indicated by graph 266 is consistent with the results shown in graph 218 of FIG. 13.
  • Graph 267 shows the predicted performance of the prototype system and a system using LFB when there is no clock drift, but the starting delay is different for different RFID tags.
  • systems using LFB are predicted to outperform the prototype system.
  • the performance of the prototype system is predicted to outperform that of systems using LFB.
  • the prototype system is further predicted to decode messages from as many as 37 RFID tags on average out of 40 RFID tags as the starting delay increases to 20 ⁇ 8. This improvement in the performance of the prototype system as the starting delay increases may be a result of the increased statistical independence among the samples received from different RFID tags.
  • the ICA based prototype system is predicted to achieve the best performance when the value of mutual information is less than 0.065.
  • LFB first determines the transitions in the colliding messages, and assumes that there is no clock drift in the RFID tags. Without clock drift, the bit durations of all RFID tags would be the same. The absence of clock drift is therefore expected to facilitate detection of all the transitions that belong to a single RFID tag. Once the transitions from a single RFID tag are separated, LFB uses these transitions to decode the message sent by the RFID tag.
  • Graph 267 of FIG. 15 shows the performance of systems using LFB without clock drift in the RFID tags.
  • the performance of the LFB based system improves.
  • the performance of the LFB based system degrades sharply when the starting delay is more than the bit duration of 10 ⁇ 8. This may be due to the starting delay of an RFID tag exceeding the duration of one bit making it difficult for LFB systems to detect exactly when the message transmission from the RFID tag starts.
  • BiGroup relies heavily on I-Q clustering which is susceptible to poor SNR and channel values of the RFID tags.
  • I-Q clustering which is susceptible to poor SNR and channel values of the RFID tags.
  • C RFID tags whose transmissions are colliding, there are 2 C clusters in the I-Q domain.
  • BiGroup first clusters I-Q domain data to identify the number of colliding RFID tags.
  • BiGroup also detects all the transitions in the colliding messages. Then, for every RFID tag, BiGroup determines the transitions associated with that RFID tag by choosing a set of transitions that fit a linear regression model.
  • I-Q clusters are partitioned into two groups using the transitions associated with an RFID tag. One of the groups corresponds to the state when the RFID tag is transmitting bit ' ⁇ ', and the other group corresponds to the state when the RFID tag is transmitting bit ⁇ ' .
  • FIG. 16 depicts graphs 300-303 illustrating the predicted performance of a BiGroup based system for different numbers of colliding RFID tags, and for different starting delays, clock drift, and message size with three colliding RFID tags.
  • the BiGroup based system included RFID tags having a data rate of 100 kbps, a receiver sampling rate of 1 MHz, a clock drift as high as 20%, and a message length of 200 bits.
  • Graph 300 includes a vertical axis 308 corresponding to the number of decoded RFID tags, a horizontal axis 310 corresponding the number of colliding tags, and a plot 312 of the number of decoded RFID tags verses number of colliding RFID tags.
  • Graph 301 includes a vertical axis 314 corresponding to the number of decoded RFID tags, a horizontal axis 316 corresponding the starting delay, and a plot 318 showing the number of decoded RFID tags verses starting delay with three colliding RFID tags.
  • Graph 302 includes a vertical axis 320 corresponding to the number of decoded RFID tags, a horizontal axis 322 corresponding clock drift, and a plot 324 showing the number of decoded RFID tags verses clock drift with three colliding RFID tags.
  • Graph 303 includes a vertical axis 326 corresponding to the number of decoded RFID tags, a horizontal axis 328 corresponding to the message size in bits, and a plot 330 of the number of decoded RFID tags verses message size with three colliding RFID tags.
  • the data used to generate the graphs 300-303 shows an average number of RFID tags decoded by BiGroup over many experiments. As can be seen from plot 312, the BiGroup based reader was unable to decode colliding messages when there were more than three colliding RFID tags.
  • Graph 301 shows the average number of RFID tags decoded by BiGroup for different starting delays but zero clock drift. As the starting delay increases, BiGroup can decode more RFID tags because it is easier to detect the transitions from each RFID tag.
  • Graph 302 shows the performance of BiGroup for different amounts of clock drift and zero starting delay. The performance degrades as the clock drift increases because with a large value for clock drift, it is difficult to detect the transitions from an RFID tag using linear regression.
  • Graph 303 shows the performance of BiGroup for different message sizes. In this case, maximum starting delay is 10 ⁇ 8 and maximum clock drift is 20%. The performance degrades as the message size increases because it is increasingly difficult to detect the transitions associated with a single RFID tag using linear regression as the message size grows larger.
  • FIG. 17 depicts a scatter graph 340 illustrating exemplary clustering errors that may be responsible for the performance of the BiGroup protocol depicted by graphs 300-303.
  • the BiGroup protocol relies on an I-Q clustering method to decode collided messages in the presence of clock drifts. Based on graphs 300-303, it may be inferred that when there are more than three colliding RFID tags, there are significant clustering errors, and these clustering errors lead to a high BER.
  • the scatter graph 340 includes a vertical axis 342 corresponding to the Quadrature value of a received message, and a horizontal axis 344 corresponding to the In-Phase value of the received message.
  • the prototype system can decode colliding messages from a large number of RFID tags even when the transmitted messages are affected by the hardware heterogeneity of the RFID tags. Due to similarities of RFID communication with other modalities of backscatter
  • RFID tags harvest ambient RF signal from the excitation signal or other ambient electromagnetic energy (e.g., television, cellular, and Wi-Fi transmissions), it is envisioned that embodiments of the invention may also be used with these types of RFID tags. It is further contemplated that embodiments of the invention may be used with other ambient electromagnetic energy (e.g., television, cellular, and Wi-Fi transmissions).
  • ambient electromagnetic energy e.g., television, cellular, and Wi-Fi transmissions
  • Embodiments of the invention thus provide a novel solution for decoding colliding messages in environments having high densities of tags, such as RFID tags.
  • inventions described above, or portions thereof may be implemented using one or more computer devices or systems, such as exemplary computer 500.
  • the computer 500 may include a processor 502, a memory 504, an input/output (I/O) interface 506, and a Human Machine Interface (HMI) 508.
  • the computer 500 may also be operatively coupled to one or more external resources 510 via the network 512 and/or I/O interface 506.
  • External resources may include, but are not limited to, servers, databases, mass storage devices, peripheral devices, cloud-based network services, or any other resource that may be used by the computer 500.
  • the processor 502 may include one or more devices selected from microprocessors, micro-controllers, digital signal processors, microcomputers, central processing units, field programmable gate arrays, programmable logic devices, state machines, logic circuits, analog circuits, digital circuits, or any other devices that manipulate signals (analog or digital) based on operational instructions that are stored in memory 504.
  • Memory 504 may include a single memory device or a plurality of memory devices including, but not limited to, read-only memory (ROM), random access memory (RAM), volatile memory, non-volatile memory, static random access memory (SRAM), dynamic random access memory (DRAM), flash memory, cache memory, and/or data storage devices such as a hard drive, optical drive, tape drive, volatile or non-volatile solid state device, or any other device capable of storing data.
  • ROM read-only memory
  • RAM random access memory
  • volatile memory volatile memory
  • non-volatile memory volatile memory
  • SRAM static random access memory
  • DRAM dynamic random access memory
  • flash memory cache memory
  • data storage devices such as a hard drive, optical drive, tape drive, volatile or non-volatile solid state device, or any other device capable of storing data.
  • the processor 502 may operate under the control of an operating system 514 that resides in memory 504.
  • the operating system 514 may manage computer resources so that computer program code embodied as one or more computer software applications, such as an application 516 residing in memory 504, may have instructions executed by the processor 502.
  • the processor 502 may execute the application 516 directly, in which case the operating system 514 may be omitted.
  • One or more data structures 518 may also reside in memory 504, and may be used by the processor 502, operating system 514, or application 516 to store or manipulate data.
  • the I/O interface 506 may provide a machine interface that operatively couples the processor 502 to other devices and systems, such as the external resource 510 or the network 512.
  • the application 516 may thereby work cooperatively with the external resource 510 or network 512 by communicating via the I/O interface 506 to provide the various features, functions, applications, processes, or modules comprising embodiments of the invention.
  • the application 516 may also have program code that is executed by one or more external resources 510, or otherwise rely on functions or signals provided by other system or network components external to the computer 500. Indeed, given the nearly endless hardware and software
  • embodiments of the invention may include applications that are located externally to the computer 500, distributed among multiple computers or other external resources 510, or provided by computing resources (hardware and software) that are provided as a service over the network 512, such as a cloud computing service.
  • computing resources hardware and software
  • the HMI 508 may be operatively coupled to the processor 502 of computer 500 in a known manner to allow a user to interact directly with the computer 500.
  • the HMI 508 may include video or alphanumeric displays, a touch screen, a speaker, and any other suitable audio and visual indicators capable of providing data to the user.
  • the HMI 508 may also include input devices and controls such as an alphanumeric keyboard, a pointing device, keypads, pushbuttons, control knobs, microphones, etc., capable of accepting commands or input from the user and transmitting the entered input to the processor 502.
  • a database 520 may reside in memory 504, and may be used to collect and organize data used by the various systems and modules described herein.
  • the database 520 may include data and supporting data structures that store and organize the data.
  • the database 520 may be arranged with any database organization or structure including, but not limited to, a relational database, a hierarchical database, a network database, or combinations thereof.
  • a database management system in the form of a computer software application executing as instructions on the processor 502 may be used to access the information or data stored in records of the database 520 in response to a query, which may be dynamically determined and executed by the operating system 514, other applications 516, or one or more modules.
  • routines executed to implement the embodiments of the invention may be referred to herein as "computer program code,” or simply "program code.”
  • Program code typically comprises computer-readable instructions that are resident at various times in various memory and storage devices in a computer and that, when read and executed by one or more processors in a computer, cause that computer to perform the operations necessary to execute operations and/or elements embodying the various aspects of the embodiments of the invention.
  • Computer- readable program instructions for carrying out operations of the embodiments of the invention may be, for example, assembly language or either source code or object code written in any combination of one or more programming languages.
  • the program code embodied in any of the applications/modules described herein is capable of being individually or collectively distributed as a program product in a variety of different forms.
  • the program code may be distributed using a computer-readable storage medium having computer-readable program instructions thereon for causing a processor to carry out aspects of the embodiments of the invention.
  • Computer-readable storage media which is inherently non-transitory, may include volatile and non-volatile, and removable and non-removable tangible media implemented in any method or technology for storage of data, such as computer-readable instructions, data structures, program modules, or other data.
  • Computer-readable storage media may further include RAM, ROM, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other solid state memory technology, portable compact disc read-only memory (CD-ROM), or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired data and which can be read by a computer.
  • a computer- readable storage medium should not be construed as transitory signals per se (e.g., radio waves or other propagating electromagnetic waves, electromagnetic waves propagating through a transmission media such as a waveguide, or electrical signals transmitted through a wire).
  • Computer-readable program instructions may be downloaded to a computer, another type of programmable data processing apparatus, or another device from a computer-readable storage medium or to an external computer or external storage device via a network.
  • Computer-readable program instructions stored in a computer-readable medium may be used to direct a computer, other types of programmable data processing apparatuses, or other devices to function in a particular manner, such that the instructions stored in the computer- readable medium produce an article of manufacture including instructions that implement the functions, acts, and/or operations specified in the flow-charts, sequence diagrams, and/or block diagrams.
  • the computer program instructions may be provided to one or more processors of a general purpose computer, a special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the one or more processors, cause a series of computations to be performed to implement the functions, acts, and/or operations specified in the flow-charts, sequence diagrams, and/or block diagrams.
  • the functions, acts, and/or operations specified in the flow-charts, sequence diagrams, and/or block diagrams may be re-ordered, processed serially, and/or processed concurrently consistent with embodiments of the invention.
  • any of the flow -charts, sequence diagrams, and/or block diagrams may include more or fewer blocks than those illustrated consistent with embodiments of the invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Toxicology (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Electromagnetism (AREA)
  • General Health & Medical Sciences (AREA)
  • Near-Field Transmission Systems (AREA)

Abstract

L'invention concerne des systèmes, des procédés et des produits programmes d'ordinateur pour lire des étiquettes électroniques, telles que des étiquettes RFID (66). Un lecteur d'étiquette (58) transmet un signal d'excitation (62) ayant une pluralité de fréquences porteuses (64). Chaque étiquette (66) réfléchit ou utilise autrement le signal d'excitation (62) pour émettre un signal de réponse (68) ayant un message modulé sur chacune des fréquences porteuses (64). Le lecteur d'étiquette (58) reçoit les signaux de réponse (68) et filtre le signal reçu (74) en une pluralité de canaux de fréquence (90) correspondant chacun à l'une des fréquences porteuses (64). Le lecteur d'étiquette (58) annule la fréquence porteuse (64) dans le signal occupant chaque canal de fréquence (90). Les signaux restants peuvent être traités à l'aide d'une analyse de composantes indépendantes pour récupérer des messages en collision (70). Les messages récupérés (94) peuvent ensuite être décodés pour extraire les données qu'il contient.
PCT/US2018/058167 2017-10-30 2018-10-30 Communication rfid échelonnable utilisant une excitation multifréquence WO2019089557A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201762578682P 2017-10-30 2017-10-30
US62/578,682 2017-10-30

Publications (1)

Publication Number Publication Date
WO2019089557A1 true WO2019089557A1 (fr) 2019-05-09

Family

ID=64457090

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2018/058167 WO2019089557A1 (fr) 2017-10-30 2018-10-30 Communication rfid échelonnable utilisant une excitation multifréquence

Country Status (1)

Country Link
WO (1) WO2019089557A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113810069A (zh) * 2020-06-16 2021-12-17 华为技术有限公司 用于传输信号的通信装置及信号传输方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2001050407A1 (fr) * 2000-01-06 2001-07-12 Samsys Incorporated Systeme pour etiquettes rfid multistandards
EP1840789A1 (fr) * 2006-03-31 2007-10-03 Assa Abloy Identification Technology Group AB Circuit de génération un signal de détection pour un lecteur RFID
KR100867718B1 (ko) * 2007-07-13 2008-11-10 부산대학교 산학협력단 항만 물류를 위한 확장성을 가지는 이중 주파수 대역능동형 ｒｆｉｄ 시스템
US7741956B1 (en) * 1996-11-29 2010-06-22 X-Cyte, Inc. Dual mode transmitter-receiver and decoder for RF transponder tags

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7741956B1 (en) * 1996-11-29 2010-06-22 X-Cyte, Inc. Dual mode transmitter-receiver and decoder for RF transponder tags
WO2001050407A1 (fr) * 2000-01-06 2001-07-12 Samsys Incorporated Systeme pour etiquettes rfid multistandards
EP1840789A1 (fr) * 2006-03-31 2007-10-03 Assa Abloy Identification Technology Group AB Circuit de génération un signal de détection pour un lecteur RFID
KR100867718B1 (ko) * 2007-07-13 2008-11-10 부산대학교 산학협력단 항만 물류를 위한 확장성을 가지는 이중 주파수 대역능동형 ｒｆｉｄ 시스템

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
P. ZHANG ET AL.: "Proceedings of the 1st ACM Workshop on Hot Topics in Wireless", 2014, article "Leveraging Interleaved Signal Edges for Concurrent Backscatter"
PAN HU ET AL.: "Laissez-Faire: Fully Asymmetric Backscatter Communication", PROCEEDINGS OF THE ACM SIGCOMM CONFERENCE, 2015
WANG ET AL.: "Efficient and Reliable Low-Power Backscatter Networks", PROCEEDINGS OF THE ACM SIGCOMM CONFERENCE, 2012, pages 61 - 72, XP058007988, DOI: doi:10.1145/2342356.2342364

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113810069A (zh) * 2020-06-16 2021-12-17 华为技术有限公司 用于传输信号的通信装置及信号传输方法
CN113810069B (zh) * 2020-06-16 2023-10-20 华为技术有限公司 用于传输信号的通信装置及信号传输方法

Similar Documents

Publication Publication Date Title
KR100752946B1 (ko) 통신 시스템에서 소스 디바이스 동기화를 위한 방법 및장치
Schepker et al. Compressive sensing multi-user detection with block-wise orthogonal least squares
Fyhn et al. Multipacket reception of passive UHF RFID tags: A communication theoretic approach
US20070025475A1 (en) Method and apparatus for data signal processing in wireless RFID systems
US8340222B2 (en) Parameter and scattered pilot based symbol timing recovery
CN101273563A (zh) 用于无线射频识别（rfid）系统中前导同步的方法和装置
EP3091477B1 (fr) Lecteur rfid et procédé de réglage d'une longueur de trame d'un système rfid comprenant un lecteur rfid
WO2007087041A2 (fr) Procede et appareil pour traitement du signal dans des recepteurs d'identification rfid
CN102545926A (zh) 识别噪声源的类型的通信系统
KR20100020891A (ko) 광대역 무선통신 시스템에서 고속 피드백 채널을 이용한 정보 송수신 장치 및 방법
JP6272574B2 (ja) 通信チャネルを介して受信されたデータブロックを復号するための方法および受信機
WO2019089557A1 (fr) Communication rfid échelonnable utilisant une excitation multifréquence
Liu et al. Comparison of reliability, delay and complexity for standalone cognitive radio spectrum sensing schemes
Sharma et al. Spectral efficient compressive transmission framework for wireless communication systems
KR102067050B1 (ko) 스파스 코드 기반 엠비언트 백스캐터 시스템에서 액세스 포인트의 통신 방법 및 그 시스템
US10312977B2 (en) Complexity reduction for receiver decoding
EP3113399B1 (fr) Détecteur de bit de trames dans des communications en champ proche
Seol et al. Collision-resilient multi-state query tree protocol for fast RFID tag identification
Das et al. ADS: Accurate decoding of RFID tags at scale
US10257004B2 (en) Inter-block interference suppression using a null guard interval
Ji et al. Compressed sensing based multi-user detection with modified sphere detection in machine-to-machine communications
Duan et al. Channel estimation for recovery of UHF RFID tag collision on physical layer
Li et al. Signal coding in physical layer separation for RFID tag collision
Yue et al. A hybrid CDMA-IDMA anti-collision protocol for dense RFID system
EP3437200A1 (fr) Transmission à deux phases pour la communication de type machine

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18808157

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18808157

Country of ref document: EP

Kind code of ref document: A1