CN113630209B

CN113630209B - WiFi-to-Bluetooth Low Energy (BLE) cross-technology communication method based on narrowband decoding

Info

Publication number: CN113630209B
Application number: CN202010381453.9A
Authority: CN
Inventors: 李林刚; 陈永锐; 李治军
Original assignee: University of Chinese Academy of Sciences
Current assignee: University of Chinese Academy of Sciences
Priority date: 2020-05-08
Filing date: 2020-05-08
Publication date: 2023-11-07
Anticipated expiration: 2040-05-08
Also published as: CN113630209A

Abstract

The invention provides a WiFi-to-Bluetooth Low Energy (BLE) cross-technology communication method based on narrowband decoding, namely NBee. The method is characterized in that: when the 802.11b signal sent by the WiFi passes through a Low Pass Filter (LPF) at the BLE receiving end, although the WiFi signal is distorted, the amplitude of the signal after filtering at the receiving end includes a part of the characteristics of the information sent by the WiFi. By utilizing the characteristics, the phase difference meeting the decoding condition of the BLE receiving end can be constructed by selecting proper data content at the WiFi transmitting end, so that the receiving end can decode accurately; the invention does not need to modify any hardware or firmware of the sending end and the receiving end, and realizes high-speed and reliable cross-technology communication from WiFi to BLE full channels under the condition of not influencing the receiving and transmitting of standard signals by equipment.

Description

WiFi-to-Bluetooth Low Energy (BLE) cross-technology communication method based on narrowband decoding

Technical Field

The invention relates to a heterogeneous device cross-technology communication method based on narrowband decoding. Specifically, when the WiFi transmits the 802.11b signal by DQPSK modulation or DBPSK modulation, the phase difference between adjacent sampling points of the WiFi signal at each interval of 1us after passing through a 1MHz low-pass filter of a BLE receiving end can meet BLE decoding conditions by reasonably selecting the transmitted data content, so that the transmission of the cross-technology information from the WiFi to the BLE full channel is realized, and the method belongs to the technical field of wireless communication.

Background

The explosive growth of wireless internet of things devices has led our wireless ecosystem to become increasingly diverse. For example, over 40 hundred million Bluetooth devices and over 15 hundred million WiFi devices coexist in the ISM band at 2.4 GHz. Such coexisting heterogeneous devices cause severe cross-technology interference (CTI) and have become a major obstacle to improving network scalability and spectral efficiency.

Some research has emerged to alleviate CTI, with cross-technology communication (CTC) as a very promising area of research, enabling direct communication between heterogeneous devices. CTCs, as an emerging technology, can bring many benefits to internet of things applications. First, conventional gateways incur additional hardware costs, deployment complexity, and data ingress and egress to these gateways result in additional latency, while CTCs avoid these drawbacks. Second, CTCs enable cross-technology collaboration between heterogeneous wireless devices. Third, CTCs help to reduce CTI and improve spectrum utilization.

The physical layer CTC (PHY-CTC) that has emerged in recent years realizes high-throughput communication through signal simulation, which closely simulates the waveform of a signal at the receiving end by selecting an appropriate payload at the transmitting end. However, these analog-based CTC techniques also face their own challenges. First, they are less reliable, and OFDM-based simulations are easily distorted by inherent errors (e.g., analog errors due to cyclic prefixes), thereby losing nearly half of the CTC packets. Secondly, due to quantization errors during simulation, only high-order modulation schemes such as 64QAM and the like can be adopted to simulate the target waveform, which limits the use of the signal simulation method in many application occasions. Third, physical layer cross-technology communication can only support few channels due to the tight constraints of the simulation process. For example, in WEBee, one WiFi channel can only support two ZigBee channels, while many other ZigBee channels cannot achieve cross-technology communication.

These drawbacks of CTCs based on signal simulation severely limit their application in many scenarios. Taking WiFi to Bluetooth Low Energy (BLE) CTCs as an example. For a BLE receiver it does not have the fault tolerance capability like ZigBee and therefore cannot tolerate any bit errors within a frame. Even with error correction codes, BLE loses the entire analog frame when an analog error occurs in its preamble or Access Address (AA). In addition, since BLE employs Adaptive Frequency Hopping (AFH) to prevent interference, when the BLE device hops to a channel that is not supported by CTCs, it cannot receive any CTC data packets. For these reasons, current signal simulation methods do not achieve reliable full channel physical layer CTC from WiFi to BLE well yet.

Disclosure of Invention

The invention provides a WiFi-to-Bluetooth Low Energy (BLE) cross-technology communication method based on narrowband decoding, namely NBee. The method can realize direct communication with high speed and high reliability from WiFi to BLE full channels.

The invention comprises the following contents:

1. CTC realization by DQPSK modulation

When the 802.11b signal sent by the WiFi single path (I path or Q path) passes through the 1MHz low-pass filter of the BLE receiving end, three modes are adopted for output signals: (a) When WiFi transmits alternating bits '1' and '0', the amplitude of the bit '1' filtered signal is almost always positive, while the bit '0' filtered amplitude is almost always negative. (b) When the WiFi transmits continuous bit '1', the signal amplitude after the filtering of the receiving end is always positive. (c) When the WiFi transmits continuous bit '0', the signal amplitude after the filtering of the receiving end is always negative. Therefore, when we send WiFi signals with DQPSK modulation, a very special phenomenon can be obtained: although the received signal is significantly distorted after passing through the BLE low pass filter, the original WiFi symbol (QPSK) and its corresponding BLE sample points are still in virtually the same quadrant within the constellation.

According to the above characteristics, as long as one WiFi symbol is given, we can infer the quadrant where the corresponding sampling point is located at the BLE receiving end. Therefore, given two consecutive WiFi symbols, we can determine whether the phase difference between two consecutive sampling points at the receiving end is positive or negative. In other words, if we choose to send the proper WiFi symbol sequence, we can get the phase difference needed for correct decoding at BLE receiver.

2. CTC realization by DBPSK modulation

Since the DBPSK signal has only a real part, the BLE filtered adjacent sample point phase difference can only be 0 or pi. This phase difference is BLE and cannot be decoded accurately, so we need to adjust the receive-side phase difference under DBPSK modulation.

Firstly, after passing through the BLE receiving end ADC module, the value of an nth sampling point used for calculating the phase difference is as follows:

s[n]＝(w(nTs)ej2π(fw-fB)nTs)*h (1)

where s n is the nth sampling point used by the BLE receiving end to calculate the phase difference. w (t) is a WiFi transmitting end baseband analog signal, fw is a WiFi carrier frequency, and fB is a BLE carrier frequency. Ts is the sampling point interval (1 us), x represents convolution, h represents the effect of the channel and filter on the signal. Since the BLE demodulator compares the phase difference between two adjacent samples, we bring n-1 into equation 1 to obtain the n-1 th sample value as:

s[n-1]＝(w((n-1)Ts)ej2π(fw-fB)(n-1)Ts)*h

＝(w((n-1)Ts)ej2π(fw-fB)nTs ej2π(fB-fW)Ts)*h (2)

by comparing s [ n ] and s [ (n-1) ], we found that the phase difference between these two sampling points consists of two parts: (1) The phase difference between the transmit side baseband signals w (nTs) and w ((n-1) Ts) (i.e., pi or 0), (2) s [ n-1] is an additional phase difference due to Carrier Frequency Offset (CFO) (i.e., 2pi (fw-fB) Ts). Thus, we can select the appropriate CFO (i.e., fw-fB) to construct a phase difference that can be accurately decoded at the BLE receiver.

In equation 2, when the carrier frequency offset of the WiFi transmitting end/BLE receiving end is adjusted to 250KHz (i.e. fw-fb=250 KHz), the n-1 th sampling point additional phase difference is pi/2 (=2pi×250khz×1/1 MHz). In this case, when WiFi transmits alternating bits, the phase difference between two consecutive sampling points at the receiving end is pi/2 (=pi-pi/2). Similarly, when WiFi transmits the same bit, the receive-side phase difference is-pi/2 (=0-pi/2). These two phase differences enable the BLE decoder to accurately decode bit '1' and bit '0', respectively.

The CFO cannot be set too large in order not to interfere with the normal reception of WiFi or BLE devices. The maximum limit of CFO specified by the 802.11 standard is 232KHz, and the carrier frequency allowable offset of some BLE chips can reach 250KHz, which has enough redundancy to realize correct decoding of the BLE receiving end. Considering the versatility of the chip and the inherent CFO, we set fw-fB to 125KHz. Such CFO values have little impact on adjacent channels due to the presence of guard bands between channels.

3. Full channel communication

According to the IEEE 802.11 and BLE4.0 standards, the center frequency of the WiFi channel and the center frequency of the BLE channel are not always identical. For example, the center frequencies of BLE channel 6 (2416 MHz) and WiFi channel 2 (2417 MHz) are offset by 1MHz. However, we have found that the NBee signal can still be correctly decoded by the BLE receiver even if there is a frequency offset, as long as the operating channel of the BLE receiver is within the coverage of the WiFi channel. The reason for this is as follows: as can be seen from equation (1), since the CFO between WiFi and BLE (i.e., fw-fB) is an integer multiple of 1MHz, the phase change due to the presence of the CFO (i.e., 2pi (fw-fB) nTs) will be an integer multiple of 2pi. For example, if CFO is nMHz, the phase of the nth sampling point will be rotated counter-clockwise by 2pi×nmhz by 1/1 mhz=2n pi. When a phase is rotated by an integer multiple of 2 pi, the position of the phase in the constellation diagram does not change, so the phase difference between adjacent sampling points does not change. Therefore, even if there is a carrier center frequency offset between the transmitting end and the receiving end, decoding of the receiving end is not affected as long as the BLE channel is within the coverage of the WiFi channel.

Since the WiFi channel is 22MHz and the BLE channel is 1MHz, one WiFi channel can cover 10 or 11 BLE channels. This means that each BLE channel can be covered by one or more WiFi channels. Based on this characteristic, for BLE devices in any channel, the WiFi transmitting end can find a channel to communicate with.

The WiFi-BLE cross-technology communication method based on narrowband decoding has the following advantages:

(1) Transparency: the NBee does not need to change the hardware and firmware of the commodity chip, only the payload of the WiFi frame needs to be carefully selected. This mechanism can ensure that neither WiFi device nor BLE device has an impact on the reception of the respective normal signals. In contrast, other CTCs (such as DopplerFi) at least need to modify the firmware of WiFi or BLE.

(2) High throughput: because BLE receives the NBee frame as a standard BLE frame, its transmission rate can reach 1Mbps, 3400 times higher than existing WiFi-BLE cross-technology communication technologies.

(3) High reliability: since this approach does not require signal simulation, inherent simulation errors are avoided and the reliability of NBee is far higher than that of simulation-based physical level CTCs.

(4) Full channel communication: NBee supports cross-technology communication of WiFi to BLE all channels, regardless of which channel BLE hops to, the WiFi device can remain connected to it all the time.

The invention designs a WiFi-BLE cross-technology communication method of a physical level by utilizing narrowband decoding, realizes high-speed and reliable information transmission of all channels from WiFi to BLE, and effectively avoids the limitation caused by gateway communication between heterogeneous devices. The invention provides new possibility for application of cross-technology communication in the Internet of things, and opens up a new idea for gateway-free direct communication between other heterogeneous devices.

Drawings

Fig. 1 is a design framework of the NBee proposed by the present invention.

Fig. 2 is a diagram comparing a signal after narrowband filtering at a receiving end with an original WiFi signal.

Fig. 3 is a feature of BLE filtered signals when WiFi transmits different signals.

Fig. 4 is a phase difference suitable for BLE receiver decoding using DQPSK modulation configuration.

Fig. 5 is a diagram of constructing a phase difference suitable for BLE receiver decoding using DBPSK modulation.

Fig. 6 is a diagram of full channel communication.

Fig. 7 is an experimental setup in an embodiment of the present invention.

Fig. 8 is a transmission rate comparison diagram of the present invention and the prior art WiFi to BLE cross technology communication.

Fig. 9 is a graph of frame reception rate at different transmission powers and transmission distances according to the present invention.

Fig. 10 is a diagram of frame reception rate for all BLE channels according to the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings, in order to make the objects, technical solutions, etc. of the present invention more apparent.

(1) Design framework of NBee: as shown in fig. 1, NBee uses narrowband decoding to realize WiFi to BLE high-speed cross-technology communication, and its principle is that a WiFi end selects appropriate data packet content to send, and then BLE can realize correct decoding according to appropriate phase difference between sampling points after receiving signals passing through a narrowband filter. The header and the trailer of the WiFi frame cannot be selected manually, but they are automatically ignored by the BLE end as noise, and the reception of the BLE frame is not affected.

(2) Comparison of the signal after the BLE receiving end narrowband filtration and the original WiFi signal: according to the 802.11 protocol and the BLE4.0 protocol, the bandwidth of WiFi is 22MHz, and the bandwidth of BLE is only 1MHz. The wideband WiFi signal must be distorted after passing through the BLE narrowband filter, however, the distorted signal still has some characteristics for transmitting information, and the receiving end can extract the information. After the WiFi signal has been spread, bit '1' is spread to 10110111000 (Barker code) and bit '0' is spread to 01001000111, as specified by DSSS in 802.11 b. Notably, the two sequences are complementary, in that they are in a symmetrical relationship about the x-axis in the time domain. Furthermore, the duration of each sequence is 1us, and the throughput of BLE is 1Mbps (=1/1 us), which means that the BLE receiving end performs a phase difference calculation at each barker code sequence to decode. When two consecutive sequences transmitted by WiFi are complementary, the sampling point values of BLE receiving end are opposite to each other every 1us, i.e. the phase difference is pi (as shown in fig. 2, the phase difference between sampling points s n-1 and s n is pi). When two continuous sequences sent by the WiFi sending end are the same, the phase difference between sampling points of the receiving end is 0. This difference in phase enables us to transmit information from WiFi to BLE.

(3) Characteristics of BLE filtered signals when WiFi transmits different signals: as shown in fig. 3, we find that, regardless of the bit sequence transmitted by WiFi, after passing through a low pass filter of 1MHz at the BLE receiving end, the amplitude of the bit '1' part of the signal is almost always positive, while the amplitude of the bit '0' part is almost always negative. We can use this feature to achieve cross-technology information transfer.

(4) Phase difference suitable for BLE receiving end decoding is constructed by DQPSK modulation: as shown in fig. 4, if the symbol transmitted by WiFi is '11' (in the first quadrant), then the sampling point corresponding to BLE will also be in the first quadrant (see red point at time t 1). This is because the bits of the in-phase (I) branch and the quadrature (Q) branch of the WiFi transmitting end are both '1', and the I-path amplitude and the Q-path amplitude of the sampling point after filtering by the receiving end are both positive values. According to the feature that the symbol sent by the WiFi and the sampling point of the BLE receiving end are in the same quadrant of the constellation diagram, only one WiFi symbol is given, and the quadrant in which the corresponding sampling point of the BLE receiving end is located can be deduced. Therefore, given two consecutive WiFi symbols, we can determine whether the phase difference between two consecutive sampling points at the receiving end is positive or negative. In other words, if we choose to send the proper WiFi symbol sequence, we can get the phase difference needed for correct decoding at BLE receiver.

For example, in fig. 4, when the I-path of the WiFi transmitting end transmits '1,0,1', and the Q-path transmits '1, 1' (i.e. the WiFi sequentially transmits symbols '11', '01', '11'), after passing through the 1MHz low pass filter of the BLE receiving end, the t1 and t2 sampling points are respectively located in the first and second quadrants of the constellation diagram according to the amplitude of the I/Q signal. Therefore, the phase difference between the sampling points t1 and t2 (1 us interval) is greater than 0 and less than pi, which is decoded by the BLE receiving end as bit '1'. Similarly, the phase difference between the sampling points t2 and t3 (1 us interval) is greater than-pi and less than 0, and the ble receiving end can decode bit '0'.

(5) Constructing a phase difference suitable for BLE receiving end decoding by using DBPSK modulation: fig. 5 shows the phase difference change before and after the addition of a 125KHz carrier frequency offset. In fig. 5a, wiFi sequentially transmits bit '1' and bit '0' through DBPSK modulation. When the carrier frequency is not shifted, the phase difference between two adjacent sampling points of the BLE receiving end is pi, and the BLE receiving end cannot be decoded correctly. When a CFO of 125KHz (fB-fw=125 KHz) is added to the WiFi transmitting/BLE receiving, the s [ n-1] sampling point is rotated by pi/4 more counterclockwise than the s [ n ] sampling point. Therefore, the phase difference between the two sampling points is changed from pi to 3 pi/4, and the BLE receiving end can accurately decode the bit '1'. In fig. 5b, wiFi continuously transmits the same bits through DBPSK modulation. When the CFO is not added, the phase difference between adjacent sampling points of the receiving end is 0, and the decoding cannot be correctly performed. However, when a CFO of 125KHz is added, the phase difference is changed from 0 to-pi/4, and BLE can accurately decode bit '0'.

(6) Full channel communication schematic: as shown in fig. 6, wiFi channel 6 covers 10 BLE channels, and WiFi channel 11 covers 11 BLE channels. Each BLE channel is within the coverage of one or more WiFi channels, so WiFi can remain in communication with it regardless of which channel the BLE hops to.

(7) Experimental equipment in the embodiment of the invention: the NBee transmitting end is USRP N210 working under the 802.11b/g protocol. The NBee receiving end is a TI CC2650 chip and works in a BLE mode. In the experiment, each CTC BLE frame consists of a one byte preamble (10101010 b), a four byte Access Address (AA) (0 x8E89BED 6), a variable length PDU and a three byte CRC. The WiFi transmitting end modulates the transmission signal using DBPSK (1 Mbps) and DQPSK (2 Mbps), respectively. We set the default BLE channel to 39 (broadcast channel) and the default WiFi channel to 13. Experimental evaluation results include Bit Error Rate (BER), frame Reception Rate (FRR), and throughput. To ensure statistical validity, we have 10 average results for the same experimental scenario for each data. Under various experimental condition settings of indoor/corridor, short distance/long distance, moving scene and the like, 1000 NBee data packets are sent for each experiment. The experimental platform setup is shown in fig. 7.

(8) In contrast to the transmission rates of the prior art: as shown in fig. 8, the data rate of freebees is only 31.5bps, which is low because it uses the periodicity of WiFi beacons to communicate information. The data rates of DCTC and C-Morse are 41bps and 50.4bps, respectively, which take advantage of the time of the data packets to carry information. DopplerFi is a recently proposed WiFi to BLE CTC that uses Carrier Frequency Offset (CFO) to transmit information, increasing throughput to 250bps. NBee, as a physical level CTC from WiFi to BLE, can reach a transmission rate of 1Mbps, the same as a standard BLE rate. The throughput of NBee is more than 3400x higher than other WiFi to BLE CTC schemes because the other schemes are packet-level, they can only carry small amounts of bit data in one or more packets.

(9) Frame reception rate at different transmission power and transmission distance: we analyze the FRR of NBee as a function of transmission distance and transmission power. In the experimental process, the distance between the WiFi transmitting end and the BLE receiving end is increased from 5m to 20m, and the power of the transmitting end is increased from 0dBm to 20dBm. Fig. 9 shows FRR of NBee at different transmission distances and transmission powers. When the transmission distance is short (e.g., 5 m), the change in transmission power has little effect on FRR (remains at 95% or more throughout). When the distance is increased to 20m, the FRR drops sharply with a drop in transmit power, with FRR only 20% at a transmit power of 0dBm. This means that when the transmission distance is more than 10m, the transmission power of NBee has a large influence on the frame reception rate. However, the transmission power of WiFi is typically greater than 10dbm (e.g., mobile phone), so even at a distance of 20m, the FRR of NBee can be higher than 80%.

(10) Frame reception rate under all BLE channels: to verify that NBee supports full channel communication, we tested the performance of NBee on 40 BLE channels. During the test, wiFi was operated in channel 1, channel 5, channel 9 and channel 13 in sequence. These four WiFi channels may cover all BLE channels. As can be seen from the results of fig. 10, NBee works stably on all BLE channels (FRR > 95%), even when BLE is located on broadcast channels (channel 37, channel 38, channel 39). This is because the power of the WiFi transmitter is much higher than the BLE device, so other BLE signals on the broadcast channel have less impact on the NBee signal.

The experiment proves that: the WiFi-BLE cross-technology communication method based on narrowband decoding can realize reliable transmission of the WiFi-BLE full channel. The frame receiving rate can reach more than 95%, the throughput can reach 1Mbps, which is 3400 times of the prior art.

Claims

1. A WiFi to Bluetooth Low Energy (BLE) physical layer cross-technology communication method based on narrowband decoding is characterized in that: by reasonably selecting the data content of the WiFi transmitting end, constructing a phase difference which accords with the decoding condition of a decoder after passing through a 1MHz low-pass filter of the BLE receiving end, and correctly decoding the BLE receiving end can be realized on all BLE channels;

when the 802.11b signal modulated by DBPSK is transmitted, after the WiFi/BLE is added with proper carrier center frequency offset, adjacent bits selected to be transmitted are transmitted and pass through a BLE receiving end low-pass filter, and the phase difference of corresponding adjacent sampling points accords with BLE decoding conditions;

after a proper WiFi sending bit is selected by a sending end, after the WiFi sending bit passes through a low-pass filter of 1MHz of a receiving end, the phase difference between adjacent sampling points of each 1us interval of a distorted WiFi signal meets the decoding condition of a BLE decoder, namely the phase difference is positive and is decoded into bit '1', and the phase difference is negative and is decoded into bit '0';

the specific steps for realizing CTC by DBPSK modulation include:

s[n]＝(w(nTs)ej2π(fw-fB)nTs)*h (1)；

s [ n ] is the nth sampling point used by the BLE receiving end to calculate the phase difference, w (t) is the WiFi transmitting end baseband analog signal, fw is WiFi carrier frequency, fB is BLE carrier frequency, ts is sampling point interval, x represents convolution, h represents the influence of a channel and a filter on the signal;

since the BLE demodulator compares the phase difference between two adjacent samples, we bring n-1 into equation (1), resulting in the n-1 th sample value being:

s[n-1]＝(w((n-1)Ts)ej2π(fw-fB)(n-1)Ts)*h＝(w((n-1)Ts)ej2π(fw-fB)nTsej2π(fB-fW)Ts)*h(2)

by comparing s [ n ] with s [ (n-1) ], the phase difference between these two sampling points is composed of two parts: (1) The phase difference between the transmit side baseband signals w (nTs) and w ((n-1) Ts) (i.e., pi or 0), (2) s [ n-1] an additional phase difference due to Carrier Frequency Offset (CFO), i.e., 2pi (fw-fB) Ts; thus, the appropriate CFO, fw-fB, is selected to construct a phase difference that is accurately decodable at the BLE receiver;

constructing a phase difference suitable for BLE receiving end decoding by using DBPSK modulation: wiFi sequentially transmits bit '1' and bit '0' through DBPSK modulation; when no carrier frequency offset exists, the phase difference between two adjacent sampling points of the BLE receiving end is pi, and the BLE receiving end cannot be decoded correctly; when a CFO of 125KHz (fB-fw=125 KHz) is added to the WiFi transmitting end/BLE receiving end, the s [ n-1] sampling point is rotated by pi/4 anticlockwise more than the s [ n ] sampling point; therefore, the phase difference between the two sampling points is changed from pi to 3 pi/4, and the BLE receiving end can accurately decode the bit '1'.