A method for eliminating the influence of BLE direction finding carrier frequency deviation to obtain phase delay

Abstract

The present invention discloses a method for eliminating the influence of BLE direction finding carrier frequency deviation to obtain phase delay, so that the phase delay only related to the direction of the incident signal source can be obtained with extremely low computing resources without estimating the carrier frequency deviation, thereby improving the accuracy of the BLE direction finding system in the presence of carrier frequency deviation and improving the stability of the BLE direction finding system under various operating temperature conditions.

Description

Technical Field

[0001] The present invention relates to the field of wireless high-precision direction finding, and specifically to an antenna array switch switching technology for resisting carrier frequency deviation in ultra-low power Bluetooth (BLE) high-precision direction finding, and a method for eliminating the influence of BLE direction finding carrier frequency deviation and obtaining phase delay based on the switching technology. Background Art

[0002] The global positioning navigation system has brought many conveniences to people's production and life. With the development of Internet of Things technology, location-based services are gradually expanding from outdoor to indoor. Warehousing logistics, personnel tracking, indoor navigation, etc. all rely on high-precision indoor positioning services. Since satellite navigation positioning signals cannot be correctly received indoors, indoor positioning relies on various positioning solutions deployed indoors.

[0003] How to achieve centimeter-level positioning accuracy has become a hot topic in indoor positioning research in recent years. With the help of low-power, low-cost IoT positioning technology, large-scale applications have been realized in large commercial complexes such as indoor smart parking, reverse car search, store navigation, and museum exhibition hall interaction. However, in the scenarios where large-scale applications have been implemented, most of them use the received signal strength indication (RSSI) to achieve positioning. The current average positioning accuracy can only reach the meter level, which cannot meet the high-precision positioning requirements in application scenarios such as emergency search and rescue, valuables, safe production, warehousing and sorting, and competitive sports.

[0004] How to achieve high-precision array direction finding on battery-powered and resource-constrained low-power IoT nodes is an urgent problem to be solved. Currently, the main technical means to achieve centimeter-level high-precision positioning are angle of arrival (AoA), time of flight (ToF) and time difference of arrival (TDoA). In order to further improve positioning accuracy, the Bluetooth standard organization introduced AoA / AoD direction finding technology combined with array antennas in versions after 5.1. At present, companies such as Quuppa in Finland have launched direction finding solutions based on BLE AoA technology.

[0005] Positioning technology based on array signal processing has been studied for nearly half a century, and many achievements have been made in related super-resolution, anti-multipath direction finding, array error calibration, etc. However, the related algorithms have high requirements on system software and hardware resources, and are difficult to implement on resource-constrained low-power, low-cost IoT nodes. In order to apply array direction finding theory to battery-powered IoT devices to achieve high-precision positioning, many practical problems still need to be solved. One of them is that the current BLE AoA / AoD direction finding accuracy is seriously affected by carrier frequency deviation. The following briefly describes the impact of carrier frequency deviation on direction finding accuracy in BLE AoA / AoD direction finding.

[0006] After a single wireless signal source TX transmits a narrowband signal, the signal can be propagated in a far-field manner. At this time, there is a propagation delay between the wireless signals received in parallel by the two antennas A1 and A2 with a spacing of l, which is related to the direction of the incident signal, such as Figure 1 Assume Figure 1 The signal received by A1 is:

[0007] s1(t)=x(t)

[0008] Then ideally the signal received by A2 is:

[0009] s2(t)=x(t-τ)

[0010] Where τ is the propagation delay caused by the distance difference d projected from the antenna spacing l onto the wireless signal propagation path. If the narrowband signal properties are used, we can further obtain:

[0011] s2(t)=x(t-τ)≈x(t)e -j2πfτ

[0012] because And the speed of wireless signal propagation in the air is approximately the speed of light c, so the propagation delay can be obtained as:

[0013]

[0014] Therefore, if the target source is found in a single direction, the conjugate multiplication of the two antenna signals received in parallel can be obtained:

[0015] s2(t)s1 * (t)=|x(t)| 2 e -j2πfτ

[0016] We can further obtain that the phase of the signal after conjugate multiplication satisfies:

[0017]

[0018] Wherein λ=c / f is the wavelength of the wireless signal.

[0019] Through the above derivation, when the distance l (l≤λ / 2) between the two parallel receiving antennas, the wavelength λ (or frequency f) of the direction-finding signal are determined, and the phase θ after the conjugate multiplication of the parallel receiving signals of the two antennas is obtained, the incident direction of the signal can be obtained by calculation

[0020] In the above model, when l=λ / 2, the value range of θ can be [-π,π], and the corresponding incident signal direction range is [0,π].

[0021] Assume that the center frequency f of the above direction finding signal has a center frequency deviation of εppm. Then the direction finding signal frequency at this time can be expressed as:

[0022] f′=f(1+ε×10 -6 )

[0023] If the antenna spacing is still half the wavelength of the direction finding signal, the phase error caused can be:

[0024]

[0025] Therefore, if the center frequency has a 10ppm deviation in 2.4GHz direction finding, that is, the maximum carrier frequency deviation is 24kHz, the maximum direction finding phase error caused by this is no more than 1.8×10 -3 °, can be ignored.

[0026] The above direction finding principle is extended to an array antenna composed of multiple antennas (taking a linear array antenna as an example), and an array antenna direction finding model composed of parallel receivers is obtained, as shown in Figure 2 shown.

[0027] Therefore, the traditional array direction finding model can be obtained, namely:

[0028] X=AS+N

[0029] However, in the above parallel direction finding system, since multiple receivers are used in parallel to receive antenna array signals, the overall system cost is high and the system power consumption is high, which is not suitable for battery-powered low-power IoT application scenarios. At the same time, based on the above analysis, the direction finding phase error caused by carrier frequency deviation in the parallel direction finding system is small and can be ignored.

[0030] In order to achieve high-precision direction finding in IoT scenarios, BLE adopts Figure 3 The system architecture shown in Figure 1. There is only one receiver in this solution, and the received signal of the array antenna is obtained by switching the RF switch at the front end. Figure 3 The direction finding system in the paper is aimed at low-power and low-cost IoT application scenarios, and uses RF switches to reuse RF receivers. However, it also introduces some issues that are not considered in parallel direction finding systems, one of which is the frequency deviation of the carrier of the direction finding signal.

[0031] Taking two antennas as an example, we establish a switch switching signal model and assume that the switch switching time interval is t slot . Then ideally, the receiver is in [0,t slot ) time, the antenna A1 is turned on by the switch to complete the signal acquisition and obtain the signal s1(t); then, in [t slot ,2t slot) time, the antenna A2 is turned on by the switch to complete the signal acquisition and obtain the signal s2(t), so we can get:

[0032] s1(t)=x(t),t∈[0,t slot )

[0033]

[0034] Therefore, if the two signals are still conjugate multiplied, we can get:

[0035]

[0036] We can further obtain that the phase of the signal after conjugate multiplication satisfies

[0037] angle(s2(t)s1 * (t)) = θ = -2πf(τ-t slot )=-2πfτ+2πft slot

[0038] According to the same error model, if the error of the previous item is ignored, then when there is a center frequency deviation of εppm at the carrier center frequency f, the direction finding phase error is:

[0039] |θ′-θ|=|2πf′t slot -2πft slot |=2πεft slot ×10 -6

[0040] According to the BLE direction finding standard, when t slot = 4us, under the condition of 10ppm frequency accuracy in 2.4G band, the maximum direction finding phase error can reach about 34°. Such a large direction finding phase error will seriously affect the direction finding accuracy, so it is necessary to focus on how to eliminate it.

[0041] In order to solve the serious direction finding error caused by the direction finding carrier frequency deviation in the switch switching system, the structure of the direction finding data packet is defined in the Bluetooth direction finding standard, such as Figure 4As shown. The direction finding data packet in the standard adds a CTE waveform after the traditional BLE physical layer data packet. This waveform is a continuous bit 1 that passes through the GFSK modulator of the BLE specification, and ideally generates a 250kHz sinusoidal signal in the baseband. Before each slot that the switch allows to switch, the standard specifically defines an 8us reference period. This waveform switch cannot be switched, and is mainly used to estimate the actual frequency deviation based on the signal of this waveform after acquisition. However, the resources of BLE devices used in IoT scenarios are limited, and the number of calculation points obtained within 8us specified in the standard is limited. Therefore, no matter what method is used to estimate the direction finding frequency, the performance is poor under low signal-to-noise ratio. Summary of the invention

[0042] In order to eliminate the direction finding phase error caused by carrier frequency deviation at a lower computing cost and time on IoT nodes with limited computing resources, the present invention proposes a method for switching an antenna switch, and further proposes a method for eliminating the influence of BLE direction finding carrier frequency deviation on obtaining phase delay. The direction finding phase error caused by frequency deviation can be directly eliminated without knowing the current carrier signal frequency deviation, and the phase information only related to the incident direction can be directly obtained.

[0043] To achieve the above object, the present invention adopts the following technical solutions:

[0044] A method for eliminating the influence of BLE direction finding carrier frequency deviation on obtaining phase delay comprises the following steps:

[0045] Step 1: Switch antenna m and antenna n in different slots to collect direction finding signals;

[0046] Step 2: Perform pairwise conjugate operations on the signals collected from different slots;

[0047] Step 3: Take the square root of the conjugate operation result of the same antenna and compensate it to the conjugate operation result between different antennas;

[0048] Step 4: After eliminating the frequency offset, obtain the phase delay of the collected signal between antennas that is only related to the incident direction.

[0049] Preferably, the switching mode of step 1 is: using m,n,m or n,m,n as the basic switching mode to switch antenna m and antenna n.

[0050] Preferably, the switching method of step 1 is: switching the antenna m and the antenna n by cycling the basic form multiple times.

[0051] Preferably, the conjugate operation in step 2 is conjugate multiplication followed by averaging.

[0052] Preferably, the data obtained after taking the square root of the conjugate operation result in step 3 remains in the same quadrant as the original data.

[0053] Beneficial effects: The present invention can directly eliminate the direction-finding phase error caused by the frequency deviation without knowing the current carrier signal frequency deviation, and directly obtain the phase information only related to the incident direction, and has the advantages of low calculation cost and short calculation time. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a schematic diagram of the antenna array receiving signals in parallel to estimate the direction of the incident signal when the wireless signal is in the far field approximation;

[0055] Figure 2 Direction finding model for linear array antenna parallel receiver;

[0056] Figure 3 Direction finding systems for low-cost receiver multiplexing;

[0057] Figure 4 It is the direction finding signal waveform specified in the Bluetooth direction finding standard;

[0058] Figure 5 The present invention proposes a two-antenna switching method to resist carrier frequency deviation;

[0059] Figure 6 The invention discloses an antenna switching method for resisting carrier frequency deviation in a four-antenna switch direction finding system. DETAILED DESCRIPTION

[0060] The following is a receiver multiplexing switch antenna array direction finding system. If you need to obtain the incident signal direction-dependent phase delay between antenna m and antenna n that is independent of the carrier frequency offset, you need:

[0061] Step 1: Using the antenna switching method proposed in the present invention, the antennas are switched in slot i, slot (i+1) and slot (i+2) in the order of m, n, m or n, m, n to collect direction finding signals;

[0062] Step 2: conjugate and multiply the data collected by slot i and slot (i+1) and take the average, recorded as p1;

[0063] Step 3: conjugate and multiply the data collected by slot i and slot (i+2) and take the average, recorded as p2;

[0064] Step 4: Take the square root of p2 and keep it in the same quadrant as p2 to get p3;

[0065] Last step: Finally, by multiplying p1 and p3, we can get the phase delay that is only related to the direction of the incident signal after eliminating the carrier frequency offset.

[0066] The following is an explanation of the above method using two antennas as an example. According to the method definition, the switching of the two antennas is as follows Figure 5 shown.

[0067] In order to obtain the phase difference between the two antennas caused by the incident direction of the signal source to be measured, it is necessary to collect the signal of antenna 1 on slot 1, collect the signal of antenna 2 on slot 2, and finally switch back to antenna 1 to collect the signal again on slot 3.

[0068] At this point, three groups of signals can be obtained respectively:

[0069] s1(t)=x(t),t∈[0,t slot )

[0070]

[0071]

[0072] Therefore, by conjugating the antennas in pairs, we can get:

[0073]

[0074]

[0075] In the BLE direction finding standard, t slot It can be 2us or 4us, and the ideal frequency for direction finding satisfies:

[0076] f(n)=(2402+2n+0.25)MHz,n=0,1,…,39

[0077] In the above formula, the highest direction finding carrier is in the 2480MHz channel, and the actual direction finding frequency with frequency deviation satisfies:

[0078] f=f(n)+f error

[0079] According to the capabilities of the current BLE wireless chip phase-locked loop and passive crystal oscillator, f error can be controlled within 10ppm of f(n), then we can get:

[0080] 4πft slot =2mπ+4πf error t slot

[0081] Wherein m is an integer.

[0082] If we further consider the periodicity of the phase, we only consider the phase caused by the frequency offset in [-v,π). In addition, under the same frequency accuracy of 10ppm, the maximum frequency offset error is at 2480MHz. Therefore:

[0083] f error ≤2480MHz×10ppm=24.8kHz

[0084] And t slot The maximum is 4us, so [-π,π) can be obtained:

[0085] |4πft slot |=|4πf error t slot |<0.4π

[0086] Therefore, directly s1(t)s3 * The result of (t) only needs to consider the solution of [-0.4π, 0.4π), so it can be ensured that:

[0087]

[0088] Similarly, we can get:

[0089]

[0090] After multiplication, we get:

[0091] s2(t)s1 * (t)(s1(t)s3 * (t)) 0.5 =|x(t)| 3 e -j2πf

[0092] Therefore, after finding the phase, we can get:

[0093]

[0094] It can be found through observation that after the above processing, the direction finding error caused by the carrier frequency offset is directly eliminated in the case of unclear carrier frequency offset.

[0095] If it is a multi-antenna switch direction finding system, taking four antennas as an example, a switching method provided by the present invention is:

[0096] Using the data collected by slot 1, slot 2, and slot 3, the phase delay related to the direction of the incident signal is obtained after the carrier frequency deviation between antenna 1 and antenna 2 is resisted;

[0097] Using the data collected by slot 3, slot 4, and slot 5, the phase delay related to the direction of the incident signal is obtained after the carrier frequency deviation between antenna 1 and antenna 3 is resisted;

[0098] Using the data collected by slot 5, slot 6, and slot 7, the phase delay related to the direction of the incident signal is obtained after the carrier frequency deviation between antenna 1 and antenna 3 is resisted;

[0099] Subsequently, the direction finding target can be estimated with high precision through traditional direction finding methods, such as MUSIC.

Claims

1. A method for eliminating the influence of BLE direction finding carrier frequency deviation on obtaining phase delay, characterized in that: The steps include: Step 1: Switch antenna m and antenna n in different slots to collect direction finding signals respectively. The switching mode is: use m,n,m or n,m,n as the basic switching mode to switch antenna m and antenna n, or use m,n,m or n,m,n as the basic switching mode and perform multiple cycles in the basic switching mode to switch antenna m and antenna n. Step 2: Perform pairwise conjugate operations on the signals collected from different slots; Step 3: Take the square root of the conjugate operation result of the same antenna and compensate it to the conjugate operation result between different antennas; Step 4: After eliminating the frequency offset, obtain the phase delay of the collected signal between antennas that is only related to the incident direction.

2. According to claim 1, a method for eliminating the influence of BLE direction finding carrier frequency deviation to obtain phase delay is characterized in that: The conjugate operation in step 2 is conjugate multiplication followed by averaging.

3. According to claim 1, a method for eliminating the influence of BLE direction finding carrier frequency deviation to obtain phase delay is characterized in that: In step 3, the data obtained after taking the square root of the conjugate operation result remains in the same quadrant as the original data.

Images