CN106878218A - A Highly Reliable Demodulation Method for IEEE802.15.4 - Google Patents

Abstract

A high-reliability BPSK demodulation method for IEEE802.15.4, first using the channel reception data corresponding to the 32-bit preamble to extract useful frequency offset information , and then perform bit-level differential processing on the complex baseband received sampling signal corresponding to the PSDU to obtain the judgment observation value , and then use the extracted frequency offset information to compensate the judgment observation value and then perform detection and judgment, and finally transmit the received PSDU data to the MAC layer for CRC check. The present invention is based on the Taylor series expansion of the arcsine function and the theory that the influence of noise on the useful signal can be ignored when the signal-to-noise ratio is large, and provides a non-coherent BPSK suitable for IEEE 802.15.4 with low power consumption, high reliability and low cost receiver.

Description

A Highly Reliable Demodulation Method for IEEE802.15.4

technical field

The invention relates to the technical field of communication signal waveform detection, in particular to a high-reliability BPSK demodulation method for IEEE802.15.4.

Background technique

IEEE 802.15.4 is the basis of ZigBee, WirelessHART and other specifications, and describes the physical layer and media access control protocol of low-speed wireless personal area network. It initially works in the 868/915MHz and 2.4GHz ISM frequency bands, and the data transmission rate can reach up to 250kbps. The advantages of low power consumption and low cost make it widely used in many fields such as data collection, processing and analysis, remote control precision farming automation, environmental protection and monitoring. In the latest standard provided in 2011, 314–316MHz, 430–434MHz, 779–787MHz and 950–956MHz operating frequency bands were added.

As shown in Figure 1, the 802.15.4 protocol uses different modulation methods and data transmission rates on different carrier frequency bands. A total of 48 channels are provided in four typical frequency bands: 1 channel in the 868MHz frequency band, 10 channels in the 915MHz frequency band, 16 channels in the 2450MHz frequency band, and 21 channels in the 950MHz frequency band. As shown in Figure 2, on the 868/915/950-MHz bands, the signal processing is the same, only the data rate is different. The sender first differentially encodes the binary data of the physical layer data protocol unit (PPDU), and then converts each bit after differential encoding into a chip sequence with a length of 15, and finally modulates it onto the channel using BPSK. Differential encoding is to XOR each original bit of data with the bit generated by the previous differential encoding: Among them, E _n is the result of differential encoding, R _n is the original bit to be encoded, and E _n-1 is the result of the last differential encoding. For each data packet sent, R ₁ is the first original bit, and it is assumed that E ₀ =0 when calculating E ₁ . The differential demodulation process is similar to the encoding process: For each received data packet, E ₁ is the first bit to be demodulated, and it is assumed that E ₀ =0 when calculating E ₁ . As shown in Figure 3, each bit after differential encoding is converted into a slice sequence with a length of 15. The sequence after spreading is modulated onto the carrier using BPSK modulation.

As shown in Figure 4, the first field of the IEEE 802.15.4 protocol physical layer data frame structure is a four-byte 32-bit all-zero preamble. When the transceiver receives the preamble, it will Complete slice synchronization and symbol synchronization. The length of the Start of Frame Delimiter (SFD) field is one byte, and its value is fixed at 0xA7, which indicates the beginning of a physical frame. After the transceiver receives the preamble, it can only achieve bit synchronization of the data. By searching the SFD field The value 0xA7 can be synchronized to the byte. The frame length is represented by the lower 7 bits of a byte, and its value is the length of the physical frame payload, so the length of the physical frame payload will not exceed 127 bytes. The payload length of the physical frame is variable, which is called the physical layer service data unit (PSDU), and is generally used to carry the MAC frame.

There are two traditional demodulation methods for IEEE802.15.4. One is the traditional typical complex baseband non-coherent demodulation method in the 868/915/950-MHz frequency band as shown in FIG. 5 . use Represents the complex baseband sampling signal received after channel transmission, where s(k) is the transmitted data to be detected, s(k)∈{+1,-1}, ω ₀ =2πf ₀ , f ₀ and θ are respectively Frequency offset and phase offset remain constant in the entire data frame, _Tc represents the chip period of the spreading code, and η ₀ (k) is the complex baseband additive white Gaussian noise, and the demodulation process can be summarized as:

Step 1, using the complex baseband receiving sample signal corresponding to the 32-bit preamble to calculate the observation value Y containing the frequency offset information:

Wherein, J represents the total number of bits of the preamble, J=32, N represents the spreading length, N=15, 1≤m≤J-1,

0≤n≤N-1, p[n+Nm] represents the channel reception value of the n-th chip corresponding to the m-th bit of the preamble, (·) ^* represents the conjugate operation, η ₁ represents all the noise item.

Step 2: Perform bit-level differential processing on the complex baseband received sampling signal corresponding to the PSDU to obtain the judgment observation value A[m]:

Among them, r[n+Nm] represents the channel reception value of the n-th chip corresponding to the m-th bit of the PSDU, η ₂ [m] represents all noise items, and E[m] represents the m-th bit data sent .

Step 3. Use Y ₀ in step 1 to extract frequency offset information, and perform detection and judgment after compensating A[m] in step 2:

in, Indicates the judgment result of the mth bit data, q(·) is a quantization function, |·| represents a modulo operation, Re(·) represents a real part operation, and Im(·) represents an imaginary part operation. Bloch, MR, Hayashi, M., and Thangaraj, A. in the article "IEEE802.15.4BPSK receiver architecture based on a new efficient detection scheme" published in "IEEE Transactions on Signal Processing" in September 2010 provide a A Calculation Method of Quantization Function of Y body, is the phase of Y and is also an estimate of Nω ₀ T _c , The calculation method is specifically described as:

The detection and decision process of formula (3) needs to extract the estimated value of frequency offset information Nω ₀ T _c from Y in advance A[m] is then compensated. As mentioned above, the disadvantage of the traditional typical non-coherent demodulation method is that: from the formula (4), it is necessary to obtain the estimated value of the frequency offset information Nω ₀ T _c through division and complex arctangent operations in step 3 For 802.15.4 network terminals whose energy supply is strictly limited, the calculation complexity is relatively large, the energy consumption is relatively large, and the implementation cost is relatively high.

Another demodulation method is to reduce the implementation complexity of extracting frequency offset compensation information Nω ₀ T _c from Y. It was proposed by Lee, S., Kwon, H., Jung, Y., and Kim, JS in 2007 on August A simplified form of demodulation method proposed in the article "Efficient non-coherent demodulation scheme for IEEE 802.15.4LR-WPANsystems" published in "Electronics Letter" in July, improves the quantization function of Y, which can be described as:

It can be seen from formula (5) that in this simplified scheme, the estimated value of the frequency offset information Nω ₀ T _c Can be specifically described as:

It can be seen that this simplified demodulation method also needs to extract the estimated value of the frequency offset information Nω ₀ T _c from Y in advance A[m] is then compensated. Use formula (5) to The approximation process of , will inevitably produce a large error, that is, the estimation process of the frequency offset information Nω ₀ T _c in formula (6) has a relatively serious "over-estimation" or "under-estimation" phenomenon, which will lead to a large drop in reliability. There is no good balance between implementation complexity and performance. As shown in Figure 6, compared with the traditional typical complex baseband non-coherent demodulation method, the performance loss of the simplified demodulation method is serious. The carrier frequency used in the simulation is 924MHz, and the frequency offset is specified in the IEEE 802.15.4 protocol The maximum value of 80ppm, the phase offset θ obeys the uniform distribution in (0, 2π], the data length of PSDU is 20 bytes (160 bits), and at least 3000 frame errors are collected under each SNR. And 802.15.4 The network MAC layer uses cyclic redundancy check (CRC) to judge the correctness of the transmission frame, and the automatic retransmission request (ARQ) protocol determines whether the transmission frame needs to be retransmitted based on this, without using the forward error correction (FEC) mechanism. Therefore, the performance of the physical layer demodulation method will have a huge impact on energy consumption. When the channel condition is poor and the communication distance is long, the power loss of the received signal is large. At this time, if the demodulation method of this simplified form is adopted , the same PSDU data frame may be successfully received by the MAC layer after multiple retransmissions. If the amount of data is huge, the communication process of multiple retransmissions will also consume huge energy, which will reduce the use of 802.15.4 networks with insufficient energy supply life.

Contents of the invention

In order to solve the deficiencies in the prior art, the present invention is based on the Taylor series expansion of the arcsine function and the theory that the influence of noise on the useful signal can be ignored when the signal-to-noise ratio is large, and provides a low computational complexity suitable for IEEE 802.15.4 , low power consumption, high reliability and low cost non-coherent BPSK demodulation method.

In order to achieve the above object, the specific scheme adopted by the present invention is:

A high-reliability BPSK demodulation method for IEEE802.15.4. The data frame of the physical layer of the sending end is transmitted to the receiving end through a channel after spread spectrum and BPSK modulation. The data frame includes a 32-bit preamble and a physical layer service Data unit PSDU; the complex baseband sampling signal received by the receiving end is expressed as Where s(k) is the transmitted data to be detected, s(k)∈{+1,-1}, ω ₀ =2πf ₀ , f ₀ and θ are frequency offset and phase offset respectively and in the whole data frame Keeping unchanged, _T represents the spreading code chip period, and η ₀ (k) is the complex baseband additive white Gaussian noise; the concrete demodulation steps are:

Step 1. Use the channel received data corresponding to the 32-bit preamble to extract the frequency offset observation value Y including the frequency offset information:

Among them, J represents the total number of bits of the preamble, J=32, N represents the spreading length, N=15, 1≤m≤J-1, 0≤n≤N-1, p[n+Nm] represents the preamble The channel reception value of the n-th chip corresponding to the m-th bit of , ( ) ^* represents taking a conjugate operation, and η ₁ represents all noise terms;

Step 2: Perform bit-level differential processing on the complex baseband sampling signal corresponding to the PSDU to obtain the judgment observation value A[m]:

Among them, r[n+Nm] represents the channel reception value of the n-th chip corresponding to the m-th bit of the PSDU, η ₂ [m] represents all noise items, and E[m] represents the m-th bit data sent ;

Step 3: Use the frequency offset observation value Y in step 1 to extract frequency offset information, compensate A[m] in step 2, and then perform detection and judgment:

in, Indicates the mth bit data obtained by detection and judgment, q( ) is a quantization function, and q(Y) is specifically expressed as:

Among them, |·| represents the modulo operation, represents the estimated value of the frequency offset Nω ₀ T _c in A[m];

Step 4, after the detection is completed, the received PSDU data is sent to the MAC layer for CRC check;

In said step three, The calculation formula is:

Among them, Re(·) represents the operation of taking the real part, and Im(·) represents the operation of taking the imaginary part.

As a preferred solution, in the step 3 The calculation method is simplified to get:

Beneficial effect:

1. The demodulation method provided by the present invention and further simplified forms can fully meet the performance requirements of the IEEE 802.15.4 protocol. In the IEEE 802.15.4 protocol, it is stipulated that the signal-to-noise ratio is 5～6dB, and the PSDU is 20 bytes (160 bits), the packet error rate is less than 1%, that is, the PER is less than 1×10 ^-2 , especially when the signal-to-noise ratio is about 1dB, it can fully meet the requirements;

2. Compared with the traditional typical complex baseband non-coherent demodulation method, the present invention has lower computational complexity, lower energy consumption and lower implementation cost. The traditional typical complex baseband non-coherent demodulation method passes a division and An arc tangent operation is used to estimate the frequency offset information Nω ₀ T _c , as can be seen from formula (10), the estimation method of the frequency offset information Nω ₀ T _c of the present invention requires at most three comparisons, two multiplications, and one division and two addition operations; the estimation method of further simplified form of frequency offset information Nω ₀ T _c needs at most three comparisons, one multiplication, one division and two addition operations, so it has lower computational complexity and lower performance power consumption and lower implementation costs;

3, compared with the demodulation method of traditional simplified form, the present invention has higher reliability, and the demodulation method of traditional simplified form carries out approximate estimation to Nω ₀ T _c with formula (6), promptly uses 0, -π and Four kinds of phase pairs Nω ₀ T _c are estimated, and the errors caused are relatively large. The present invention uses formula (10) to estimate the frequency offset information Nω ₀ T _c . The "over-estimation" or "under-estimation" existing in the estimation process Weaker phenomenon, better performance, higher reliability.

Description of drawings

Figure 1 is a basic characteristic diagram of four frequency bands in the physical layer of the IEEE 802.15.4 protocol;

Figure 2 is a diagram of the IEEE 802.15.4 protocol 868/915/950-MHz frequency band physical layer data transmission process;

Figure 3 is a diagram of the IEEE 802.15.4 protocol 868/915/950-MHz frequency band spread spectrum mapping method;

Fig. 4 is a frame structure diagram of the IEEE 802.15.4 protocol physical layer;

Figure 5 is a structural diagram of a typical traditional non-coherent demodulation method applicable to the 868/915/950-MHz frequency band;

Figure 6 is a performance comparison diagram of the traditional demodulation method and the simplified demodulation method applicable to the 868/915/950-MHz frequency band;

Fig. 7 is two kinds of demodulation methods provided by the present invention and existing two kinds of typical demodulation method performance comparison diagrams;

FIG. 8 is a probability distribution diagram of the frequency offset offset f ₀ .

detailed description

Embodiments of the present invention will be specifically described below with reference to the accompanying drawings.

A high-reliability BPSK demodulation method for IEEE802.15.4. The data frame of the physical layer of the sending end is transmitted to the receiving end through a channel after spread spectrum and BPSK modulation. The data frame includes a 32-bit preamble and a physical layer service Data unit PSDU; the complex baseband sampling signal received by the receiving end is expressed as Where s(k) is the transmitted data to be detected, s(k)∈{+1,-1}, ω ₀ =2πf ₀ , f ₀ and θ are frequency offset and phase offset respectively and in the whole data frame Keeping unchanged, _T represents the spreading code chip period, and η ₀ (k) is the complex baseband additive white Gaussian noise; the concrete demodulation steps are:

Step 1. Use the channel received data corresponding to the 32-bit preamble to extract the frequency offset observation value Y including the frequency offset information:

Wherein, J represents the total number of bits of the preamble, J=32, N represents the spreading length, N=15, 1≤m≤J-1,

0≤n≤N-1, p[n+Nm] represents the channel reception value of the n-th chip corresponding to the m-th bit of the preamble, (·) ^* represents the conjugate operation, η ₁ represents all the noise item;

Step 2: Perform bit-level differential processing on the complex baseband sampling signal corresponding to the PSDU to obtain the judgment observation value A[m]:

Among them, r[n+Nm] represents the channel reception value of the n-th chip corresponding to the m-th bit of the PSDU, η ₂ [m] represents all noise items, and E[m] represents the m-th bit data sent ;

Step 3: Use the frequency offset observation value Y in step 1 to extract frequency offset information, compensate A[m] in step 2, and then perform detection and judgment:

in, Indicates the mth bit data obtained by detection and judgment, q( ) is a quantization function, and q(Y) is specifically expressed as:

Among them, |·| represents the modulo operation, represents the estimated value of the frequency offset Nω ₀ T _c in A[m];

Step 4, after the detection is completed, the received PSDU data is sent to the MAC layer for CRC check;

In said step three, The calculation formula is:

As a further simplified form, the quantization function described in step 3 is in:

The theoretical basis of the present invention is set forth below.

First, Taylor series expansion of tan ^-1 x at x=0 gives:

When |x| is small, there is an approximate relationship tan ^-1 x≈x.

When the frequency offset Nω ₀ T _c is small and the signal-to-noise ratio is large, The value is small. Formula (4) can be directly simplified by tan ^{- 1} x≈x to get:

But when the frequency offset Nω ₀ T _c is large or the signal-to-noise ratio SNR is not high, The value is larger. The approximate calculation of tan ^-1 x≈x to formula (4) will bring a large error, resulting in a great loss of final performance. That is, formula (16) is only applicable to the detection process under the two constraints of small frequency offset and large signal-to-noise ratio.

In order to obtain a low-complexity frequency offset estimation method under unconstrained conditions, we use the equivalent relationship between the arcsine function and the arctangent function to obtain:

If Y falls in the first and fourth quadrants of the complex plane coordinate system, that is Then there are:

If Y falls in the second and third quadrants of the complex plane coordinate system, ie Then there are:

According to formula (17) and formula (18), formula (4) can be equivalently transformed into:

Second, the Taylor series expansion of sin ^-1 x at x=0 gives:

Then when |x| is small, there is an approximate relationship sin ^-1 x≈x.

because It is constant under any frequency offset and signal-to-noise ratio conditions, so the sin ^-1 x≈x can be directly used in the equation (19) Items can be simplified to get:

Third, the calculation process of the fraction of formula (20) can be considered to be further simplified. When the SNR is large, we have the following approximate relationship:

The 4th, use formula (21) only to the molecule Im (Y) of formula (20) fractional formula and can get after approximation:

As shown in FIG. 7 , the demodulation method provided by the present invention has superior performance and higher reliability compared with the traditional typical simplified form demodulation method. The carrier frequency used in the simulation is 924MHz, the frequency offset f ₀ obeys the triangular distribution shown in Figure 8, the phase offset θ obeys the uniform distribution in (0, 2π], and the data length of PSDU is 20 bytes (160 bit), gather at least 3000 frame errors under each signal-to-noise ratio. Visible, two kinds of demodulation method performances provided by the present invention are between two kinds of traditional demodulation methods.Compared with traditional simplified demodulation method, in packet error rate When it is 1×10-3, a gain of not less than 1.3dB can be obtained. Therefore, the two schemes provided by the present invention achieve a better balance between complexity and performance.

The theoretical basis for further simplifying the form is: using formula (21) to approximate the numerator Re(Y) and the denominator Im(Y) in the formula (20) at the same time:

The calculation process can be further simplified and energy consumption can be reduced.

Claims (2)

1. A high reliability BPSK demodulation method used in IEEE802.15.4, the data frame of the physical layer of the sending end is transmitted to the receiving end through the signal channel after spreading and BPSK modulation, the data frame includes the lead code of 32 bits and physical layer service data unit PSDU; the complex baseband sampling signal received at the receiving end is represented asWhere s (k) is the transmitted data to be detected, s (k) ∈ { +1, -1}, ω₀＝2πf₀,f₀And theta is eachFrequency offset and phase offset and remain constant throughout the data frame, T_cIndicating spreading code chip period, η₀(k) Complex base band additive white gaussian noise; the specific demodulation steps are as follows: step one, extracting a frequency offset observation value Y containing frequency offset information by using channel receiving data corresponding to a preamble of 32 bits:

wherein J represents the total number of bits of the preamble, J is 32, N represents the spreading length, N is 15, 1 ≦ m ≦ J-1, 0 ≦ N ≦ N-1, p [ N + Nm ≦ N-1]Denotes a channel reception value of the nth chip corresponding to the mth bit of the preamble, (· denotes a conjugation operation, η₁Represents all noise terms;

step two, carrying out bit-level differential processing on the complex baseband sampling signals corresponding to the PSDU to obtain a judgment observation value A [ m ]:

wherein r [ n + Nm]Indicating the channel reception value of the nth chip corresponding to the mth bit of the PSDU, η₂[m]Representing all noise terms, E [ m ]]Represents the mth bit data transmitted;

step three, extracting frequency offset information by using the frequency offset observation value Y in the step one, compensating the A [ m ] in the step two, and then performing detection judgment:

wherein,the mth bit data obtained by detection decision is represented, q (-) is a quantization function, and q (y) is specifically represented as:

wherein, | - | represents a modulo operation,is represented by A [ m ]]Medium frequency offset N omega₀T_cAn estimated value of (d);

after the detection is finished, transmitting the received PSDU data to an MAC layer for CRC;

the method is characterized in that:

in the third step, the first step is that,the calculation formula of (2) is as follows:

where Re (·) represents the real part operation, and Im (·) represents the imaginary part operation.

2. The high reliability BPSK demodulation method for ieee802.15.4 of claim 1, wherein: for the third stepThe calculation method of (a) is simplified to obtain:

。