Channel estimation method with embedded pilot assisted by guard band for OTFS system

Abstract

The application relates to a channel estimation method of an OTFS system without a protection band embedded pilot assistance, and belongs to the technical field of wireless communication, and comprises the following steps: S1: in an orthogonal time-frequency-space modulation system (OTFS) system, data symbols are mapped into a delay-Doppler (DD) domain, a pilot structure is inserted in the data symbols in the DD domain, and a received signal under the DD domain is obtained by using an OTFS system model; and S2: according to the obtained received signal, an iterative interference cancellation technology is carried out at a receiving end, and estimation of a channel is completed while data detection is completed. The application can effectively improve the comprehensive performance of the system.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wireless communication, and relates to a channel estimation method of an OTFS system without a guard band embedded pilot assisted. BACKGROUND

[0002] An orthogonal frequency division multiplexing (OFDM) system flexibly copes with a frequency selective fading channel through adaptive modulation of each subband, and has good anti-multipath time delay capability. However, a future 6G communication network will realize an air, space and ground integrated communication target when facing scenes such as low-orbit satellites, aircraft and high-speed railways. This means that a large number of dynamic communication nodes will exist in the future communication network, and higher requirements are put forward for the robustness of a mobile communication system. When facing a high mobility scene, channel fading will become time-varying, that is, color dispersion occurs in both time and frequency dimensions. At the same time, with the increase of a communication frequency, especially in a terahertz frequency band, high Doppler spread caused will cause more serious damage to the orthogonality of subcarriers in the current OFDM system, and then making up for the defect of the OFDM system will become a key problem of 6G research.

[0003] In 2017, R. Hadani et al. proposed an orthogonal time-frequency space modulation (OTFS) technology, and the essential difference between the OTFS technology and an OFDM system is that the OTFS technology maps communication transmission from a time-frequency (TF) domain to a delay-Doppler (DD) domain, so that a wireless channel presents a series of excellent properties such as stability and sparsity, and is used to resist a double dispersion phenomenon of a wireless channel. As a new generation of communication system, researches on the OTFS system mainly focus on basic theories, key technologies and application scenes. As a basis for key technologies such as data detection and multiple access, channel estimation in the DD domain is particularly important. Since DD domain data can obtain time-frequency full diversity, the main target of estimation is to obtain an attenuation coefficient and a delay-Doppler parameter of a channel in the DD domain.

[0004] Since OTFS system was proposed, many methods have been tried to complete channel estimation in DD domain, such as based on compressed sensing, based on deep learning and based on Bayesian theory. Their core idea is to insert a small number of pilots to complete channel estimation by using the sparse characteristics of DD domain channel. Most researches only insert one pilot in DD domain, and then derive two different pilot configuration schemes, namely embedded and superimposed. Embedded embeds a single pilot in the transmitted DD domain data, and sets a guard interval between the pilot and the data to isolate the interference of DD domain data on the pilot information. This method can complete channel estimation at one time, making the algorithm simple and efficient, but it needs to adjust the guard interval according to the channel, which limits the effective use of the frequency band. Superimposed superimposes a single high-power pilot with DD domain data, and introduces iterative interference cancellation technology to separate the pilot information at the receiving end. Without setting a guard interval, it achieves similar estimation performance as embedded. This method performs channel estimation in two steps, the first step estimates the delay and Doppler characteristic parameters, and the second step estimates the channel attenuation coefficient. In summary, in the research of OTFS channel estimation, embedded pilot and superimposed pilot design have their own advantages and disadvantages. SUMMARY

[0005] Therefore, the present application aims to provide a pilot-aided channel estimation method without guard interval, which can not only cancel the guard interval and improve the spectrum utilization, but also reduce the pilot power.

[0006] To achieve the above object, the present application provides the following technical scheme:

[0007] A pilot-aided channel estimation method without guard interval in an OTFS system, comprising the following steps:

[0008] S1: In an OTFS system, data symbols are mapped to a DD domain, a pilot structure is inserted in the data symbols in the DD domain, and the OTFS system model is used to obtain a received signal in the DD domain;

[0009] S2: According to the obtained received signal, iterative interference cancellation technology is performed at the receiving end to complete data detection and channel estimation at the same time.

[0010] Further, the DD domain resource is divided into a grid of M x N, where there are M rows of data in the delay dimension and N columns of data in the Doppler dimension, and a data symbol sequence of length M x N is mapped into it to form the DD domain transmitted signal x[k, l]; the transmitted signal is transformed from the DD domain to the TF domain by the Inverse Symmetric Finite Fourier Transform (ISFFT), as shown in the following formula:

[0011]

[0012] where 0≤n≤N-1, 0≤m≤M-1;

[0013] The multicarrier modulator converts the signal X[n, m] in the TF domain into a continuous-time waveform s(t) using a transmit waveform g tx (t), also known as the Heisenberg transform, as shown in the following formula:

[0014]

[0015] where Δf is the subcarrier spacing and T is the signal duration;

[0016] Further, after s(t) is transmitted through a multipath time-varying channel, the received signal is given by the following formula:

[0017] r(t) = ∫∫h(τ, v)s(t-τ)e j2πv(t-τ) dτdv + w(t)

[0018] where w(t) is additive white Gaussian noise and h(τ, v) is the DD domain channel response;

[0019] At the receiving end, the TF domain received symbol Y[n , m] is obtained by performing traditional multicarrier demodulation on r(t), also known as the Wigner transform; finally, the TF domain symbol Y[n, m] is transformed to the DD domain by the Symmetric Finite Fourier Transform (SFFT), as shown in the following formula:

[0020]

[0021] where 0≤n≤N-1, 0≤m≤M-1;

[0022] Further, since fewer parameters are required to model the channel in the delay-Doppler domain, the sparse representation of the channel h(τ, v) is:

[0023]

[0024] where P is the number of propagation paths, h i , τ i and v i denote the path gain, delay and Doppler shift associated with the ith path, respectively, and δ(·) denotes the impulse function. i , τ i , v i and δ(·) are the main targets of channel estimation; the delay and Doppler parameters of the ith path are mapped into the DD domain grid, and their mapping relationship is shown in the following formula:

[0025]

[0026] where l τi and k vi denote the delay tap and Doppler tap of the ith path, respectively, and satisfy the constraint condition TΔf = 1.

[0027] Further, considering a specific DD domain transmitted signal x[k, l] and received signal y[k, l], assuming that the transmitted waveform is ideal, i.e., it satisfies biorthogonality, then the input-output relationship in the DD domain is:

[0028]

[0029] where (.) x denotes the modulo operation of the divisor x, h eff is the DD domain equivalent channel, and w(k, l) is Gaussian white noise with a mean of 0 and a variance of σ 2 . Using the sparse representation of the channel, the input-output relationship is further written as:

[0030]

[0031] Further, as in the conventional OFDM system, the pilot design needs to consider the balance between the pilot placement at the transmitting end and the channel estimation performance. For a given channel maximum delay tap parameter l max and maximum Doppler tap parameter k max , the embedded pilot scheme uses the following formula to describe the characteristic parameters of the pilot:

[0032]

[0033] where 0 denotes a guard interval, subscript p denotes a pilot, subscript d denotes data, x p denotes a pilot symbol, x d denotes a data symbol, k p and l p denote the positions of the pilot placement, respectively.

[0034] The present application cancels the restriction of the guard interval, and the design method is improved as follows:

[0035]

[0036] Compared with the superposition type, the pilot design of the present application does not superimpose data at the pilot, which can reduce the interference between the superposition pilot and the data. However, because the guard interval is cancelled, the scheme must consider the interference of the surrounding data on the pilot information.

[0037] Further, there will be interference of data near the pilot at the receiving end. Because the time delay parameter is generally positive, the Doppler parameter can be positive or negative, so the part of data will only affect the positive direction along the time delay axis and the positive and negative directions along the Doppler axis, and the range is the maximum time delay parameter l max and the maximum Doppler parameter k max . At the same time, this part will not only be used for data detection, but also be used for channel estimation, while the other part is only used for data detection.

[0038] According to the pilot structure and combining the input-output relationship formula of the OTFS system in the DD domain, the following formula is obtained in the red area at the receiving end:

[0039] y[k,l]=P k,l +I k,l +w[k,l]

[0040] Where the received signal is composed of three parts of pilot response P k,l , data interference I k,l and Gaussian noise w[k,l], wherein the pilot response P k,l , data interference I k,l is as follows:

[0041] P k,l =x p [k p ,l p ]h eff [(k-k p ) N ,(l-l p ) M ]

[0042]

[0043] Wherein, x p represents the pilot symbol, x d represents the data symbol, k p and l p represent the position of the pilot placement, k v and l τ represent the position of the data.

[0044] Since the pilot needs to reflect the unknown channel characteristic parameters as a known signal, it is determined that the pilot power cannot be set too low, so it is considered here that the pilot response is larger than the energy of the data. However, considering the sparsity of the DD domain channel and the two-dimensional convolution characteristics of the input and output, the pilot response will only appear discretely in the affected area, and the number of affected grids is related to the channel sparsity P.

[0045] Further, at the receiving end, a channel estimation algorithm is proposed according to the iterative interference cancellation technology, including:

[0046] Pre-estimation, used for channel delay and Doppler parameter estimation;

[0047] Iterative part, used for estimation of channel attenuation coefficient, including interference cancellation, data detection and updating estimation;

[0048] First, the interference cancellation uses the estimated channel attenuation to cancel the pilot signal in the received signal; then the signal is sent into the data detector to obtain the transmitted data; finally, the estimation is updated according to the transmitted data and the pilot signal.

[0049] The beneficial effects of the present application are that: based on the OTFS system, a channel estimation algorithm is proposed, through the comprehensive design of the pilot, the protection interval of the embedded pilot is broken through, and the pilot power of the superimposed pilot is reduced under the premise of ensuring the channel estimation performance. For the pilot structure of the existing OTFS system, the method proposed in the present application can effectively improve the comprehensive performance of the system.

[0050] Other advantages, objects and features of the present application will be set forth in part in the following specification, and in part will become apparent to those skilled in the art from the following, or can be learned from the practice of the present application. The objects and other advantages of the present application can be realized and obtained by the following description. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to make the objects, technical solutions and advantages of the present application clearer, the preferred detailed description of the present application will be made below in combination with the drawings, in which:

[0052] Figure 1 It is a schematic diagram of OTFS system structure;

[0053] Figure 2 It is a pilot design comparison diagram, in which (a) is a conventional embedded pilot design, and (b) is a pilot design of the present embodiment;

[0054] Figure 3 It is a receiving end interference schematic diagram;

[0055] Figure 4 It is an algorithm framework schematic diagram;

[0056] Figure 5 for estimating performance comparison chart;

[0057] Figure 6 for NMSE performance comparison chart;

[0058] Figure 7 for iteration performance diagram;

[0059] Figure 8 for BER performance comparison chart. DETAILED DESCRIPTION

[0060] The present application can be implemented or applied in other different embodiments, and the details in the specification can be modified or changed based on different views and applications without departing from the spirit of the present application. It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present application in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0061] The accompanying drawings are only used for exemplary illustration, and the representation is only a schematic diagram, not a physical diagram, and should not be understood as a limitation of the present application; in order to better illustrate the embodiments of the present application, some components in the drawings are omitted, enlarged or reduced, and do not represent the actual product size; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings can be omitted.

[0062] The same or similar reference numerals in the drawings of the embodiments of the present application correspond to the same or similar components; in the description of the present application, it should be understood that if the terms "upper", "lower", "left", "right", "front", "back" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, therefore the terms describing the positional relationship in the drawings are only used for exemplary illustration, and should not be understood as a limitation of the present application, for those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0063] Please refer to Figures 1-8 , the present application provides an OTFS system channel estimation algorithm. The algorithm starts from pilot design, designs a new structure, and improves the overall performance of the system. In view of the interference problem caused by the pilot structure, the present application uses iterative interference cancellation technology to eliminate data interference at the receiving end, and realizes accurate estimation of the channel.

[0064] 1. Receive signal generation

[0065] To ensure compatibility with the OFDM system, the OTFS system adds signal processing modules before and after the OFDM system. Figure 1 The figure shows a block diagram of an uncoded single-input single-output OTFS system. Figure 2 (b) generates the transmission signal x[k,l]( Figure 2 (a) in the figure is for comparison only). The transmitted signal obtains a continuous-time waveform through the ISFFT and Heisenberg transform. The receiving end obtains the waveform after passing through the channel and obtains the interfered received waveform by performing Wigner and SFFT transforms on the waveform. Obviously, the receiving end is the inverse process of the transmitting end, where the ISFFT and SFFT transforms establish the connection between the DD domain and the TF domain. For ease of understanding, it can be simply considered here that based on the ISFFT in the DD domain, a Fourier transform is performed along the delay axis to obtain the frequency axis, and an inverse Fourier transform is performed on the Doppler axis to obtain the time axis, and vice versa. In this way, the received signal in the DD domain can be obtained. However, the pilot of the received signal at this stage is affected by the data signal, so iterative interference cancellation is required to obtain accurate channel estimation.

[0066] 2. Iterative interference elimination

[0067] At the receiving end, in order to eliminate the aliasing interference between the pilot information and the surrounding data, the present invention proposes an iterative interference elimination technology. Figure 4 The channel estimation algorithm shown in Figure 1 is divided into two parts. The first part is the pre-estimation part, which mainly estimates the channel delay and Doppler parameters. The second part is the iterative part, which mainly estimates the channel attenuation coefficient. It consists of interference cancellation, data detection, and updated estimation. First, the interference cancellation part uses the estimated channel attenuation to remove the pilot signal from the received signal. Then, the signal is fed into the data detector to obtain the transmitted data. Finally, the estimate is updated based on the transmitted data and the pilot signal. This iterative process completes the estimation of the channel attenuation.

[0068] 2.1 Pre-estimation and Interference Cancellation

[0069] The pre-estimation phase needs to complete the estimation of channel delay and Doppler parameters. This phase effectively utilizes the two-dimensional convolution characteristics of the OTFS system to determine the delay and Doppler parameters related to the DD channel. Due to the large difference between the pilot power and data energy, based on this characteristic, a threshold can be set for detection. The detection threshold is set to E s is the symbol energy. If |y(k,l)|≥γ, then it is considered that there is a DD domain path at this location, and the interference area is as follows Figure 3 shown.

[0070] In Figure 4 interference cancellation requires the received signal y(k, l) and the equivalent channel Two parameters, since the delay and Doppler parameters have been determined in the pre-estimation stage, only the channel attenuation is unknown, so an initial channel attenuation value is needed before the iteration begins. Because the pilot power is usually set large, the data interference I k,l and the noise w[k, l] together as noise, a rough estimate of the channel gain can be obtained, as shown in the following formula:

[0071]

[0072] The superscript Λ represents the estimated value, so represents the estimated value of h eff . To obtain a more accurate channel estimate, steps 1 and 2 shown in the following formula are needed to eliminate pilot interference and data detection. Figure 4 The pilot interference elimination formula is as follows.

[0073]

[0074] The next step is to send the received signal after pilot interference elimination to the data detector.

[0075] 2.2 Data detection

[0076] After step 1, step 2 data detection is needed for the received signal. Compared with the traditional ZF (Zero-Forcing) and MMSE (Minimum Mean Square Error) detection methods, the MP (Message Passing) algorithm in the DD domain reduces complexity while improving detection performance by constructing a factor graph, so the MP detector is used as the data detection scheme in this invention. The received signal is rewritten as follows:

[0077]

[0078] The pre-estimated value of the channel is used in step 1 to eliminate the pilot response. Whether it is the pre-estimation at the beginning of the iteration or the estimation during the iteration, there is a small error, so the impact of the estimation error is considered as part of the Gaussian noise, and is used to replace the original noise w[k, l]. The vectorized received signal is represented as follows:

[0079]

[0080] where, is a complex vector with a dimension of NM x 1, where the elements are represented as 1≤d≤MN. is an NMxNM complex matrix, is an information vector of dimension NMx1, where the elements are denoted as 1≤c≤MN. According to equation (17), the joint Maximum a Posteriori Probability (MAP) detection formula is as follows:

[0081]

[0082] Since it is difficult to deal with the joint MAP detection for the actual values of N and M, the symbol-by-symbol MAP detection is considered, as shown in the following equation:

[0083]

[0084] where A is the set of modulation symbols and it is assumed that all transmission symbols are transmitted with equal probability. Since has sparsity, it is further simplified as follows, assuming that y is independent of the transmitted symbol x[c] with each other:

[0085]

[0086] Thus, the data detection is completed according to the above equation in combination with the MP algorithm. After eliminating the interference of the pilot, the data detection problem is converted into a simple data detection problem, and the specific algorithm steps and message calculation method will not be described here.

[0087] 2.3 Update of channel estimation

[0088] Step 3 is the update of the channel gain estimation value. The present application combines the idea of adaptive signal processing to complete the update of the channel gain by constructing an expected signal. Due to the randomness of the noise, an approximate processing is performed here. The received signal is composed of three parts, and the influence of the noise can be ignored because the pilot power is set to be large. Therefore, only the received signal composed of the pilot and the transmitted data is considered here. The expected signal y e [k, l] is reconstructed according to the estimation value.

[0089]

[0090] The estimation error is obtained, as shown in the following equation:

[0091]

[0092] Thus, the channel estimation update equation is obtained

[0093]

[0094] The updating of the channel estimation is completed according to the updating formula, and the iteration step is executed again with the received signal, so that the channel estimation and data detection are completed at the same time.

[0095] 3. Simulation experiment

[0096] 3.1 Experimental data and settings

[0097] The OTFS system is built by using the Matlab software.

[0098] Table 1

[0099]

[0100] The maximum moving speed in this scenario is 150 km / h, and the maximum Doppler index k max = 5 is obtained. The channel gain obeys a Gaussian distribution with a mean of 0 and a variance of 1 / P, and the channel number P is 4. SNR = E S / N0 is defined as the signal-to-noise ratio of the transmitted symbol, and is referred to as the signal-to-noise ratio below. SNR p = E P / N0 is defined as the signal-to-noise ratio of the pilot symbol, where E p , E s are the pilot and data symbol energies, respectively. In order to illustrate the accuracy of the channel estimation, the normalized mean square error (NMSE) is used to describe the channel estimation accuracy, and is defined as shown in the following formula.

[0101]

[0102] 3.2 Performance analysis

[0103] The embedded and superimposed pilot estimation algorithms are compared. The performance of the pre-estimation of the present scheme and the superimposed pilot is compared first, then the estimation accuracy and convergence of the three schemes are compared and analyzed, and finally the experimental analysis of the bit error performance of the present scheme is given.

[0104] 3.2.1 Performance analysis of pre-estimation

[0105] Both the present scheme and the superimposed pilot have pilot pre-estimation, in order to illustrate the influence of different pilot structures on the pre-estimation performance, Figure 5 the pre-estimation performance of the two schemes under different signal-to-noise ratios is given. It is not difficult to see that the pre-estimation performance decreases in the range of 10 dB to 16 dB of the signal-to-noise ratio SNR, because under the condition of fixed pilot signal-to-noise ratio, with the increase of the signal-to-noise ratio, the interference of the data symbol to the pilot symbol increases, affecting the pre-estimation performance. The embedded pilot does not require pre-estimation operation, but requires SNR p35dB, and 40dB. The superposition scheme uses pilot and data superposition configuration, and the SNR p is set to at least 40dB. The scheme of the present application can obtain similar estimation results when the SNR p is set to 38dB. This shows that, compared with the superposition pilot, the scheme of the present application weakens the requirement for the SNR p .

[0106] 3.2.2 NMSE performance analysis

[0107] Figure 6 The estimation performance of the three schemes under different SNRs is given. As shown in the figure, when the SNR increases from 10dB to 16dB, the errors of the three schemes are basically stable. The scheme of the present application completes estimation when the SNR p is 38dB, and the performance is in the middle. Compared with the embedded scheme, the scheme of the present application cancels the guard interval; compared with the superposition scheme, the scheme of the present application weakens the requirement for the SNR p .

[0108] 3.2.3 Analysis of algorithm convergence

[0109] In order to illustrate the convergence of the algorithm, Figure 7 the influence of the number of iterations on the NMSE performance is shown when other conditions are unchanged. It can be seen that when the SNR is 13dB, after one iteration, the NMSE curve has a significant decline, and then converges quickly. This shows that the use of the detection result of the data detector can quickly eliminate interference, thereby improving the estimation accuracy of the iterative algorithm. According to the figure, the NMSE performance of the scheme of the present application and the superposition scheme is basically stable after 3 iterations, so the number of iterations of the algorithm is set to 3. The embedded scheme does not use the iterative algorithm, and the curve is only illustrative. Figure 7

[0110] 3.2.4 System bit error rate performance analysis

[0111] In order to illustrate the performance of the detector, Figure 8 the bit error rate performance comparison chart of the MMSE detector and the MP detector under different SNRs is given. As shown in the figure, in the process of increasing the SNR from 10dB to 16dB, the bit error rates of the two detectors are in sharp decline, but the MP detector shows better performance than the MMSE. Even in the non-ideal state, the performance of the MP detector is not much different from that in the ideal state.

[0112] ​Finally, it is to be explained that the above embodiments are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the purpose and scope of the technical solutions, and all should be covered in the scope of the claims of the present application.

Claims

1. A channel estimation method for an OTFS system without guard band embedded pilot assistance, characterized by: The following steps are involved: S1: In the Orthogonal Time-Frequency Space Modulation (OTFS) system, data symbols are mapped into the Delay-Doppler (DD) domain. A pilot structure is inserted into the DD domain data symbols, and the OTFS system model is used to obtain the received signal in the DD domain. S2: Based on the received signal, iterative interference cancellation technology is performed at the receiving end to complete the channel estimation while completing data detection; For a given channel maximum delay tap parameter l max and the maximum Doppler tap parameter k max , the embedded pilot scheme uses the following formula to describe the characteristic parameters of the pilot: Where 0 represents the guard interval, subscript p represents the pilot, subscript d represents the data, and x p represents the pilot symbol, x d Represents the data symbol, k p and l p Respectively represent the locations where the pilots are placed; Remove the guard interval restriction as follows: At the receiving end, a channel estimation algorithm is proposed based on iterative interference cancellation technology, including: Pre-estimation, used to estimate channel delay and Doppler parameters; The iterative part is used to estimate the channel attenuation coefficient, including interference elimination, data detection and update estimation; First, interference cancellation uses the estimated channel attenuation to remove the pilot signal from the received signal; then the signal is fed into a data detector to obtain the transmitted data; finally, the estimate is updated based on the transmitted data and the pilot signal.

2. The OTFS system non-guard-band embedded pilot-assisted channel estimation method according to claim 1, characterized in that: The DD domain resources are divided into an M×N grid, where the delay dimension has M rows of data and the Doppler dimension has N columns of data. An M×N data symbol sequence is mapped into this grid to form the DD domain transmit signal x[k,l]. The transmit signal is transformed from the DD domain signal to the TF domain using the inverse symplectic Fourier transform (ISFFT), as shown in the following equation: Among them, 0≤n≤N-1, 0≤m≤M-1; The multi-carrier modulator uses the transmit waveform g tx (t), convert the signal X[n,m] in the TF domain into a continuous-time waveform s(t), as shown in the following formula: Where Δf is the subcarrier spacing and T is the signal duration.

3. The OTFS system channel estimation method without guard band embedded pilot assistance according to claim 1, characterized in that: After s(t) is transmitted through a multipath time-varying channel, the received signal is given by: r(t)=∫∫h(τ,v)s(t-τ)e j2πv(t-τ) dτdv+w(t) Where w(t) is additive white Gaussian noise, h(τ,v) is the DD domain channel response; At the receiving end, the TF domain received symbol Y[n,m] is obtained by performing traditional multi-carrier demodulation on r(t); finally, the TF domain symbol Y[n,m] is transformed into the DD domain through the symplectic finite Fourier transform SFFT, as shown in the following equation: Among them, 0≤n≤N-1, 0≤m≤M-1.

4. The OTFS system channel estimation method without guard band embedded pilot assistance according to claim 1, characterized in that: The sparse representation of the channel h(τ,v) is: Where P is the number of propagation paths, h i , τ i and v i denote the path gain, delay, and Doppler shift associated with the ith path, δ(·) represents the impulse function, and h i , τ i , v i and δ(·) are the main objectives of channel estimation; the delay and Doppler parameters of the i-th path are mapped to the DD domain grid, and their mapping relationship is shown in the following formula: Among them, l τi and k vi denote the delay tap and Doppler tap of the i-th path respectively, satisfying the constraint condition of TΔf=1.

5. The channel estimation method for an OTFS system without guard band embedded pilot assistance according to claim 1, characterized in that: Consider a specific DD domain transmit signal x[k,l] and receive signal y[k,l], assuming that the transmit waveform is ideal, that is, it satisfies biorthogonality, then the input and output relationship in the DD domain is: in,(.) x Represents the modulo operation of the divisor x, h eff is the DD domain equivalent channel, w(k,l) has a mean of 0 and a variance of σ 2 Gaussian white noise, using the sparse representation of the channel, the input-output relationship is further written as:

6. The OTFS system channel estimation method without guard band embedded pilot assistance according to claim 1, characterized in that: According to the pilot structure and combined with the input and output relationship of the OTFS system in the DD domain, the following formula is obtained in the red area at the receiving end: y[k,l]=P k,l +I k,l +w[k,l] The received signal consists of the pilot response P k,l , data interference I k,l and Gaussian noise w[k,l] are aliased into three parts, where the pilot response P k,l , data interference I k,l As shown in the following formula: P k,l =x p [k p ,l p ]h eff [(k-k p ) N ,(l-l p ) M ] Among them, x p represents the pilot symbol, x d Represents the data symbol, k p and l p Respectively represent the location of the pilot placement, k v and l τ They represent the location of the data respectively; the pilot response will only appear discretely in the affected area, and the number of affected grids is related to the channel sparsity P.

Images