A semi-deterministic channel simulation method based on finite-difference time-domain method
By adopting a semi-deterministic channel simulation method based on the finite-difference time-domain method, the problem of lack of randomness in traditional FDTD simulation is solved, a channel model suitable for multi-dimensional environments is established, and the accuracy and applicability of channel simulation are realized.
Patent Information
- Application Number
- CN202111431082.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-11-29
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2041-11-29
AI Technical Summary
Traditional FDTD simulations lack randomness and fail to reflect channel characteristics.
A semi-deterministic channel simulation method based on the finite-difference time-domain method is adopted. Through full-space electromagnetic simulation and random fine-tuning of scattering objects, a semi-deterministic channel model applicable to one-dimensional, two-dimensional and three-dimensional environments is established.
It achieves accuracy and applicability in channel simulation, reflects the randomness of channels, and is suitable for channel calculations in complex media.
Smart Images

Figure CN114117860B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a semi-deterministic channel simulation technology based on a finite difference time domain method, and belongs to the technical field of computational electromagnetics and wireless communication channel models. Background Art
[0002] As people's demand for mobile communication performance increases day by day, in order to cope with the future needs of massive mobile data and the connection of various electronic devices, and to further improve the performance of communication systems, channel simulation will definitely move towards a more precise direction.
[0003] Channel simulation is usually divided into statistical models, deterministic models, and semi-deterministic models. Statistical models are usually based on simple random mathematical formulas, and only simple inputs are needed to get results, such as the Rayleigh channel model and the Rice channel model. Deterministic models are based on electromagnetic theory and usually use ray tracing or the finite difference time domain method (FDTD) for simulation. FDTD is more accurate. Semi-deterministic models, such as the spatial channel model (SCM), perform statistical analysis on each path of the scatterer and finally superimpose each path at the receiving point.
[0004] FDTD simulation is a deterministic model, which can provide full information of the signal for the entire area. However, as a channel, complete determinism will limit its application. In order to reflect the channel characteristics, we need to introduce the randomness of scatterers and establish a semi-deterministic channel model through analysis. Summary of the invention
[0005] Technical problem: In order to solve the problem that traditional FDTD simulation is not random and difficult to reflect channel characteristics, the present invention proposes a semi-deterministic channel simulation method based on the traditional finite-difference time-domain method. This method uses full-space electromagnetic simulation and can be applied to one-dimensional, two-dimensional and three-dimensional environments. The method has accurate calculation results and wide applicability, and can also be used for channel calculations in complex media.
[0006] Technical solution: A semi-deterministic channel simulation method based on finite-difference time-domain method of the present invention comprises the following steps:
[0007] Step 1. According to the requirements, set the center frequency and bandwidth to be solved, and calculate the corresponding wavelength according to the relationship between frequency and wavelength;
[0008] Step 2. Set the corresponding solution area according to the environment to be simulated, determine the size of the required space, determine the processing method of the space boundary, determine the position of the scattering object in the space and its corresponding electromagnetic constant, determine the excitation point and the receiving point; then discretize the space according to the Yee grid, determine the number of grids in the space, the grid position of the scattering object, and the grid where the excitation point and the receiving point are located; finally, obtain the maximum time step Δt according to the CFL stability condition CFL , and set it to be no greater than Δt for the simulation CFL The time step and the total number of time steps for the simulation are determined;
[0009] Step 3. According to the number of Yee grids, apply for memory space for each electromagnetic field component and initialize it, and then save and initialize the iteration coefficients needed in the update iteration equation in an indexed manner;
[0010] Step 4. Update and iterate the electromagnetic field in the entire solution area. First, update the magnetic field value in the area. After the magnetic field calculation is completed, add the magnetic excitation source and process the magnetic field on the boundary. Then update the electric field value, add the electric excitation source and process the electric field on the boundary. Finally, save the signal at the receiving point as the output at that moment. Calculate the magnetic field and electric field continuously and stop after the set total time step number, and finally obtain the receiving signal at the receiving point for the entire time.
[0011] Step 5. Multiply the received signal by Perform frequency shifting, and then pass the frequency shifted signal through a cutoff frequency of f c , a low-pass filter with an amplitude of 2 obtains the corresponding envelope signal;
[0012] Step 6. Cross-correlate the received signal envelope with the transmitted signal envelope, and use correlation peak detection to obtain the complex attenuation and corresponding delay of each path contained in the received signal;
[0013] Step 7. In order to reflect the random fluctuation of the channel during mobile communication, the scatterer is fine-tuned, and then the process of steps 2-6 is repeated for multiple solutions;
[0014] Step 8. Perform statistical analysis on the results of multiple solutions to obtain the attenuation and delay statistics of each multipath component at the receiving point, thereby obtaining a semi-deterministic channel model.
[0015] in:
[0016] The scattering objects are fine-tuned as described in step 7, and the attenuation and delay statistics of the multipath components are analyzed through multiple FDTD simulations.
[0017] The iteration coefficients described in step 3 are stored in an indexed manner, which saves memory space and improves memory utilization efficiency.
[0018] In step 8, the results of multiple solutions are statistically analyzed to obtain the average value and standard deviation.
[0019] The discrete time step Δt
[0020]
[0021] The discrete time step Δt needs to be less than or equal to the maximum time step Δt CFL , where Δx is the discrete grid size in the x direction, Δz is the discrete grid size in the z direction, and c is the speed of light.
[0022] In step 4, the electromagnetic field is updated and iterated, wherein the magnetic field H y The update formula is:
[0023]
[0024] Among them, i and k are the spatial discrete coordinates in the x direction and the z direction respectively, n is the current time step, and the coefficient c is hyhy ,c hyez ,c hyex as follows
[0025]
[0026]
[0027]
[0028] Where μ is the magnetic permeability, σ m is the magnetic permeability coefficient, Δx is the discrete grid size in the x direction, and Δz is the discrete grid size in the z direction.
[0029] In step 4, the electromagnetic field is updated and iterated, wherein the electric field E x and E z The update formula is:
[0030]
[0031]
[0032] Coefficient c exex ,c exhy ,c ezez ,c ezhy as follows
[0033]
[0034]
[0035]
[0036]
[0037] Among them, ε is the dielectric constant and σ is the conductivity.
[0038] In step 4, the received signal envelope is cross-correlated with the transmitted signal envelope, and the envelope cross-correlation formula is:
[0039]
[0040] Among them, s l (t) is the complex envelope of the transmitted signal, r l (t) is the complex envelope of the received signal, R rs (τ) is r l (t) and s l The cross-correlation function of (t), R ss (τ) is s l (t) autocorrelation function, α n is the complex attenuation coefficient of the nth path, τ n is the transmission delay of the nth path.
[0041] Beneficial effects: The present invention establishes a semi-deterministic channel simulation model based on the finite difference time domain method, which can be applied to one-dimensional, two-dimensional and three-dimensional environments. The method adopts electromagnetic simulation of the whole space, and the calculation results are accurate and widely applicable. It can also be used for channel calculation under complex media. The randomness of scatterers is added to obtain a semi-deterministic channel model, which reflects the random characteristics of the channel. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Figure 1 It is a flow chart of a semi-deterministic channel simulation method based on finite-difference time-domain method of the present invention;
[0043] Figure 2 A schematic diagram of a simulation environment of an embodiment;
[0044] Figure 3 for Figure 2 Relative amplitude of multipath components at the middle receiving point 1 (15m, 2m);
[0045] Figure 4 for Figure 2 Relative amplitude of multipath components at the middle receiving point 2 (15m, 7m);
[0046] Figure 5 for Figure 2 Relative amplitude of multipath components at the middle receiving point 3 (15m, 10m);
[0047] Figure 6 for Figure 2 Relative amplitude of multipath components at the middle receiving point 4 (15m, 13m);
[0048] Figure 7 for Figure 2 Relative amplitude of multipath components at the middle receiving point 5 (15m, 20m);
[0049] Figure 8 for Figure 2 Relative amplitude of multipath components at receiving point 6 (15m, 25m). DETAILED DESCRIPTION
[0050] The present invention proposes a semi-deterministic channel simulation method based on the traditional finite-difference time-domain method. The specific scheme is as follows:
[0051] 1. Set the center frequency and bandwidth to be solved as needed, and calculate the corresponding wavelength based on the relationship between frequency and wavelength.
[0052] 2. Set the corresponding solution area according to the environment to be simulated, determine the size of the space to be solved, determine the processing method of the space boundary, determine the position of the scattering object in the space and its corresponding electromagnetic constant, determine the excitation point and the receiving point. Then discretize the space according to the Yee grid, determine the number of grids in the space, the grid position of the scattering object, and the grid where the excitation point and the receiving point are located. Finally, the maximum time step is calculated by the CFL stability condition, and a time step not greater than this number is set for the simulation, and the number of time steps required for simulation is determined.
[0053] 3. After step 2 is completed, apply for memory space for each electromagnetic field component and initialize it according to the number of Yee grids. Then save and initialize the coefficients needed in the update iterative equation in an indexed manner.
[0054] 4. Update and iterate the electromagnetic field. First, update the magnetic field value in the area. After the magnetic field calculation is completed, add the magnetic excitation source and process the magnetic field on the boundary. Then update the electric field value, add the electric excitation source, and process the electric field on the boundary. Finally, save the signal at the receiving point as the output at that moment. Through the continuous iteration of the magnetic field and electric field, stop calculating after the set total time step. Finally, the receiving signal at the receiving point for the entire time can be obtained.
[0055] 5. Multiply the received signal by The frequency is shifted and then passed through a cutoff frequency of f c , a low-pass filter with an amplitude of 2 can obtain the corresponding envelope signal.
[0056] 6. Cross-correlate the received signal envelope with the transmitted signal envelope, and use correlation peak detection to obtain the complex attenuation and corresponding delay of each path contained in the received signal.
[0057] 7. In order to reflect the random fluctuation of the channel during mobile communication, the scatterer is fine-tuned, and then the process of steps 2-6 is repeated for multiple solutions.
[0058] 8. Perform statistical analysis on the results of multiple solutions to obtain the attenuation and delay statistics of each multipath component at the receiving point. Thus, a semi-deterministic channel model is obtained.
[0059] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
[0060] The embodiment performs channel analysis on receiving points at different heights in front of a building in a vertical plane, such as Figure 2 The specific implementation method is as follows:
[0061] 1. Set the solution frequency, bandwidth and calculate the response wavelength.
[0062] Assume the simulation center frequency f c is 3.5GHz, and the bandwidth BW is 200MHz. Therefore, the center frequency wavelength λ c =c / f c is 85.7mm, where c is the speed of light.
[0063] 2. Set the solution region and discretize it in space and time.
[0064] like Figure 2 As shown in the figure, the computational environment is 45m long and 30m high, including the ground and a building with a width of 15m and a height of 18m located on the ground as scatterers. The ground is replaced by a PEC boundary, the relative dielectric constant of the building is 5.31, the relative magnetic permeability is 1, and the electrical conductivity is 0.0899S / m. The excitation point is located at point (10m, 10m), and the observation points are located at (15m, 2m)(15m, 7m)(15m, 10m)(15m, 13m)(15m, 20m)(15m, 25m). The discrete grid size Δx=Δz=5mm in the x-direction and z-direction is taken, and the FDTD simulation is performed in TE mode, and the solution area is discretized according to the Yee grid. The number of spatial grids is N. x ×N z =9000×6000, the building is located at the grid coordinates [4000, 7000]×[0, 3600], the excitation point is located at the grid coordinates (2000, 2000), and the six observation points are located at the grid coordinates (3000, 400) (3000, 1400) (3000, 2000) (3000, 2600) (3000, 4000) (3000, 5000). According to the CFL stability condition of FDTD, as shown in formula (1):
[0065]
[0066] The discrete time step Δt needs to be less than or equal to Δt CFL , for this embodiment, Δt CFL is 1.1785×10 - 2 ns, here we take Δt = 0.99Δt CFL =1.1667×10 -2 ns. According to the calculation environment, the length is 45m and the height is 30m, so its diagonal length is 70.71m. It takes 20,000 time steps for the electromagnetic wave to propagate with the above time step Δt. Under normal circumstances, the total number of time steps should not be less than this value. In this embodiment, the total number of time steps is N t =30000 steps.
[0067] 3. Apply for memory space and initialize it.
[0068] In TE mode, the electromagnetic field includes three components: Ex, Ez and Hy. Each component requires a 9001×6001 double type memory space and is initialized to 0. The medium parameters change less and are relatively fixed at different positions. The index method can be used to save memory space. Apply for 9001×6001 unsigned char type memory space for the index of Ex, the index of Ez and the index of Hy respectively. The coefficients required for Ex are as shown in formula (2)
[0069]
[0070] Among them, ε is the dielectric constant and σ is the conductivity. Calculate the parameters at each position one by one, and access the index number corresponding to the current parameter group in the Ex index space. If a new parameter group appears, save it and assign a new index number. Finally, the index number of the parameter group at each position is stored in the index space, and the index number corresponds to a set of parameters that have been calculated by formula (2). In this way, the original two double-type memory spaces can be saved into one unsigned char-type memory space, greatly improving the memory efficiency. Next, similar operations are performed on the indexes of Ez and Hy. The coefficients required are shown in formulas (3) and (4).
[0071]
[0072]
[0073] Where μ is the magnetic permeability, σ m is the magnetic permeability coefficient.
[0074] 4. Update and iterate the electromagnetic field.
[0075] The excitation is taken as magnetic field Hy excitation, and its waveform is a sinusoidally modulated Gaussian pulse, as shown in formula (5):
[0076]
[0077] Where n = 1, ..., N t is the current time step, π is the circumference of a circle, τ and t 0 Take separately and 0.9τ.
[0078] In each time step, the following operations are performed in sequence:
[0079] (1) Magnetic field calculation
[0080]
[0081] Among them, i and k are the spatial discrete coordinates in the x direction and the spatial discrete coordinates in the z direction respectively, and the coefficient c hyhy ,c hyez ,c hyex As shown in formula (4).
[0082] (2) Magnetic field excitation treatment
[0083]
[0084] This step is only performed in the magnetic field excitation position.
[0085] (3) Magnetic field boundary condition processing
[0086] The magnetic field boundary is processed according to the PML magnetic field calculation method.
[0087] (4) Electric field calculation
[0088]
[0089]
[0090] The coefficients are as shown in equations (2) and (3).
[0091] (5) Electric field excitation treatment
[0092] This embodiment does not perform electric field excitation, so no processing is performed.
[0093] (6) Electric field boundary condition processing
[0094] The electric field boundary is processed according to the PML electric field calculation method.
[0095] (7) Save the value of the observation point at the current time.
[0096] In this way, N t= 30000 steps, the signal situation r at 6 observation points over time can be obtained i (t), i=1,2,3,4,5,6.
[0097] 5. Get the envelope signal.
[0098] Since the environment is a multipath channel, assume that the complex envelope signal of the transmitted signal is s l (t), then the complex envelope of the received signal r l (t) can be expressed as
[0099]
[0100] where α n is the complex attenuation coefficient of the nth path, τ n is the transmission delay of the nth path. The complex envelope signal and the original signal have the following relationship:
[0101]
[0102]
[0103] Among them, Re[] is the real part operator, j is the imaginary unit, and e is a natural constant. In order to obtain the envelope signal from the original signal, we have
[0104]
[0105] So we can transform equation (13) through a cutoff frequency of f c , a low-pass filter with an amplitude of 2 can obtain the corresponding envelope signal r l (t).
[0106] 6. Use correlation analysis to analyze each path component.
[0107] Cross-correlating the received signal envelope with the transmitted signal envelope yields
[0108]
[0109] Therefore, the cross-correlation function of the received signal envelope and the transmitted signal envelope is the autocorrelation function of the transmitted signal envelope. n And add the path complex attenuation α n Therefore, correlation peak detection can be used to determine the complex attenuation and delay information of each path.
[0110] 7. Fine-tune the scatterer and solve the problem again.
[0111] In order to reflect the random fluctuation of the channel during mobile communication, in this embodiment, the horizontal position and vertical height of the building are fine-tuned, and the fine-tuning amount is evenly distributed within ±2m. Repeat 2-6 again to obtain the fine-tuned result.
[0112] 8. Obtain the statistics for each path.
[0113] The attenuation and delay information obtained in the above steps is statistically analyzed. In this embodiment, only the mean and standard deviation of the attenuation and delay under each path are statistically analyzed. The results are shown in Table 1 and Table 2.
[0114] Table 1 Relative (*) amplitude statistics of complex attenuation coefficients of multipath components
[0115]
[0116]
[0117] *: Attenuation coefficient normalized to free space.
[0118] Table 2 Delay statistics of multipath components (unit: seconds)
[0119]
[0120] In order to more vividly display the multipath situation, the amplitude attenuation data in Table 1 can be plotted in the form of error bars as shown in Figures 3 to 8 shown.
[0121] The present invention establishes a semi-deterministic channel simulation model based on the finite difference time domain method, which can be applied to one-dimensional, two-dimensional and three-dimensional environments. The method performs electromagnetic simulation on the whole space to obtain the response between the transmitter and the receiver, and fine-tunes the scatterers to cause the fluctuation of the channel, thereby obtaining the multipath statistical characteristics of the channel.
[0122] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with this technology to understand the contents of the present invention and implement them accordingly. They cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made according to the spirit of the present invention should be included in the protection scope of the present invention.
Claims
1. A semi-deterministic channel simulation method based on finite-difference time-domain method, characterized in that The method comprises the following steps: Step 1. According to the requirements, set the center frequency and bandwidth to be solved, and calculate the corresponding wavelength according to the relationship between frequency and wavelength; Step 2. Set the corresponding solution area according to the environment to be simulated, determine the size of the required space, determine the processing method of the space boundary, determine the position of the scattering object in the space and its corresponding electromagnetic constant, determine the excitation point and the receiving point; then discretize the space according to the Yee grid, determine the number of grids in the space, the grid position of the scattering object, and the grid where the excitation point and the receiving point are located; Finally, the maximum time step Δt is obtained from the CFL stability condition CFL , and set it to be no greater than Δt for the simulation CFL The time step and the total number of time steps for the simulation are determined; Step 3. According to the number of Yee grids, apply for memory space for each electromagnetic field component and initialize it, and then save and initialize the iteration coefficients needed in the update iteration equation in an indexed manner; Step 4. Update and iterate the electromagnetic field in the entire solution area. First, update the magnetic field value in the area. After the magnetic field calculation is completed, add the magnetic excitation source and process the magnetic field on the boundary. Then update the electric field value, add the electric excitation source and process the electric field on the boundary. Finally, save the signal at the receiving point as the output at that moment. Calculate the magnetic field and electric field continuously and stop after the set total time step number, and finally obtain the receiving signal at the receiving point for the entire time. Step 5. Multiply the received signal by Perform frequency shifting, and then pass the frequency shifted signal through a cutoff frequency of f c , a low-pass filter with an amplitude of 2 obtains the corresponding envelope signal; Step 6. Cross-correlate the received signal envelope with the transmitted signal envelope, and use correlation peak detection to obtain the complex attenuation and corresponding delay of each path contained in the received signal; Step 7. In order to reflect the random fluctuation of the channel during mobile communication, the scatterers are fine-tuned, and the horizontal position and vertical height of the building are fine-tuned. The fine-tuning amount is evenly distributed within ±2m, and then the process of steps 2-6 is repeated for multiple solutions; Step 8. Perform statistical analysis on the results of multiple solutions to obtain the attenuation and delay statistics of each multipath component at the receiving point, thereby obtaining a semi-deterministic channel model.
2. The semi-deterministic channel simulation method based on finite-difference time-domain method according to claim 1, characterized in that: The scattering objects are fine-tuned as described in step 7, and the attenuation and delay statistics of the multipath components are analyzed through multiple FDTD simulations.
3. The semi-deterministic channel simulation method based on finite-difference time-domain method according to claim 1, characterized in that: The iteration coefficients described in step 3 are stored in an indexed manner, which saves memory space and improves memory utilization efficiency.
4. The semi-deterministic channel simulation method based on the finite difference time domain method according to claim 1, characterized in that: In step 8, the results of multiple solutions are statistically analyzed to obtain the average value and standard deviation.
5. The semi-deterministic channel simulation method based on finite-difference time-domain method according to claim 1, characterized in that: The discrete time step Δt The discrete time step Δt needs to be less than or equal to the maximum time step Δt CFL , where Δx is the discrete grid size in the x direction, Δz is the discrete grid size in the z direction, and c is the speed of light.
6. The semi-deterministic channel simulation method based on finite-difference time-domain method according to claim 1, characterized in that: In step 4, the electromagnetic field is updated and iterated, wherein the magnetic field H y The update formula is: Among them, i and k are the spatial discrete coordinates in the x direction and the z direction respectively, n is the current time step, and the coefficient c is hyhy ,c hyez ,c hyex as follows Where μ is the magnetic permeability, σ m is the magnetic permeability coefficient, Δx is the discrete grid size in the x direction, Δz is the discrete grid size in the z direction, and Δt is the discrete time step.
7. The semi-deterministic channel simulation method based on finite-difference time-domain method according to claim 1, characterized in that: In step 4, the electromagnetic field is updated and iterated, wherein the electric field E x and E z The update formula is: Coefficient c exex ,c exhy ,c ezez ,c ezhy as follows Among them, ε is the dielectric constant and σ is the conductivity.
8. The semi-deterministic channel simulation method based on finite-difference time-domain method according to claim 1, characterized in that: In step 4, the received signal envelope is cross-correlated with the transmitted signal envelope, and the envelope cross-correlation formula is: Among them, sl(t) is the complex envelope of the transmitted signal, rl(t) is the complex envelope of the received signal, R rs (τ) is the cross-correlation function between rl(t) and sl(t), R ss (τ) is the autocorrelation function of sl(t), α n is the complex attenuation coefficient of the nth path, τ n is the transmission delay of the nth path.
