A method for generating orthogonal waveform of ultrasonic signal based on genetic algorithm
Through the orthogonal waveform design based on genetic algorithm, the frequency encoding matrix of ultrasonic signals is optimized to generate high orthogonal OFDM-LFM signals, which solves the signal interference problem in ultrasonic multi-access positioning system and improves positioning accuracy and calculation efficiency.
Patent Information
- Application Number
- CN202210637855.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-07
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2042-06-07
AI Technical Summary
In an ultrasonic multi-access positioning system, multiple access interference of multiple ultrasonic signals leads to a decrease in the accuracy of signal delay estimation, which in turn affects the positioning accuracy. The prior art has high computational complexity and low efficiency when optimizing orthogonal waveforms.
Using the orthogonal waveform design method based on genetic algorithm, the frequency encoding matrix of the signal is optimized to generate OFDM-LFM signals with high orthogonality, reduce the autocorrelation side lobe peak and cross-correlation peak of the signal, reduce the number of iterations, and improve the calculation. efficiency.
The waveform orthogonality is improved, multiple access interference between transmitted signals is reduced, calculation complexity is reduced, signal optimization efficiency is improved, and the number of iterations is stable when multiple signals is applied.
Smart Images

Figure CN115032589B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of wireless positioning and signal processing, and in particular relates to a method for generating orthogonal waveforms of ultrasonic signals based on a genetic algorithm. Background Art
[0002] In the ultrasonic multiple access positioning system, multiple ultrasonic signals are transmitted simultaneously, and there is multiple access interference (MAI) between the signals, which will reduce the accuracy of signal delay estimation and thus affect the final positioning accuracy. Therefore, effective orthogonal waveform design of the transmitted signal is the key to improving the accuracy of ultrasonic multiple access positioning. Ideally, the autocorrelation sidelobe value and cross-correlation value of the completely orthogonal signal are both zero, but this cannot be achieved in engineering. Therefore, in practical applications, the waveform is usually optimized with the goal of reducing the autocorrelation sidelobe peak (ASP) and cross-correlation peak (CP) of the signal.
[0003] The document "Design of Chirp Waveforms for Multiple-Access Ultrasonic Indoor Positioning" adopts the method of generating multiple-access excitation frequency modulation sequences for ultrasonic positioning from the field of radar communication, so that the signal produces pulse-like autocorrelation and relatively flat cross-correlation. When optimizing the generated waveform, the scheme uses the minimum distance formula to ensure that the autocorrelation sidelobe peak and the cross-correlation peak are jointly minimized. In order to limit the computational complexity, a trial-and-error strategy is adopted. However, as the number of transmitted signals increases, the computational complexity of the optimization method will still increase greatly, and the algorithm efficiency is low. The document "A Narrowband Ultrasonic Ranging Method for Multiple Moving Sensor Nodes" uses an exhaustive search method for waveform optimization, that is, traversing every possible result and then selecting the best result according to the optimization criterion, which also increases the computational complexity of signal optimization.
[0004] In order to reduce the number of iterations in the waveform optimization process and improve the waveform optimization efficiency, the present invention discloses an orthogonal waveform design method based on a genetic algorithm. Genetic Algorithm (GA) is an optimization algorithm first proposed and developed by Professor Holland JH of the University of Michigan in the mid-1960s. It simulates the biological evolution mechanism in nature, is based on natural selection and genetic theory, draws on Darwin's theory of evolution and Mendel's genetic theory, and combines the survival of the fittest rule in the biological evolution process with the random information exchange mechanism of chromosomes within the group. It is an efficient, parallel, global optimization search algorithm.
[0005] Genetic algorithms simulate the replication, crossover, and mutation phenomena that occur in natural selection and heredity. Starting from any initial population, they produce a group of individuals that are more adaptable to the environment through random selection, crossover, and mutation operations, allowing the population to evolve to better and better areas in the search space. They evolve from generation to generation and finally converge to a group of individuals that are most adaptable to the environment, thereby obtaining the optimal solution to the problem. Summary of the invention
[0006] The present invention provides an ultrasonic signal orthogonal waveform generation method based on a genetic algorithm, the purpose of which is to improve the orthogonality of the waveform, reduce the computational complexity in the waveform optimization process, and improve the computational efficiency.
[0007] The technical solution adopted by the present invention comprises the following steps:
[0008] Step 1: The system transmits M ultrasonic signals simultaneously, and the generated orthogonal signal set is S. Each signal consists of N consecutive subcarriers, and the duration of each subcarrier is T. z , each subcarrier of the same signal is modulated at a different frequency, orthogonal frequency division multiplexing OFDM signal s m Expressed as follows:
[0009]
[0010] Where m = 1, 2, ..., M, n = 1, 2, ..., N; t is the time parameter of the subcarrier, u(t) = 1, 0 ≤ t ≤ T z is a rectangular window function; P m,n (t) represents the nth subcarrier of the mth signal, and the linear frequency modulation signal LFM signal is used as the carrier, so the representation of the subcarrier is as follows:
[0011]
[0012] Among them, T z is the duration of the subcarrier; △f is the bandwidth of the subcarrier; f m,n represents the initial frequency of the nth subcarrier of the mth signal, fm,n =f0+c m,n △f, f0 is the initial frequency of the OFDM signal band, c m,n ∈{0,1,…,N-1}, the coding frequency of the mth signal can be expressed as {f m,1 ,f m,2 ,…,f m,N}, for M signals, a total of M×N frequency values need to be optimized and screened to obtain the subcarrier frequency that minimizes the correlation output value of the orthogonal signal. Therefore, the coding format of the orthogonal signal set S can be expressed in the form of the following frequency coding matrix F:
[0013]
[0014] Step 2: Design an orthogonal transmit waveform to generate an OFDM-LFM signal;
[0015] Step 3: OFDM-LFM signals obtained with different coding frequencies have different orthogonality. In order to obtain OFDM-LFM signals with good orthogonality, the frequency coding sequence of the signal must be optimized and screened. The frequency coding matrix F of the orthogonal signal set is optimized using a genetic algorithm to obtain the optimal frequency coding matrix of the signal that minimizes the signal correlation value output.
[0016] Step 4: Use the optimized optimal frequency coding matrix Generates quadrature waveforms for ultrasound signals.
[0017] Step 2 of the present invention comprises:
[0018] Step 2.1: Generate a linear frequency modulation signal, and generate two linear frequency modulation signals s according to the frequency increase and decrease u (t) and s d (t);
[0019]
[0020]
[0021] In the formula, f s and f e are the starting frequency and ending frequency of the linear frequency modulation signal, respectively, 0≤t≤T, T is the total duration of the linear frequency modulation signal;
[0022] Step 2.2: Generate a linear sub-FM signal using the linear FM signal s u (t) and s d (t) Generate N linearly increasing frequency sub-frequency modulation signals s respectively zu (t) and N linearly decreasing frequency sub-FM signals s zd (t):
[0023]
[0024]
[0025] Where m = 1, 2, ..., M, M is the number of signals; T z is the subcarrier duration, N is the number of subcarriers of each signal; f sm and f em are the starting frequency and ending frequency of each sub-FM signal, respectively, and the two are calculated as follows:
[0026]
[0027]
[0028] Step 2.3: Create the base FM signal using the sub-FM signal zu (t) and s zd (t) Create a frequency modulated base signal φ of the mixed frequency h,m (t):
[0029]
[0030] Step 2.4: Construct an orthogonal unit pulse basis. Use the cyclic shift sequence and the unit pulse to construct M orthogonal pulse bases. The mth unit pulse basis η m (t) is generated by the following formula:
[0031] η m,i (t) = δ(t)⊙(R(i)-1)T z i=1,2,…,N
[0032] Where R is a set of random permutations of {1,2,…,N}, R(i) is the i-th value of R; “⊙” represents a cyclic shift operation, and the sequence R is subjected to 0,1,2,…,M-1 cyclic shift operations respectively. The resulting M×N-order sequence matrix R is the frequency coding matrix F corresponding to the orthogonal signal;
[0033]
[0034] δ(t) represents the pulse signal;
[0035] Step 2.5: Create an orthogonal signal set using the frequency modulation basis φ h,m (t) and the orthogonal unit pulse basis η m (t) Construct and generate an orthogonal signal frequency modulation sequence, and then generate an orthogonal signal set. The formula for generating the frequency modulation sequence is as follows:
[0036]
[0037] FM Sequences h (t) represents the subcarrier frequency of different signals, according to the frequency modulation sequence s h The frequency correspondence of (t) generates LFM signals of different frequencies to obtain OFDM-LFM signals.
[0038] Step 3 of the present invention comprises:
[0039] Step 3.1: Encoding and generating the initial population;
[0040] Step 3.2: Calculate the fitness function. In orthogonal waveform design, the optimization criterion is the minimum of the maximum autocorrelation sidelobe peak and the maximum cross-correlation peak of the signal. In the genetic algorithm, the greater the fitness, the higher the individual survival rate and the greater the probability of being selected. Therefore, the inverse of the optimization criterion is used as the fitness function.
[0041] Step 3.3: Perform selection operation to select parent individuals based on the fitness of individuals in the population. Individuals with high fitness are superior individuals, and individuals with low fitness are inferior individuals. Superior individuals are selected and inferior individuals are eliminated.
[0042] Step 3.4: Perform a crossover operation and use the two selected individuals as parents. c Decide whether to perform a crossover operation. If so, exchange the same position of the two individuals in the parent generation and check whether the two sequences after the exchange have the same code. If so, the interval with the same code will not participate in the crossover.
[0043] Step 3.5: According to the mutation probability p m Decide whether to perform a mutation operation. If so, perform mutations at a certain position in the two individuals of the parent generation and flip the gene order of the mutated part.
[0044] Step 3.6: Update the population. After the individuals in the population have undergone the above selection, crossover, and mutation, new offspring are generated. The above operations are performed on all individuals in the population to generate a new population and realize population update. Repeat steps 3.2 to 3.5 until all genetic generations are completed, and the optimal individual is output, that is, the optimal frequency coding matrix of the signal, and the genetic algorithm optimization is completed.
[0045] In step 3.1 of the present invention, the frequency coding matrix F corresponding to the entire signal set S is taken as an individual, and each signal s in the signal set mAs a chromosome of an individual, each subcarrier frequency is a gene, then an individual consists of M chromosomes, and a chromosome consists of N genes. The signal is encoded according to the frequency characteristics of the LFM signal, and the gene range of the chromosome is set to {1, 2, ..., N}, and an initial population containing k individuals is randomly generated.
[0046] The genetic algorithm in step 3.2 of the present invention aims to reduce the autocorrelation sidelobe peak and cross-correlation peak of the signal as much as possible. In the orthogonal waveform design, the autocorrelation function and cross-correlation function are used to evaluate the performance of the designed orthogonal waveform signal. In actual calculation, the correlation function of the discrete sequence of the signal is used to replace the correlation function of the signal. The autocorrelation function of the transmitted signal can be expressed as:
[0047]
[0048] A(s m ,τ) is the autocorrelation function of the mth signal at time τ, and N is the sequence length;
[0049] The cross-correlation function of the transmitted signal can be expressed as:
[0050]
[0051] C(s p ,s q ,τ) is the cross-correlation function between the p-th signal and the q-th signal at time τ;
[0052] In order to reduce the interference between signals, the cross-correlation value of the transmitted signal needs to be as low as possible; at the same time, in order to improve the resolution of each signal, sharp autocorrelation is required, that is, the autocorrelation sidelobe of the signal is as low as possible. Therefore, the waveform optimization criterion of the genetic algorithm is that the autocorrelation sidelobe peak value and the cross-correlation peak value of the signal are minimized together, which is expressed as follows:
[0053]
[0054] In the genetic algorithm, the greater the fitness, the higher the individual survival rate and the greater the probability of being selected. The waveform design requires that the correlation between signals is as low as possible. Therefore, the inverse of the optimization criterion is used as the fitness function of the genetic algorithm. The fitness function expression is as follows:
[0055]
[0056] The roulette wheel selection method is used in step 3.3 of the present invention, that is, according to the fitness f of each individual in the population i The total fitness of the population The proportion of is used for selection. A high proportion means that it has a strong ability to produce offspring. The calculation of the roulette wheel selection method is as follows:
[0057]
[0058] Select the two individuals with the largest fitness ratio as parents, and then proceed to the next step.
[0059] Step 3.4 of the present invention comprises:
[0060] Step 3.4.1: Predetermine the chromosome crossover probability p c , and generate a random number rand in the range of 0 to 1. <p c When rand>p c When , execute step 3.4.3;
[0061] Step 3.4.2: Take the two individuals selected by the roulette wheel selection method as the parents, select a chromosome from each of the two parents, exchange the corresponding segments of the two chromosomes, that is, exchange a segment of the two frequency coding sequences, obtain two crossover offspring individuals, and then proceed to the next step; if the same coding exists on the same chromosome of the crossover offspring individuals, then the segment does not participate in the crossover, and the next step is directly performed;
[0062] Step 3.4.3: The current individual remains unchanged and proceed to the next step.
[0063] Step 3.5 of the present invention comprises:
[0064] Step 3.5.1: Predetermine the probability of gene mutation p m , and generate a random number rand in the range of 0 to 1. <p m When rand>p m When , execute step 3.5.3;
[0065] Step 3.5.2: Take the two individuals after the above crossover as the parents, and perform gene mutation operations on the two chromosomes respectively, that is, flip the coding order of the sequence variation fragments;
[0066] Step 3.5.3: The current individual remains unchanged.
[0067] In step 3.6 of the present invention, the best individuals of each generation are recorded in turn, and the final best individual is selected from the best individuals of all generations, that is, the frequency coding matrix corresponding to the minimum signal correlation value, which is the optimal frequency coding matrix obtained by genetic algorithm optimization.
[0068] Step 4 of the present invention comprises:
[0069] Step 4.1: Use the optimal frequency coding matrix obtained after optimization Instead of randomly permuting R in step 2.4 to construct an orthogonal unit pulse basis, the formula is as follows:
[0070]
[0071] in, is the optimal frequency coding matrix A sequence in For sequence The i-th value of ;
[0072] Step 4.2: Using the frequency modulation basis φ h,m (t) and the orthogonal unit pulse basis Construct and generate orthogonal signal frequency modulation sequence The formula is as follows:
[0073]
[0074] Step 4.3: According to the frequency modulation sequence The frequency correspondence generates LFM signals of different frequencies to obtain an optimized OFDM-LFM signal.
[0075] The present invention has the following advantages:
[0076] 1. The genetic algorithm is used to optimize the orthogonal waveform of the ultrasonic signal. It is an efficient, parallel, global search algorithm that can reduce the number of iterations, reduce the amount of calculation, and improve the search efficiency. Moreover, as the number of transmitted signals increases, the number of iterations of the genetic algorithm remains stable.
[0077] 2. Using the optimized frequency coding matrix to generate orthogonal waveforms can improve the orthogonality of the waveforms and reduce the multiple access interference between transmitted signals. BRIEF DESCRIPTION OF THE DRAWINGS
[0078] Figure 1 It is the genetic algorithm flow chart;
[0079] Figure 2 It is the simulation result diagram of genetic algorithm;
[0080] Figure 3 It is a diagram of the steps of generating orthogonal signals;
[0081] Figure 4 It is a schematic diagram of orthogonal signal correlation. DETAILED DESCRIPTION
[0082] In order to further explain the technical means adopted by the present invention to achieve the expected objectives and the final effect of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0083] A method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm comprises the following steps:
[0084] Step 1: The system transmits M ultrasonic signals simultaneously, and the generated orthogonal signal set is S. Each signal consists of N consecutive subcarriers, and the duration of each subcarrier is T. z , each subcarrier of the same signal is modulated at a different frequency, orthogonal frequency division multiplexing OFDM signal s m Expressed as follows:
[0085]
[0086] Where m = 1, 2, ..., M, n = 1, 2, ..., N; t is the time parameter of the subcarrier, u(t) = 1, 0 ≤ t ≤ T z is a rectangular window function; P m,n (t) represents the nth subcarrier of the mth signal, and the linear frequency modulation signal LFM signal is used as the carrier, so the representation of the subcarrier is as follows:
[0087]
[0088] Among them, T z is the duration of the subcarrier; △f is the bandwidth of the subcarrier; f m,n represents the initial frequency of the nth subcarrier of the mth signal, f m,n =f0+c m,n △f, f0 is the initial frequency of the OFDM signal band, c m,n ∈{0,1,…,N-1}, the coding frequency of the mth signal can be expressed as {f m,1 ,f m,2 ,…,f m,N}, for M signals, a total of M×N frequency values need to be optimized and screened to obtain the subcarrier frequency that minimizes the correlation output value of the orthogonal signal. Therefore, the coding format of the orthogonal signal set S can be expressed in the form of the following frequency coding matrix F:
[0089]
[0090] Step 2: Design an orthogonal transmit waveform to generate an OFDM-LFM signal;
[0091] Furthermore, the step 2 comprises:
[0092] Step 2.1: Generate a linear frequency modulation signal, and generate two linear frequency modulation signals s according to the frequency increase and decrease u (t) and s d (t);
[0093]
[0094]
[0095] In the formula, f s and f e are the starting frequency and ending frequency of the linear frequency modulation signal, respectively, 0≤t≤T, T is the total duration of the linear frequency modulation signal;
[0096] Step 2.2: Generate a linear sub-FM signal using the linear FM signal s u (t) and s d (t) Generate N linearly increasing frequency sub-frequency modulation signals s respectively zu (t) and N linearly decreasing frequency sub-FM signals s zd (t):
[0097]
[0098]
[0099] Where m = 1, 2, ..., M, M is the number of signals; T z is the subcarrier duration, N is the number of subcarriers of each signal; f sm and f em are the starting frequency and ending frequency of each sub-FM signal, respectively, and the two are calculated as follows:
[0100]
[0101]
[0102] Step 2.3: Create a frequency modulation base signal. Relevant literature has verified that the frequency modulation base signal based on the mixed frequency has a better cross-correlation suppression effect. Therefore, the present invention uses the sub-frequency modulation signal s zu (t) and s zd (t) Create a frequency modulated base signal φ of the mixed frequency h,m (t):
[0103]
[0104] Step 2.4: Construct an orthogonal unit pulse basis. Use the cyclic shift sequence and the unit pulse to construct M orthogonal pulse bases. The mth unit pulse basis η m (t) is generated by the following formula:
[0105] η m,i (t) = δ(t)⊙(R(i)-1)T z i=1,2,…,N
[0106] Where R is a set of random permutations of {1,2,…,N}, R(i) is the i-th value of R; “⊙” represents a cyclic shift operation, and the sequence R is subjected to 0,1,2,…,M-1 cyclic shift operations respectively. The resulting M×N-order sequence matrix R is the frequency coding matrix F corresponding to the orthogonal signal;
[0107]
[0108] δ(t) represents the pulse signal;
[0109] Step 2.5: Create an orthogonal signal set using the frequency modulation basis φ h,m (t) and the orthogonal unit pulse basis η m (t) Construct and generate an orthogonal signal frequency modulation sequence, and then generate an orthogonal signal set. The formula for generating the frequency modulation sequence is as follows:
[0110]
[0111] FM Sequences h (t) represents the subcarrier frequency of different signals, according to the frequency modulation sequence s h The frequency of (t) generates LFM signals of different frequencies to obtain OFDM-LFM signals;
[0112] Step 3: OFDM-LFM signals obtained with different coding frequencies have different orthogonality. In order to obtain OFDM-LFM signals with good orthogonality, the frequency coding sequence of the signal must be optimized and screened. The frequency coding matrix F of the orthogonal signal set is optimized using a genetic algorithm to obtain the optimal frequency coding matrix of the signal that minimizes the signal correlation value output. Figure 1 It is the flowchart of genetic algorithm;
[0113] Furthermore, the step 3 comprises:
[0114] Step 3.1: Encoding and generating the initial population;
[0115] Step 3.2: Calculate the fitness function. In orthogonal waveform design, the optimization criterion is the minimum of the maximum autocorrelation sidelobe peak and the maximum cross-correlation peak of the signal. In the genetic algorithm, the greater the fitness, the higher the individual survival rate and the greater the probability of being selected. Therefore, the inverse of the optimization criterion is used as the fitness function.
[0116] Step 3.3: Perform selection operation to select parent individuals based on the fitness of individuals in the population. Individuals with high fitness are superior individuals, and individuals with low fitness are inferior individuals. Superior individuals are selected and inferior individuals are eliminated.
[0117] Step 3.4: Perform a crossover operation and use the two selected individuals as parents.c Decide whether to perform a crossover operation. If so, exchange the same position of the two individuals in the parent generation and check whether the two sequences after the exchange have the same code. If so, the interval with the same code will not participate in the crossover.
[0118] Step 3.5: According to the mutation probability p m Decide whether to perform a mutation operation. If so, perform mutations at a certain position in the two individuals of the parent generation and flip the gene order of the mutated part.
[0119] Step 3.6: Update the population. After the individuals in the population have undergone the above selection, crossover, and mutation, new offspring are generated. The above operations are performed on all individuals in the population to generate a new population and realize population update. Repeat steps 3.2 to 3.5 until all genetic generations are completed, and the optimal individual is output, that is, the optimal frequency coding matrix of the signal, and the genetic algorithm optimization is completed.
[0120] One implementation of the present invention is: in step 3.1, the frequency coding matrix F corresponding to the entire signal set S is taken as an individual, and each signal s in the signal set m As a chromosome of an individual, each subcarrier frequency is a gene, then an individual consists of M chromosomes, and a chromosome consists of N genes. The signal is encoded according to the frequency characteristics of the LFM signal, the gene range of the chromosome is set to {1,2,…,N}, and an initial population of k individuals is randomly generated.
[0121] One implementation of the present invention is: the genetic algorithm in step 3.2 aims to reduce the autocorrelation sidelobe peak and cross-correlation peak of the signal as much as possible. In the orthogonal waveform design, the autocorrelation function and cross-correlation function are used to evaluate the performance of the designed orthogonal waveform signal. In actual calculation, the correlation function of the signal discrete sequence is used to replace the correlation function of the signal. The autocorrelation function of the transmitted signal can be expressed as:
[0122]
[0123] A(s m ,τ) is the autocorrelation function of the mth signal at time τ, and N is the sequence length;
[0124] The cross-correlation function of the transmitted signal can be expressed as:
[0125]
[0126] C(s p ,s q ,τ) is the cross-correlation function between the p-th signal and the q-th signal at time τ;
[0127] In order to reduce the interference between signals, the cross-correlation value of the transmitted signal needs to be as low as possible; at the same time, in order to improve the resolution of each signal, sharp autocorrelation is required, that is, the autocorrelation sidelobe of the signal is as low as possible. Therefore, the waveform optimization criterion of the genetic algorithm is that the autocorrelation sidelobe peak value and the cross-correlation peak value of the signal are minimized together, which is expressed as follows:
[0128]
[0129] In the genetic algorithm, the greater the fitness, the higher the individual survival rate and the greater the probability of being selected. The waveform design requires that the correlation between signals is as low as possible. Therefore, the inverse of the optimization criterion is used as the fitness function of the genetic algorithm. The fitness function expression is as follows:
[0130]
[0131] One implementation of the present invention is: the roulette selection method is used in step 3.3, that is, according to the fitness f of each individual in the population i The total fitness of the population The proportion of is used for selection. A high proportion means that it has a strong ability to produce offspring. The calculation of the roulette wheel selection method is as follows:
[0132]
[0133] Select the two individuals with the largest fitness ratio as parents, and then proceed to the next step.
[0134] One embodiment of the present invention is: step 3.4 comprises:
[0135] Step 3.4.1: Predetermine the chromosome crossover probability p c , and generate a random number rand in the range of 0 to 1. <p c When rand>p c When , execute step 3.4.3;
[0136] Step 3.4.2: Take the two individuals selected by the roulette wheel selection method as the parents, select a chromosome from each of the two parents, exchange the corresponding segments of the two chromosomes, that is, exchange a segment of the two frequency coding sequences, obtain two crossover offspring individuals, and then proceed to the next step; if the same coding exists on the same chromosome of the crossover offspring individuals, then the segment does not participate in the crossover, and the next step is directly performed;
[0137] Step 3.4.3: The current individual remains unchanged and proceed to the next step.
[0138] One implementation mode of the present invention is: the step 3.5 comprises:
[0139] Step 3.5.1: Predetermine the probability of gene mutation p m , and generate a random number rand in the range of 0 to 1. <p m When rand>p m When , execute step 3.5.3;
[0140] Step 3.5.2: Take the two individuals after the above crossover as the parents, and perform gene mutation operations on the two chromosomes respectively, that is, flip the coding order of the sequence variation fragments;
[0141] Step 3.5.3: The current individual remains unchanged.
[0142] One implementation of the present invention is: in step 3.6, the best individuals of each generation are recorded in sequence, and the final best individual is selected from the best individuals of all generations, that is, the frequency coding matrix corresponding to the minimum signal correlation value, which is the optimal frequency coding matrix obtained by genetic algorithm optimization.
[0143] Step 4: Use the optimized optimal frequency coding matrix generating an orthogonal waveform of an ultrasonic signal;
[0144] Furthermore, the step 4 comprises:
[0145] Step 4.1: Use the optimal frequency coding matrix obtained after optimization Instead of randomly permuting R in step 2.4 to construct an orthogonal unit pulse basis, the formula is as follows:
[0146]
[0147] in, is the optimal frequency coding matrix A sequence in For sequence The i-th value of ;
[0148] Step 4.2: Using the frequency modulation basis φ h,m (t) and the orthogonal unit pulse basis Construct and generate orthogonal signal frequency modulation sequence The formula is as follows:
[0149]
[0150] Step 4.3: According to the frequency modulation sequence The frequency correspondence generates LFM signals of different frequencies to obtain an optimized OFDM-LFM signal.
[0151] The present invention will be further described below in conjunction with the accompanying drawings and specific examples:
[0152] Table 1 shows the parameter settings of the simulation experiment, including signal parameters and genetic algorithm parameters. In order to maximize the bandwidth utilization, the number of transmitted signals and the number of subcarriers are made equal.
[0153] Table 1 Waveform design simulation parameters
[0154] Signal parameters Genetic algorithm parameters Number of transmitted signals M=6 Population size k=50 Number of subcarriers N=6 Genetic Algebra n=500 Sampling time T=12ms Crossover probability Pc=0.8 Sampling frequency Fs=1MHz Mutation probability Pm=0.8 Start frequency fs=38kHz Stop frequency fe=44kHz
[0155] Figure 2 The simulation result diagram of the genetic algorithm shows the optimization process of the genetic algorithm. The horizontal axis represents the genetic generation, and the vertical axis represents the objective function value of the optimal individual in each generation during the program operation, that is, the sum of the maximum value of the autocorrelation sidelobe peak and the maximum value of the cross-correlation peak of the optimal orthogonal signal set in each generation. It can be seen that with the increase of the genetic generation, the population gradually evolves, and the objective function value of the current best individual gradually decreases.
[0156] Table 2 shows the optimal frequency coding sequence obtained using the genetic algorithm. In order to verify the accuracy of the genetic algorithm, the exhaustive search method was used to optimize the sequence. Table 3 shows the optimal frequency coding sequence obtained using the exhaustive search method. The results of the two are equal, which proves the accuracy of the genetic algorithm optimization.
[0157] Table 2 Optimal frequency coding sequence based on genetic algorithm
[0158]
[0159] Table 3 Optimal frequency coding sequence based on exhaustive search method
[0160]
[0161] Figure 3 This is a schematic diagram of generating an orthogonal waveform using the optimal frequency coding sequence in Table 2. The signal parameter settings are consistent with Table 1. Among them, 6 orthogonal unit pulse bases and orthogonal frequency modulation sequences are generated, and 1 is used here for illustration.
[0162] Figure 4Figure 1 is a schematic diagram of the orthogonal waveform generated by the optimal frequency coding sequence, where (a) to (f) represent the autocorrelation of each signal and the cross-correlation with all other signals, respectively. It can be seen that the signal autocorrelation is relatively sharp, because the LFM signal is used as the subcarrier to improve the signal resolution. At the same time, the cross-correlation of the signals is relatively flat, because the designed OFDM waveform reduces the similarity between signals by making the signal frequencies orthogonal, thereby reducing the cross-correlation interference between signals. In addition, this paper uses a genetic algorithm to determine the optimal frequency coding sequence, which further improves the orthogonality of the signal.
[0163] Table 4 shows the autocorrelation sidelobe peak value and cross-correlation peak value of each signal in the optimal individual. In the table, each signal corresponds to its own value as the autocorrelation sidelobe peak value, and corresponds to the value of other signals as the cross-correlation peak value.
[0164] Table 4 ASP and CP of the optimal individual
[0165]
[0166] It can be seen from Table 4 that the cross-correlation peaks of the signals all reach a relatively low level.
[0167] Table 5 shows the number of iterations required to obtain the optimal frequency coding sequence using the exhaustive search method and the genetic algorithm for different numbers of transmitted signals (assuming that the number of signals is always equal to the number of subcarriers).
[0168] Table 5 Comparison of iterations between genetic algorithm and exhaustive search method
[0169] M=5 M=6 M=7 M=8 M=9 M=10 Exhaustive search 120 720 5040 40320 362880 3628800 Genetic Algorithms 383 125 269 215 100 466
[0170] The number of iterations required by the genetic algorithm is much smaller than that required by the exhaustive search method; moreover, as the number of transmitted signals increases, the number of iterations of the exhaustive search method is greatly improved, while the number of iterations of the genetic algorithm can be stabilized within the set genetic generation. Therefore, compared with the exhaustive search method, the use of genetic algorithms for sequence optimization can improve the optimization efficiency.
Claims
1. A method for generating orthogonal waveforms of ultrasonic signals based on genetic algorithm, characterized in that: The following steps are involved: Step 1: The system transmits M ultrasonic signals simultaneously, and the generated orthogonal signal set is S. Each signal consists of N consecutive subcarriers, and the duration of each subcarrier is T. z , each subcarrier of the same signal is modulated at a different frequency, orthogonal frequency division multiplexing OFDM signal s m Expressed as follows: Where m = 1, 2, ..., M, n = 1, 2, ..., N; t is the time parameter of the subcarrier, u(t) = 1, 0 ≤ t ≤ T z is a rectangular window function; P m,n (t) represents the nth subcarrier of the mth signal, and the linear frequency modulation signal LFM signal is used as the carrier, so the representation of the subcarrier is as follows: Among them, T z is the duration of the subcarrier; △f is the bandwidth of the subcarrier; f m,n represents the initial frequency of the nth subcarrier of the mth signal, f m,n =f0+c m,n △f, f0 is the initial frequency of the OFDM signal band, c m,n ∈{0,1,…,N-1}, the coding frequency of the mth signal can be expressed as {f m,1 ,f m,2 ,…,f m,N }, for M signals, a total of M×N frequency values need to be optimized and screened to obtain the subcarrier frequency that minimizes the correlation output value of the orthogonal signal. Therefore, the coding format of the orthogonal signal set S can be expressed in the form of the following frequency coding matrix F: Step 2: Design an orthogonal transmit waveform to generate an OFDM-LFM signal; Step 3: OFDM-LFM signals obtained with different coding frequencies have different orthogonality. In order to obtain OFDM-LFM signals with good orthogonality, the frequency coding sequence of the signal must be optimized and screened. The frequency coding matrix F of the orthogonal signal set is optimized using a genetic algorithm to obtain the optimal frequency coding matrix of the signal that minimizes the signal correlation value output. Step 4: Use the optimized optimal frequency coding matrix Generates quadrature waveforms of ultrasound signals.
2. The method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm according to claim 1, characterized in that: The step 2 comprises: Step 2.1: Generate a linear frequency modulation signal, and generate two linear frequency modulation signals s according to the frequency increase and decrease u (t) and s d (t); In the formula, f s and f e are the starting frequency and ending frequency of the linear frequency modulation signal, respectively, 0≤t≤T, T is the total duration of the linear frequency modulation signal; Step 2.2: Generate a linear sub-FM signal using the linear FM signal s u (t) and s d (t) Generate N linearly increasing frequency sub-frequency modulation signals s respectively zu (t) and N linearly decreasing frequency sub-FM signals s zd (t): Where m = 1, 2, ..., M, M is the number of signals; T z is the subcarrier duration, N is the number of subcarriers of each signal; f sm and f em are the starting frequency and ending frequency of each sub-FM signal, respectively, and the two are calculated as follows: Step 2.3: Create the base FM signal using the sub-FM signal zu (t) and s zd (t) Create a frequency modulated base signal φ of the mixed frequency h,m (t): Step 2.4: Construct an orthogonal unit pulse basis. Use the cyclic shift sequence and the unit pulse to construct M orthogonal pulse bases. The mth unit pulse basis η m (t) is generated by the following formula: η m,i (t)=δ(t)⊙(R(i)-1)T z i=1,2,…,N Where R is a set of random permutations of {1,2,…,N}, R(i) is the i-th value of R; "⊙" represents a cyclic shift operation, and the sequence R is subjected to 0,1,2,…,M-1 cyclic shift operations respectively. The resulting M×N-order sequence matrix R is the frequency coding matrix F corresponding to the orthogonal signal; δ(t) represents the pulse signal; Step 2.5: Create an orthogonal signal set using the frequency modulation basis φ h,m (t) and the orthogonal unit pulse basis η m (t) Construct and generate an orthogonal signal frequency modulation sequence, and then generate an orthogonal signal set. The formula for generating the frequency modulation sequence is as follows: FM Sequences h (t) represents the subcarrier frequency of different signals, according to the frequency modulation sequence s h The frequency correspondence of (t) generates LFM signals of different frequencies to obtain OFDM-LFM signals.
3. The method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm according to claim 1, characterized in that: The step 3 comprises: Step 3.1: Encoding and generating the initial population; Step 3.2: Calculate the fitness function. In orthogonal waveform design, the optimization criterion is the minimum of the maximum autocorrelation sidelobe peak and the maximum cross-correlation peak of the signal. In the genetic algorithm, the greater the fitness, the higher the individual survival rate and the greater the probability of being selected. Therefore, the inverse of the optimization criterion is used as the fitness function. Step 3.3: Perform selection operation to select parent individuals based on the fitness of individuals in the population. Individuals with high fitness are superior individuals, and individuals with low fitness are inferior individuals. Superior individuals are selected and inferior individuals are eliminated. Step 3.4: Perform a crossover operation and use the two selected individuals as parents. c Decide whether to perform a crossover operation. If so, exchange the same position of the two individuals in the parent generation and check whether the two sequences after the exchange have the same code. If so, the interval with the same code will not participate in the crossover. Step 3.5: According to the mutation probability p m Decide whether to perform a mutation operation. If so, perform mutations at a certain position in the two individuals of the parent generation and flip the gene order of the mutated part. Step 3.6: Update the population. After the individuals in the population have undergone the above selection, crossover, and mutation, new offspring are generated. The above operations are performed on all individuals in the population to generate a new population and realize population update. Repeat steps 3.2 to 3.5 until all genetic generations are completed, and the optimal individual is output, that is, the optimal frequency coding matrix of the signal, and the genetic algorithm optimization is completed.
4. The method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm according to claim 3, characterized in that: In step 3.1, the frequency coding matrix F corresponding to the entire signal set S is taken as an individual, and each signal s in the signal set m As a chromosome of an individual, each subcarrier frequency is a gene, then an individual consists of M chromosomes, and a chromosome consists of N genes. The signal is encoded according to the frequency characteristics of the LFM signal, and the gene range of the chromosome is set to {1, 2, ..., N}, and an initial population containing k individuals is randomly generated.
5. The method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm according to claim 3, characterized in that: The genetic algorithm in step 3.2 aims to minimize the autocorrelation sidelobe peak and cross-correlation peak of the signal. In the orthogonal waveform design, the autocorrelation function and cross-correlation function are used to evaluate the performance of the designed orthogonal waveform signal. In actual calculation, the correlation function of the discrete sequence of the signal is used to replace the correlation function of the signal. The autocorrelation function of the transmitted signal can be expressed as: A(s m ,τ) is the autocorrelation function of the mth signal at time τ, and N is the sequence length; The cross-correlation function of the transmitted signal can be expressed as: C(s p ,s q ,τ) is the cross-correlation function between the p-th signal and the q-th signal at time τ; In order to reduce the interference between signals, the cross-correlation value of the transmitted signal needs to be as low as possible; at the same time, in order to improve the resolution of each signal, sharp autocorrelation is required, that is, the autocorrelation sidelobe of the signal is as low as possible. Therefore, the waveform optimization criterion of the genetic algorithm is that the autocorrelation sidelobe peak value and the cross-correlation peak value of the signal are minimized together, which is expressed as follows: In the genetic algorithm, the greater the fitness, the higher the individual survival rate and the greater the probability of being selected. The waveform design requires that the correlation between signals is as low as possible. Therefore, the inverse of the optimization criterion is used as the fitness function of the genetic algorithm. The fitness function expression is as follows:
6. The method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm according to claim 3, characterized in that: The roulette wheel selection method is used in step 3.3, that is, according to the fitness f of each individual in the population i The total fitness of the population The proportion of selection is used, and a high proportion represents a strong ability to produce offspring. The calculation of the roulette selection method is as follows: Select the two individuals with the largest fitness ratio as parents, and then proceed to the next step.
7. The method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm according to claim 3, characterized in that: The step 3.4 comprises: Step 3.4.1: Predetermine the chromosome crossover probability p c , and generate a random number rand in the range of 0 to 1. <p c When rand>p c When , execute step 3.4.3; Step 3.4.2: Take the two individuals selected by the roulette wheel selection method as the parents, select a chromosome from each of the two parents, exchange the corresponding segments of the two chromosomes, that is, exchange a segment of the two frequency coding sequences, obtain two crossover offspring individuals, and then proceed to the next step; if the same coding exists on the same chromosome of the crossover offspring individuals, then the segment does not participate in the crossover, and the next step is directly performed; Step 3.4.3: The current individual remains unchanged and proceed to the next step.
8. The method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm according to claim 3, characterized in that: The step 3.5 comprises: Step 3.5.1: Predetermine the probability of gene mutation p m , and generate a random number rand in the range of 0 to 1. <p m When rand>p m When , execute step 3.5.3; Step 3.5.2: Take the two individuals after the above crossover as the parents, and perform gene mutation operations on the two chromosomes respectively, that is, flip the coding order of the sequence variation fragments; Step 3.5.3: The current individual remains unchanged.
9. The method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm according to claim 3, characterized in that: In step 3.6, the best individuals of each generation are recorded in turn, and the final best individual is selected from the best individuals of all generations, that is, the frequency coding matrix corresponding to the minimum signal correlation value, which is the optimal frequency coding matrix obtained by genetic algorithm optimization.
10. The method for generating an orthogonal waveform of an ultrasonic signal based on a genetic algorithm according to claim 1, characterized in that: The step 4 comprises: Step 4.1: Use the optimal frequency coding matrix obtained after optimization Instead of randomly permuting R in step 2.4 to construct an orthogonal unit pulse basis, the formula is as follows: in, is the optimal frequency coding matrix A sequence in For sequence The i-th value of h,m (t) and the orthogonal unit pulse basis Construct and generate orthogonal signal frequency modulation sequence The formula is as follows: Step 4.3: According to the frequency modulation sequence The frequency correspondence generates LFM signals of different frequencies to obtain an optimized OFDM-LFM signal.
