JP2006109083A - Hearing-aid adjuster and hearing-aid adjusting program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce the time for calculating various parameters of a hearing-aid to obtain target hearing-aid characteristics and raise the accuracy of various parameters of the hearing-aid. <P>SOLUTION: The adjuster dynamically changes a mutation rate being a parameter used in a geometric algorithm (GA), reduces the occasion of falling to a local solution with a small number of samples in combination of huge acoustic parameters of the hearing aid and the time for calculating the acoustic parameters of the hearing aid to obtain hearing-aid characteristics required by the hearing aid user, and raises the accuracy of the hearing-aid acoustic parameters. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a hearing aid adjustment device or a hearing aid adjustment program for calculating various parameters that determine signal processing characteristics of a hearing aid.

  The hearing aid has a function of amplifying the input sound input from the microphone and outputting it from the receiver. Which frequency component of the input sound is amplified and output, that is, the gain at each frequency is one of the characteristics of the hearing aid. The characteristics of such hearing aids need to be adjusted to match the hearing ability of the person using the hearing aid.

  Conventional hearing aids include a filter using a capacitor and a variable resistor, and the characteristics of the hearing aid have been adjusted by adjusting the filter. For example, in a high-pass filter composed of a capacitor and a variable resistor included in a hearing aid, changing the resistance value of the variable resistor changes the cutoff frequency and changes the amount of attenuation of the low frequency. I was able to. Similarly, in the low-pass filter, the amount of attenuation of the high frequency can be changed. Thus, the characteristic of the hearing aid has been adjusted by adjusting the frequency attenuation of the filter. There is also a hearing aid that can change the maximum output by changing the power supply voltage of the receiver with a variable resistor, and can provide an output limit. Adjustment of these conventional hearing aids has been performed by a specialist manually adjusting a variable resistor provided on the surface of the hearing aid using a small screwdriver.

  In recent years, the hearing aid has an EEPROM (electrically erasable programmable read-only memory), and the resistance value of the variable resistor is set by the information of the EEPROM to adjust the characteristics. Digital hearing aids that can adjust characteristics more finely by setting many adjustment items have become mainstream. In order to adjust the characteristics of these hearing aids, hearing aid adjustment devices are used. The hearing aid adjustment device determines various parameters for determining the characteristics of these hearing aids, and adjusts the characteristics of the hearing aids by writing various parameter information in an EEPROM or the like included in the hearing aid.

  When digital signal processing is used in a hearing aid as described above, the items that can be adjusted by the hearing aid and the values that can be set for various parameters increase, and the number of combinations thereof becomes enormous. When all of these combinations are simulated and the optimum values of the hearing aid characteristics required by the hearing aid wearer are calculated, it takes a very long time, and the burden on the hearing aid wearer and the hearing aid adjuster increases. Therefore, by using a simulated annealing method (hereinafter referred to as SA method) or a genetic algorithm (Genetic Algorithm, hereinafter referred to as GA method) as an optimization method, the hearing aid characteristics required by the hearing aid wearer are obtained. The way to get is taken.

  For example, in a hearing aid adjustment device, as a method of bringing the characteristics of a hearing aid closer to the target characteristic, a method is proposed in which a person evaluates the characteristic and uses the GA method to bring the hearing aid closer to the target characteristic. (For example, refer to Patent Document 1 and Patent Document 2). The prior art will be described below with reference to the drawings. FIG. 4 shows the order of conventional optimization using the GA method.

  Step 1 indicates the start of optimization using the GA method, Step 2 creates an individual for use in the genetic algorithm, Step 3 writes individual information to the hearing aid, and Step 4 writes information in Step 3. Hearing aid evaluation, step 5 is evaluation of the next hearing aid, transition to the next generation in GA method, or selection of termination, step 6 selects two individuals to be crossed, step 7 is abrupt for the individual selected in step 6 Mutation, step 8 is crossover of individuals, step 9 is a comparison of the number of next generation individuals obtained using the GA method with a preset number D of individuals, step 10 adopts hearing aid parameters of the best evaluation value, step 11 shows the end of the optimization approaching the recommended characteristics of the hearing aid using the GA method.

  First, in step 1, optimization using the GA method is started. In step 2, individuals for use in the genetic algorithm are created. Specifically, among acoustic parameters that determine the characteristics of the hearing aid, acoustic parameters to be varied are randomly distributed and converted into bits, and a preset number of D individuals are created. In step 3, one individual is selected from the D individuals created in step 2, and the acoustic parameter represented by the selected individual is written to the hearing aid as an acoustic parameter that determines the characteristics of the hearing aid. In step 4, the evaluator actually uses the hearing aid in which the acoustic parameters are written in step 3, and evaluates the characteristics of the hearing aid. In step 5, a selection is made between setting the next individual, moving to the next generation using the GA method, or ending with the current information. Since the processing of Step 3 and Step 4 is repeated for all D individuals, the setting of the next individual is selected until the above processing is completed for all D individuals.

  When the above process is completed for all D, the transition to the next generation is selected, and the process of step 6 is performed. In step 6, two individuals are selected so that the probability that a good evaluation value in step 4 is selected is increased. In step 7, one bit or a plurality of bits are inverted according to the mutation rate for each selected individual. Here, the mutation rate indicates a rate at which an individual changes to completely different data in a genetic algorithm. In step 8, these two individuals are crossed. Crossover is performed by exchanging some of the bits of these two individuals. In step 9, a comparison is made to repeat the crossover until the number of next-generation individuals reaches the number D of individuals created in advance. When the number of next-generation individuals becomes D, the processing returns to step 3 and the subsequent processing is repeated.

If the end is selected in step 5, the information of the individual that is the best evaluation value is written to the hearing aid in step 10, and the process ends in step 11.
Japanese Patent Application Laid-Open No. 09-054765 JP 2001-175637 A

  In the above conventional example, it is necessary to perform the time-consuming work of writing the acoustic parameters to the hearing aid in step 3 and evaluating the characteristics of the hearing aid in step 4 for the D individuals, which takes time. Therefore, there is a problem that it is difficult to shorten the adjustment time and reduce the burden on the hearing aid wearer and the hearing aid adjuster.

  In addition, when evaluating a hearing aid wearer, there is a risk of a sudden change in volume, so it is difficult to increase the mutation rate in the processing in step 7. If the mutation rate is lowered, in the GA method, the search space is narrowed and the obtained optimal solution often falls into a local solution. In order to widen the search space and prevent the obtained optimal solution from falling into a local solution, it is necessary to increase the set number D of individual creation in step 2 or increase the number of generations. However, if the set number D for individual creation is increased, many processes are performed for each generation, which takes time. Similarly, if the number of generations is increased, the processing is increased correspondingly and it takes time. Therefore, it is difficult to shorten the time required for evaluation. That is, it takes time to calculate various parameters of the hearing aid, and there is a problem that the accuracy cannot be increased in a short time.

  Therefore, an object of the present invention is to provide a hearing aid adjustment device or a hearing aid adjustment program that can efficiently calculate various parameters that give the hearing aid characteristics required by the hearing aid wearer and can shorten the adjustment time.

  In order to solve the above-described problems, a hearing aid adjustment device according to the present invention is a hearing aid adjustment device that adjusts characteristics of a hearing aid by setting an acoustic parameter, and inputs data representing the hearing ability of a person using the hearing aid. An input unit, a recommended characteristic calculation unit that calculates a recommended characteristic that is a characteristic that the hearing aid should have, an optimal value calculation unit that calculates an optimal value of the acoustic parameter based on the recommended characteristic, and an optimal value calculation unit A setting unit configured to set the optimum value of the obtained acoustic parameter in the hearing aid.

  According to the hearing aid adjustment apparatus of the present invention, the optimum value calculation unit includes a plurality of acoustic parameters for determining the characteristics of the hearing aid based on the recommended characteristics calculated based on the data representing the hearing ability of the person using the hearing aid. Since the optimum value is obtained, it is possible to efficiently obtain an acoustic parameter suitable for the hearing ability of the hearing aid user without going through a time-consuming and laborious process in which a person evaluates the characteristics of the hearing aid. As a result, the time required for adjusting the hearing aid is shortened.

  In the hearing aid adjustment apparatus according to the present invention, the optimum value calculation unit calculates an optimum value of the acoustic parameter using a genetic algorithm, and the optimum value calculation unit has a value that the acoustic parameter can take. The data representing one pattern is determined as one individual, and an individual group creating unit that creates a plurality of individuals, and the individual group in the individual group as input data is determined by the acoustic parameter value represented by each individual in the individual group. An evaluation unit that evaluates the fitness of each individual by comparing the characteristics of the hearing aid with the recommended characteristics, and determines whether or not an individual showing the optimal value of the acoustic parameter has been obtained. If it is determined that the optimum value is not obtained by the determination unit that terminates the calculation process and the determination unit, a new generation of the next generation is determined based on the evaluation by the evaluation unit. Create a body unit, it is preferable to provide a next generation population creating section that passes the population as input data to the evaluation unit.

  The evaluation unit evaluates the fitness of each individual by comparing the characteristics of the hearing aid determined by the values of the plurality of acoustic parameters represented by the individual and the recommended characteristics. There is no need for a person to evaluate the characteristics of the hearing aid. Therefore, the optimum values of the plurality of acoustic parameters can be obtained efficiently.

  The next generation population creation unit in the present invention includes a mutation unit that mutates the data of the individual based on a mutation rate, and the mutation rate is a population created by the next generation population creation unit. It is preferable to change according to the number of generations.

  In the hearing aid adjustment apparatus according to the present invention, it is preferable that the mutation rate decreases as the number of generations increases.

  In the hearing aid adjustment device according to a preferred embodiment of the present invention, the next generation population creation unit increases the mutation rate when the generation number is small, and decreases the mutation rate as the generation number increases. Even when the number of individuals is small or when the number of generations is not increased, it becomes possible to search a wider space, and it is possible to reduce falling into a local solution. As a result, highly accurate optimum acoustic parameter values can be obtained in a short time.

  In the hearing aid adjustment apparatus according to the preferred embodiment of the present invention, various acoustic parameters that provide the hearing aid characteristics required by the hearing aid wearer can be efficiently obtained by dynamically changing the parameters used in the GA method while using the GA method. It is possible to calculate, shorten the adjustment time, and improve the accuracy of various acoustic parameters of the hearing aid.

  In order to solve the above problems, a hearing aid adjustment program according to the present invention is a hearing aid adjustment program for adjusting the characteristics of a hearing aid by setting an acoustic parameter, the data representing the hearing ability of a person using the hearing aid. An input process for input, a recommended characteristic calculation process for calculating a recommended characteristic that is a characteristic that the hearing aid should have, an optimal value calculation process for calculating an optimal value of the acoustic parameter based on the recommended characteristic, and the optimal value calculation A setting process for setting optimal values of the plurality of acoustic parameters obtained in the process to the hearing aid is executed by a computer.

  In the hearing aid adjustment program according to the present invention, the optimum value calculating process calculates an optimum value of the acoustic parameter using a genetic algorithm, and the optimum value calculating process includes a value that can be taken by the acoustic parameter. Hearing aid determined by the value of the acoustic parameter represented by each individual in the individual group using the individual group in the individual group as the input data, and the individual group creation process for creating a plurality of individuals The evaluation process for evaluating the fitness of each individual by comparing the characteristics of the characteristics and the recommended characteristics, and whether or not an individual showing the optimal value of the acoustic parameter was obtained, and determined to have been obtained In the case where the optimum value calculation process is terminated and the evaluation process determines that the optimum value is not obtained in the decision process. Based on, to create a new population of the next generation, it is preferred to include a next generation population creation process to perform the evaluation process the population as input data.

  In the hearing aid adjustment program according to the present invention, the next-generation population creation process includes a mutation process for mutating the data of the individual based on a mutation rate, and the mutation rate is the next-generation population creation process. It is preferable to change according to the number of generations of the population created in (1).

  In the hearing aid adjustment program according to the present invention, it is preferable that the mutation rate decreases as the number of generations increases.

  According to the present invention, it is possible to provide a hearing aid adjustment device or a hearing aid adjustment program capable of efficiently calculating various parameters of a hearing aid for obtaining a hearing aid characteristic required by a hearing aid wearer and shortening the adjustment time.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing an example of the configuration of the hearing aid adjustment apparatus according to the present embodiment. As shown in FIG. 1, the hearing aid adjustment device 20 includes an input unit 12, a recommended characteristic calculation unit 13, an optimum value calculation unit 14, and a setting unit 15. The optimum value calculation unit 14 includes an individual group creation unit 141, an evaluation unit 142, a next generation individual creation unit 143, and a process determination unit 145. The next generation population creation unit 143 includes a mutation unit 144.

  The input unit 12 inputs hearing data 17 representing the hearing ability of the person using the hearing aid 16. The recommended characteristic calculation unit 13 calculates a recommended characteristic that is a characteristic that the hearing aid 16 should have, based on the hearing data 17 input by the input unit 12. Based on the recommended characteristic calculated by the recommended characteristic calculation unit 13, the optimal value calculation unit 14 calculates the optimal value of the acoustic parameter that determines the characteristic of the hearing aid 16. The setting unit 15 sets the optimal value of the acoustic parameter calculated by the optimal value calculation unit 14 in the hearing aid 16.

  The hearing aid adjustment device 20 can be configured by a general-purpose device such as a personal computer, a workstation, or the like (hereinafter referred to as a personal computer or the like). In addition, the hearing aid adjustment device 20 can be configured by a dedicated device including a calculation unit such as a CPU, a data storage unit, a data input / output interface and the like. The hearing aid adjustment device 20 is used by being connected to the hearing aid 16.

  The input unit 12 includes an interface for inputting the auditory data 17 to the hearing aid adjustment device 20. Input may be performed by reading a file, or input may be performed by accepting manual input. The functions of the recommended characteristic calculation unit 13 and the optimum value calculation unit 14 can be realized, for example, when a CPU such as a personal computer executes a predetermined program. The setting unit 15 includes an output interface for storing the data calculated by the optimum value calculation unit 14 in the parameter storage unit 18 of the hearing aid 15.

  A hearing aid adjustment program for realizing the functions of the input unit 12, the recommended characteristic calculation unit 13, the optimum value calculation unit 14, and the setting unit 15 is obtained from a storage medium such as a CD-ROM or via a communication line. The hearing aid adjustment device 20 can also be constructed by installing it on an arbitrary personal computer or the like by downloading or the like.

  Next, the operation of the hearing aid adjustment device 20 in the present embodiment will be described with reference to FIGS. 1 and 2. FIG. 2 is a flowchart showing the operation of the hearing aid adjustment apparatus 20 according to the present embodiment.

  FIG. 2 shows the order of adjustment of the hearing aids. When the hearing aid adjustment is started (step 100), the input unit 12 first inputs the hearing data 17 representing the hearing level of the person wearing the hearing aid 16 (step 200). For example, as the hearing data 17, a minimum audible value (HTL) at 250 Hz, 500 Hz, 1000 Hz, 2000 Hz, and 4000 Hz is input.

  Based on the hearing data 17 of the hearing aid wearer input by the input unit 12, the recommended characteristic calculation unit 13 calculates the recommended characteristic of the hearing aid 16 (step 300). A recommended characteristic of the hearing aid 16 is a target gain. For example, the recommended characteristic calculation unit 13 calculates the output sound pressure level characteristic with respect to the input sound pressure of 40 dBSPL, 60 dBBSPL, 90 dBSPL (dBSPL: unit of sound pressure dB Sound Pressure Level) at 250 Hz, 500 Hz, 1000 Hz, 2000 Hz, and 4000 Hz, respectively. calculate.

As an example of the recommended characteristic calculation, a case where only HTL is input as the hearing data 17 will be described. In this case, the recommended characteristic is calculated by calculating a comfort level (MCL: Most Comfortable Loudness Level) and a discomfort level (UCL: Uncomfortable Loudness Level) from the HTL. For example, at 1000 Hz, the values of HTL, MCL, and UCL when HTL (input value) is 40 dBHL (dBHL: dB Hairing Level) are calculated as follows.
HTL = 40 (when HTM (input value) <115, HTL = HTL (input value), otherwise, HTL = 115)
UCL = 100 (If HTL <55, UCL = 100, otherwise UCL = 75 + HTL / 2)
MCL = 76 (MCL = (UCL−HTL) × 0.6 + HTL)
When the recommended characteristics at 40 dBSPL, 60 dBSPL, and 90 dBSPL are calculated from the above HTL, MCL, and UCL, they are as follows.
Recommended 60dBSPL input characteristic value S ₆₀ = 82.5 (S ₆₀ = MCL + 6.5)
Recommended 90dBSPL input characteristic value _{S 90 = 106.5 (S 90 =} UCL + 6.5, where (if UCL + 6.5 value of) is not S ₆₀ +7.5 or S ₆₀ +30 or less, S ₉₀ is within the range Value)
Recommended characteristic value of ₄₀ dBSPL input S ₄₀ = 63.5 (HTL + 6.5, provided that the value of (HTL + 6.5) is not S ₆₀ −20 or more and S ₆₀ −5 or less, S ₄₀ is a value within that range. To do)
As described above, the example of calculating the recommended characteristic has been described for the case where only the HTL is input as the hearing data 17, but the recommended characteristic is similarly applied even when at least one of the MCL and the UCL is input as the audio data 17 in addition to the HTL. Can also be calculated. Note that the recommended characteristic calculation is not limited to the above calculation method.

  The optimal value calculation unit 14 uses the GA method to change the acoustic parameter value so that the characteristic of the hearing aid 16 approaches the recommended characteristic obtained by the recommended characteristic calculation unit 13. Is calculated (step 400).

  The optimum value of the acoustic parameter obtained by the optimum value calculation unit 14 is stored in the acoustic parameter storage unit 18 of the hearing aid 16 by the setting unit 15 (step 500). The stored acoustic parameters are used by the signal processing unit 19 to determine output characteristics when the input signal is processed and output. This completes the hearing aid adjustment of the hearing aid 16 (step 600).

  The detailed procedure of the process (step 400 in FIG. 2) in which the optimum value calculation unit 14 calculates the optimum value of the acoustic parameter using the GA method will be described below. FIG. 3 is a flowchart showing details of processing for calculating the optimum value of the acoustic parameter.

  First, when the process is started (step 401), the individual group creation unit 141 creates an individual to be used in the genetic algorithm (step 402). For example, the individual group creation unit 141 randomly distributes the acoustic parameters to be varied among the acoustic parameters that determine the characteristics of the hearing aid, converts them into bits, and creates a predetermined number of A individuals. The acoustic parameters include, for example, volume, gain of each channel (each band obtained by dividing the frequency into four, etc.), compression ratio indicating the compression ratio between the input level and the output level, and TK level (Threshold) indicating the level at which compression is released. Kneepoint). A set of solutions that can be taken by these acoustic parameters is randomly generated. Each randomly generated solution is bitified. Each bitized solution becomes one individual. That is, one piece of data representing one pattern among values that the acoustic parameter can take. Note that the method for generating a set of acoustic parameter solutions is not limited to this. For example, a set of acoustic parameter solutions may be randomly generated with a value within a preset range. Also, it is not always necessary to convert the acoustic parameter solution into bits. Depending on the situation, you can choose how to represent an individual, such as representing an individual with a real value.

  For each individual created by the individual creation unit 141, the evaluation unit 142 calculates hearing aid characteristics (step 403). The characteristics of the hearing aid are calculated based on the acoustic parameters indicated by the bit data of each individual. For example, output sound pressure level characteristics with respect to input sound pressures of 40 dBSPL, 60 dBSPL, and 90 dBSPL are calculated at 250 Hz, 500 Hz, 1000 Hz, 2000 Hz, and 4000 Hz, respectively. Specifically, the IO characteristics of each channel are calculated from the volume, the AGCo, and the gain, compression ratio, and TK level of each channel (each band obtained by dividing the frequency into four or the like). The output for 40 dBSPL, 60 dBSPL, and 90 dBSPL inputs is obtained from the IO characteristics, and the characteristics are calculated by adding the microphone conversion efficiency, the earphone conversion efficiency, and the IC filter characteristics to this value.

Based on the hearing aid characteristics of each individual, the evaluation unit 142 calculates an evaluation value for each individual (step 404). The evaluation value calculation process includes, for example, a difference between each hearing aid characteristic corresponding to the number of individuals (A) calculated in step 403 and the recommended characteristic of the hearing aid calculated in step 300 in FIG. Each is converted into an evaluation value by the square method.
The smaller the evaluation value, the closer to the recommended characteristic. That is, the smaller the evaluation value of an individual, the higher the fitness of the individual. When A evaluation values are obtained, a new A individual group, that is, a next-generation individual group is created based on the evaluation value based on the current population (steps 407 to 411). Such a process of evaluating each individual and creating a next generation individual group is repeated over a plurality of generations.

  If the number of generations of the created population is 2 or more, in order to leave the best evaluation value of the previous generation, the individual with the lowest evaluation value of the previous generation (the individual with the highest fitness) and the maximum evaluation of the current generation Individuals of values (individuals with the lowest fitness) are exchanged (step 405).

  The process determining unit 145 compares the smallest evaluation value closest to the recommended characteristic among the evaluation values obtained in step 405 with the preset evaluation reference value B (step 406). When the minimum evaluation value is smaller than B, it is determined that the sound is optimized, the sound parameter represented by the individual is adopted as the sound parameter of the hearing aid 16 (step 412), and the optimization process is ended (step 413).

  If the minimum evaluation value is greater than B, the process determination unit 145 compares the number of generations of the created individual group with a preset maximum generation number C (step 407). When the number of generations of the created individual group is larger than C, the acoustic parameter represented by the individual having the smallest evaluation value currently obtained is adopted as the acoustic parameter of the hearing aid (step 412), and the optimization process is performed. The process ends (step 413). When the number of generations of the created population is smaller than C, the next generation population creation processing in steps 408 to 411 is performed. That is, the generation of the next generation population is performed C times.

  As a process for creating a next-generation individual group, first, the next-generation individual group creation unit 143 selects two individuals so that the probability that an evaluation value is good (a value is small) will be selected (step 408).

  The mutation unit 144 performs mutation on the two individuals selected in Step 408 (Step 409). Mutation is performed, for example, by inverting one or more bits of an individual represented by an N-bit binary number according to a preset mutation rate. For example, the mutation unit 144 determines whether to perform bit inversion based on the mutation rate for each solid bit represented by an N-bit binary number. For example, when the mutation rate is 1 (100%), the bit is always inverted, and when the mutation rate is 0.5 (50%), the bit is inverted with a probability of 1/2. The number of inversions in N bits depends on the mutation rate.

  Further, whether or not to perform bit inversion for one solid can be determined by the mutation rate. In this case, in an individual determined to perform bit inversion, one bit (or a plurality of bits) in N bits is inverted. Here, the mutation rate can be changed so as to decrease as the number of generated generations increases. The above-described mutation method is an example, and the mutation method is not limited to this.

  The next-generation population creation unit 143 crosses these two individuals (step 410). The two individuals after the crossover become the next generation individuals. Note that the order of the mutation process in step 409 and the crossover process in step 410 may be reversed. Also, either mutation or crossover can be omitted.

  The processing in steps 408 to 410 is not limited to the present embodiment, and various natural selection and genetic recombination methods that are currently known can be used.

  The next generation individual creation process (steps 408 to 410) is repeated (step 411) until the number of next generation individuals reaches the number of individuals A created in advance. When the number of next-generation individuals becomes A, the processing after step 403 is repeated.

  In the above processing, when the number A of individuals created in step 402 is increased, a larger number of individuals are scattered in a wide search space, so that a local solution is less likely to occur. However, in order to advance one generation, many processes are performed and it takes time. Also, increasing the maximum number of generations C is likely not to fall into a local solution in the same way, but when processing for the maximum number of generations C is performed, more processing occurs as the number of generations increases, and time is increased. Will take.

  In the present embodiment, when the number of generations is small in Step 409, the mutation rate is increased, and the method of decreasing the mutation rate as the number of generations increases is used. Even if the number of generations is not increased, a wider space can be searched, so that it is less likely to fall into a local solution, and an increase in processing time can be prevented.

  Further, in the present embodiment, the case where the mutation rate is changed according to the number of generations is illustrated, but the present invention is not limited to this. Hearing aid is recommended with a small number of individuals and a small number of generations in a wide search space by dynamically changing the number of individuals, the number of individuals divided, the number of generations, etc., which are parameters used in the GA method, according to the number of generations. The characteristics and the characteristics of the hearing aid can be brought close to each other efficiently, and the burden on the hearing aid wearer and the hearing aid adjuster can be reduced.

  As described above, the present invention is useful for shortening the processing time and increasing the accuracy of various parameters in a method of automatically searching for a solution using the GA method.

The block diagram which shows an example of a structure of the hearing aid adjustment apparatus in this Embodiment Flow chart representing the operation of the hearing aid in the present embodiment The flowchart showing the detail of the optimal value calculation process of the acoustic parameter in the hearing aid adjustment apparatus in this Embodiment Flow chart showing operation of conventional hearing aid adjustment device

Explanation of symbols

1 Optimization start processing using conventional genetic algorithm (GA method) 2 Individual creation processing 3 Individual information writing processing to hearing aid 4 Individual evaluation processing 5 Next-stage branch processing 6 Crossover target individual selection processing 7 Mutation Processing 8 Crossover processing 9 Next generation population comparison processing 10 Hearing aid parameter creation processing 11 Conventional hearing aid adjustment end processing 12 Input unit 13 Recommended characteristic calculation unit 14 Optimal value calculation unit 141 Population group creation unit 142 Evaluation unit 143 Next generation population Creation unit 144 Mutation unit 145 Processing judgment unit 15 Setting unit 16 Hearing aid 17 Hearing data 18 Acoustic parameter storage unit 19 Signal processing unit 20 Hearing aid adjustment device 100 Hearing aid adjustment start processing 200 Hearing aid wearer hearing level input processing 300 Recommended characteristics of hearing aid Calculation process 400 Optimization process using genetic algorithm (GA method) 40 1 Optimization start processing using genetic algorithm (GA method) 402 Individual creation processing 403 Individual hearing aid characteristic calculation processing 404 Individual evaluation value calculation processing 405 Minimum evaluation value holding processing 406 Minimum evaluation value comparison processing 407 Generation number comparison processing 408 Crossover target individual selection process 409 Mutation process 410 Crossover process 411 Next generation number comparison process 412 Acoustic parameter creation process from individual 413 Optimization end process using genetic algorithm (GA method) 500 Hearing aid adjustment end process

Claims (8)

A hearing aid adjustment device for adjusting the characteristics of a hearing aid by setting acoustic parameters,
An input unit for inputting data representing the hearing ability of the person using the hearing aid;
A recommended characteristic that is a characteristic that the hearing aid should have, a recommended characteristic calculation unit that calculates based on the data;
An optimum value calculating unit for calculating an optimum value of the acoustic parameter based on the recommended characteristics;
A hearing aid adjustment apparatus comprising: a setting unit that sets an optimum value of the acoustic parameter obtained by the optimum value calculation unit in the hearing aid. The optimal value calculation unit calculates an optimal value of the acoustic parameter using a genetic algorithm,
The optimum value calculation unit includes:
An individual group creation unit that creates a plurality of individuals, with one piece of data representing one pattern of values that the acoustic parameter can take,
Evaluation that evaluates the fitness of each individual by comparing the recommended characteristics with the characteristics of the hearing aid determined by the value of the acoustic parameter represented by each individual in the individual group using the individual group as input data And
It is determined whether or not an individual indicating the optimal value of the acoustic parameter has been obtained, and if it is determined that it has been obtained, a determination unit that ends the optimal value calculation process;
When the determination unit determines that the optimum value has not been obtained, a new generation of the next generation is created based on the evaluation by the evaluation unit, and the individual group is input to the evaluation unit. The hearing aid adjustment device according to claim 1, further comprising: The next generation population creation unit includes a mutation unit that mutates the data of the individual based on a mutation rate,
The hearing aid adjustment device according to claim 2, wherein the mutation rate changes according to the number of generations of the population created by the next-generation population creation unit.   The hearing aid adjustment apparatus according to claim 3, wherein the mutation rate decreases as the number of generations increases. A hearing aid adjustment program for adjusting the characteristics of a hearing aid by setting acoustic parameters,
An input process for inputting data representing the hearing ability of the person using the hearing aid;
Recommended characteristic calculation processing for calculating a recommended characteristic that is a characteristic that the hearing aid should have,
An optimal value calculation process for calculating an optimal value of the acoustic parameter based on the recommended characteristics;
A hearing aid adjustment program for causing a computer to execute a setting process for setting an optimum value of the acoustic parameter obtained by the optimum value calculation process in the hearing aid. The optimal value calculation process is to calculate an optimal value of the acoustic parameter using a genetic algorithm,
The optimum value calculation process includes:
Individual group creation processing for creating a plurality of individuals, with one piece of data representing one pattern of values that the acoustic parameter can take,
Evaluation that evaluates the fitness of each individual by comparing the recommended characteristics with the characteristics of the hearing aid determined by the value of the acoustic parameter represented by each individual in the individual group using the individual group as input data Processing,
It is determined whether or not an individual indicating the optimum value of the acoustic parameter has been obtained, and if it is determined that it has been obtained, a determination process for terminating the optimum value calculation process;
When it is determined that the optimum value is not obtained in the determination process, a new generation of the next generation is created based on the evaluation in the evaluation process, and the evaluation process is performed using the individual group as input data. The hearing aid adjustment program according to claim 5, further comprising: The next generation population creation process includes a mutation process for mutating the data of the individual based on a mutation rate,
The hearing aid adjustment program according to claim 6, wherein the mutation rate changes according to the number of generations of the population created by the next-generation population creation processing.   The hearing aid adjustment program according to claim 7, wherein the mutation rate decreases as the number of generations increases.

Images