JP2013195405A - Position estimation method, device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To acquire local positional information with high accuracy in a position estimation method, device and program.SOLUTION: The same original sound is outputted from three speakers provided on the same plane while adding an artificial echo of a different delay time for each speaker just for a fixed period of time, the fixed period of time in which the artificial echo is added to output of each of speakers and the fixed period of time in which the artificial echo is added to output of another speaker are in an exclusive relationship, and a fixed period of time in which the artificial echo is not added to output of all the three speakers is provided. In a position estimation device, a cepstrum is generated by cepstrum analysis of a sample train obtained by detecting and sampling the output of the three speakers by means of a microphone, a distance order indicating a magnitude relation of distances from a position of the position estimation device to each of the three speakers is acquired on the basis of a peak of the cepstrum, and the position of the position estimation device is estimated on the basis of the distance order and known positions of the three speakes.

Description

  The present invention relates to a position estimation method, apparatus, and program.

  In recent years, an information distribution system using local position information has been proposed as one form of various services provided to consumers. Such an information distribution system distributes a wide variety of information based on observer positions within a limited space. For example, by acquiring position information in a limited space such as a store, it is possible to search for a route to a target display shelf or use it as a position grasping method in the space of an autonomous robot operating in a closed space.

  Conventionally, GPS (Global Positioning System) is used as a position information service. GPS acquires observer position information based on information from a plurality of satellites orbiting the earth. GPS is used in car navigation systems and the like because it can acquire position information at any position on the earth. However, since the practical accuracy of GPS is limited to a few dozen meters, the observation accuracy is not sufficient for spaces such as stores. In addition, GPS may not be used in an environment where it is difficult to receive information from satellites such as indoors and underground, and it is difficult to apply position information acquisition in a local space. For this reason, realization of the local position information acquisition method which can acquire position information acquisition reliably in local space is desired.

  One characteristic required for the local position information acquisition method is the position observation accuracy. The local space is typically limited to a few tens of meters to a few tens of meters, but the observation accuracy of the position in the local space can be used as useful information only if it is about tens of centimeters or less. There is sex.

  Another feature required for the local position information acquisition method is cost. The provision of information services based on observer position information can be implemented in ultra-small units such as retail stores and supermarkets, so that conventional equipment can be used as much as possible or low If it cannot be realized at a low cost, an increase in service providers cannot be expected.

  Local location information acquisition methods using pseudo satellites, electronic tags, and the like have been proposed, but the burden on service providers is large due to the introduction of new hardware and the accompanying increase in costs.

JP 2000-147084 A JP 7-333331 A JP-A-5-157835

  Conventionally, it is difficult to achieve local position information acquisition in an area where GPS is difficult to use or detailed position information acquisition for GPS assistance with high observation accuracy and low cost.

  Accordingly, an object of the present invention is to provide a position estimation method, apparatus, and program capable of acquiring local position information with high accuracy.

  According to one aspect of the present invention, the same original sound is output from three speakers provided on the same plane with artificial echoes having different delay times for each speaker for a certain period of time, and the output of each speaker is artificial. There is an exclusive relationship between a certain period in which the echo is added and a certain period in which the artificial echo is added to the output of another speaker, and a certain period in which no artificial echo is added to the outputs of all three speakers. In the terminal device, a cepstrum is generated by cepstrum analysis of a sample sequence obtained by detecting the outputs of the three speakers with a microphone, and the position of the terminal device is determined based on the peak of the cepstrum in the terminal device. To obtain a distance order indicating a magnitude relationship of distances from the three speakers to each of the three speakers, and in the terminal device, the distance order and Position estimating method characterized by estimating the position of the terminal apparatus based on the known position of the serial three loudspeakers are provided.

  According to one aspect of the present invention, the same original sound is output from three speakers provided on the same plane with artificial echoes having different delay times for each speaker for a certain period of time, and the output of each speaker is artificial. There is an exclusive relationship between a certain period in which the echo is added and a certain period in which the artificial echo is added to the output of another speaker, and a certain period in which no artificial echo is added to the outputs of all three speakers. A position estimation apparatus used in a system provided with a microphone that detects an output of the three speakers, and a processor that executes a position estimation process for estimating the position of the position estimation apparatus based on the detection output of the microphone The position estimation processing generates a cepstrum by cepstrum analysis of a sample sequence obtained by sampling the detection output, and the cepstrum A distance order indicating the magnitude relationship of the distances from the position of the position estimation device to each of the three speakers based on the peak of the position estimation device, and based on the distance order and the known positions of the three speakers A position estimation apparatus is provided that estimates the position of the position estimation apparatus.

  According to one aspect of the present invention, the same original sound is output from three speakers provided on the same plane with artificial echoes having different delay times for each speaker for a certain period of time, and the output of each speaker is artificial. There is an exclusive relationship between a certain period in which the echo is added and a certain period in which the artificial echo is added to the output of another speaker, and a certain period in which no artificial echo is added to the outputs of all three speakers. Is a program that causes a computer of a terminal device to perform position estimation processing, and generates a cepstrum by cepstrum analysis of a sample sequence sampled by detecting outputs of the three speakers with a microphone of the terminal device And a distance from the position of the terminal device to each of the three speakers based on the peak of the cepstrum. A program for causing a computer to execute a procedure for obtaining a distance order indicating a small relationship, and a procedure for estimating the position of the terminal device based on the distance order and the known positions of the three speakers. Is provided.

  According to the disclosed position estimation method, apparatus, and program, local position information can be acquired with high accuracy.

It is a figure explaining an example of the position estimation method in one Example of this invention. It is a figure explaining addition of an artificial echo. It is a block diagram which shows an example of a terminal device. It is a figure explaining an example of a cepstrum analysis result. It is a figure explaining the distance order determination method from a sound source. It is a figure explaining an example of the narrowing-down process of the solution candidate of a user position. It is a flowchart explaining an example of a position estimation process. It is a flowchart explaining an example of a speaker position acquisition process.

  In the disclosed position estimation method, apparatus, and program, the same original sound from three speakers provided on the same plane is added with an artificial echo having a different delay time for each speaker for a certain period and output. There is an exclusive relationship between the fixed period in which the artificial echo is added to the output of each speaker and the fixed period in which the artificial echo is added to the output of other speakers, and the artificial echo is added to the outputs of all three speakers. A certain period of time is provided. In the position estimation device, a cepstrum is generated by cepstrum analysis of a sample sequence obtained by detecting the outputs of the three speakers with a microphone, and the distance from the position of the position estimation device to each of the three speakers based on the peak of the cepstrum. Is obtained, and the position of the position estimation device is estimated based on the distance order and the known positions of the three speakers.

  Embodiments of the disclosed position estimation method, apparatus, and program will be described below with reference to the drawings.

  An example of a position estimation method in one embodiment of the present invention will be described. FIG. 1 is a diagram for explaining an example of a position estimation method in one embodiment of the present invention, and FIG. 2 is a diagram for explaining the addition of an artificial echo.

  In the example shown in FIG. 1, a personal computer (PC: Personal Computer) 11 is an example of a sound source generation device provided on the service provider side. At least three speakers 1A, 1B, and 1C output sound of sound source data output from the PC 11 (hereinafter also referred to as a sound source). The speakers 1A, 1B, and 1C are arranged at the vertices of arbitrary triangles on the same plane indoors, for example, in a store, for example, and the terminal device 2 carried by the user (or an observer) is located within the triangular area. It shall be.

  The terminal device 2 is an example of a position estimation device. The terminal device 2 includes a microphone, a processor, a memory, and the like that detect sound output from the speakers 1A, 1B, and 1C. For example, a mobile phone including a smartphone, a notebook PC, a tablet PC, and a personal digital assistant (PDA: It can be formed by a portable electronic device including a personal digital assistant. The terminal device 2 stores a sample (data) obtained by sampling sounds output from the speakers 1A to 1C in a memory, and the terminal device 2 in a space surrounded by the speakers 1A to 1C based on the stored sound samples. Is estimated. For convenience, in the following description, the position of the terminal device 2 is also referred to as a user position. For example, the same original sound (for example, music) is output from each of the speakers 1A to 1C, but a period in which a sound delayed by a certain time with respect to the original sound is superimposed on the original sound to output a sound source including an artificial echo is provided. That is, the sound recorded by the terminal device 2 is a sound in which the artificial echo sound from each of the speakers 1A to 1C is superimposed in addition to the sound that has arrived from the three speakers 1A to 1C.

  In FIG. 2, (a) shows an example of the waveform of the sound source and the waveform of the sound source with artificial echoes from the speakers 1A, 1B, 1C, and (b) shows the sound source output from each speaker 1A, 1B, 1C. The presence / absence of an artificial echo is shown every t1 for a certain period. As shown in FIG. 2, each of the speakers 1A to 1C outputs a sound source to which an artificial echo is added only for a certain period. Further, the delay times ta, tb, tc of the artificial echo added to the outputs of the speakers 1A to 1C are different for each speaker 1A to 1C (that is, specific to each channel Ch-A to Ch-C described later). And the period in which the artificial echo is added to the output of each speaker (for example, speaker 1A) and the period in which the artificial echo is added to the output of other speakers (for example, speakers 1B and 1C) Is in an exclusive relationship. Furthermore, a period for outputting a sound source in which no artificial echo is added to all the speakers 1A to 1C, that is, a sound source without echo is provided.

  In FIG. 1, the PC 11 has a channel Ch-A playback system, a channel Ch-B playback system, and a channel Ch-C playback system in order to perform multi-channel playback of the digital sound source data 100. The reproduction system of channel Ch-A includes an adder circuit 12A, a delay circuit 13A, a selector 14A, a digital / analog converter (DAC) 15A, and an amplifier 16A. The reproduction system for channel Ch-B includes an adder circuit 12B, a delay circuit 13B, a selector 14B, a DAC 15B, and an amplifier 16B. The reproduction system for channel Ch-C includes an adder circuit 12C, a delay circuit 13C, a selector 14C, a DAC 15C, and an amplifier 16C. The outputs of the amplifiers 16A, 16B, and 16C are connected to the corresponding speakers 1A, 1B, and 1C.

  The PC 11 has a known hardware configuration including a processor such as a CPU (Central Processing Unit) and a storage unit such as a memory, and functions of a delay circuit unit and a selector unit in the channels Ch-A to Ch-C playback system ( The delay process and the selection process can be realized by the processor executing a program stored in the storage unit. The functions of the DAC unit and the amplifier unit can be realized by using a reproduction system such as an external sound card that supports multi-channel reproduction or a sound chip mounted on the motherboard of the PC 11. The program may be stored in an appropriate computer-readable storage medium such as a semiconductor storage device, a magnetic recording medium, an optical recording medium, or a magneto-optical recording medium. Needless to say, the functions of the delay circuit section and the selector section of the channels Ch-A to Ch-C reproduction system may be realized by hardware.

  In the reproduction system of channel Ch-A, the adding circuit 12A adds the digital sound source data 100 and the digital sound source data 100 delayed by the delay time ta by the delay circuit 13A. The selector 14A selects the digital sound source data 100 and the sound source data with artificial echoes from the addition circuit 12A obtained by adding the digital sound source data 100 delayed by the delay time ta to the digital sound source data 100 in a predetermined order and interval. Select to output. The output sound source data of the selector 14A is digital / analog (D / A) converted by the DAC 15A, amplified by the amplifier 16A, and then supplied to the speaker 1A. Since the operations of the channel Ch-B playback system and the Ch-C playback system are the same as the processing of the channel playback system Ch-A, description thereof will be omitted. However, the delay circuit 13B has a delay time tb, and the delay circuit 13C has a delay time tc.

  For example, when the predetermined period (or time) t1 is 100 milliseconds (msec), the PC 11 determines that “the artificial echo is not added to the outputs of all the speakers 1A, 1B, and 1C” and “the artificial echo is output to the speaker 1A. The four states of “with addition”, “with artificial echo added to the output of the speaker 1B”, and “with artificial echo added to the output of the speaker 1C” are repeated at a cycle of 400 msec. In this case, the order of the speakers 1A to 1C that output the sound source to which the artificial echo is added is arbitrary, and three such as “speaker 1B”, “speaker 1C”, “no echo”, and “speaker 1A”. The order in which the artificial echo addition periods of the outputs of the speakers 1A to 1C are not continuous may be used. In the following description, for the sake of convenience of explanation, the delay time ta = 0.3 msec of the artificial echo of the speaker 1A, the delay time tb of the artificial echo of the speaker 1B, and the delay time tc = 0.7 msec of the artificial echo of the speaker 1C. Shall. The artificial echo delay time ta to tc of each speaker 1A to 1C is not limited, but in general, considering that it is difficult for humans to distinguish echoes of 1 msec or less, the delay of the artificial echo to be added It is desirable to keep the time within a range that humans cannot discern.

  FIG. 3 is a block diagram illustrating an example of a terminal device. 3 includes a processor 21 such as a CPU, a direct memory access controller (DMAC) 22, timers 23 and 24, a memory 25, a microphone 26, a temperature sensor 27, a camera 28, a communication unit 29, and an operation unit 30. Are connected by a bus 40. In this example, the terminal device 2 forms a smart phone, but the terminal device 2 is not limited to a smart phone.

  In this example, the processor 21 is connected to the instruction memory 31 and the data memory 32. However, the processor 21 and the memory used by the processor 21 are not limited to the configuration and connection shown in FIG. The structure which is not used may be sufficient. If the DMA is not used, the DMAC 22 can be omitted. Further, the instruction memory 31 and the data memory 32 can be omitted by using the memory 25. Furthermore, the timers 23 and 24 can be omitted by using the internal timer of the processor 31. As will be described later, the temperature sensor 27 can be omitted if the sound speed is not corrected according to the environmental temperature, and the timer 24 can be omitted if the temperature sensor 27 is omitted. The microphone 26 is, for example, a monaural microphone.

  The camera 28, the communication unit 29, and the operation unit 30 are an example of an input unit for the user to acquire the positional information of the speakers 1A to 1C, and the camera 28 and the communication unit 29 can be omitted. The operation unit 30 includes an input unit such as a keyboard and a display unit such as a display, and may have a configuration in which the input unit and the display unit are integrated as in a touch panel.

  The processor 21 executes the position estimation program stored in the instruction memory 31, thereby estimating the position of the terminal device 2 based on the sound from the speakers 1A, 1B, and 1C detected by the microphone 26. Data such as parameters used in calculations performed by the processor 21 and intermediate data of the calculations are temporarily stored in the data memory 32. The processor 21 samples the sound from the speakers 1 </ b> A to 1 </ b> C detected by the microphone 26 at a timing determined by the interrupt signal output from the DMAC 22 based on the first and second transfer requests from the timers 23 and 24. Thus, the analog / digital (A / D) conversion is performed, and the sample is transferred to the memory 25 and stored. The timer 23 generates a first transfer request at a cycle for sampling the output of the microphone 26, and the timer 24 generates a second transfer request at a cycle for sampling the output of the temperature sensor 27. The period for sampling the output of the microphone 26 and the period for sampling the output of the temperature sensor 27 can be set from the processor 21 to the timers 23 and 24, for example. Since the environmental temperature of the terminal device 2 rarely fluctuates greatly, the cycle set in the timer 24 is, for example, 1 second, which is longer than the cycle set in the timer 23. Further, the processor 21 sets an interrupt signal generation cycle in the DMAC 22 so that an interrupt signal is generated when a predetermined number of samples are transferred to the memory 25. When the DMAC 22 receives the first transfer request from the timer 23 or receives the second transfer request from the timer 24, the DMAC 22 transfers the data detected by the microphone 26 to the memory 25, so that the sampled sample is obtained. Store in the memory 25.

  When the processor 21 receives the interrupt signal from the DMAC 22, the processor 21 starts the following position estimation process based on the samples stored in the memory 25. That is, the processor 21 performs cepstrum analysis on the superimposed sound sample stored in the memory 25. Cepstrum analysis is performed based on the following equation.

Cepstrum = IFFT (log (abs (FFT (X))))
Here, X is a vector of digitally recorded samples, FFT is a fast Fourier transform, abs is an absolute value acquisition, log is a logarithm (the base of the logarithm is arbitrary), and IFFT is a function that performs an inverse fast Fourier transform. The vector sequence obtained by the above formula is called a cepstrum, and in this example, it is used to analyze whether or not a superimposed sound is included in the input sound source.

  FIG. 4 is a diagram for explaining an example of a cepstrum analysis result. In FIG. 4, (a) shows the original sound by a one-dot chain line, the echo sound by a broken line, and the sound source including the echo by a solid line. When echo sound delayed by time T is superimposed on the input sound source, the result of cepstrum analysis shows that the position of T in the cepstrum space (the horizontal axis is relative to the time axis as shown in FIG. 4B). A sharp peak in the cepstrum amplitude can be observed. From this result, it can be seen that an echo sound delayed by time T is superimposed on the original sound. The processor 21 can acquire the difference (or difference distance) in the distance between the user position and each of the speakers 1A to 1C from the cepstrum analysis result. This is because a minute arrival delay determined by the user position and sound speed occurs in the sound output from the plurality of speakers 1A to 1C, and the sound recorded by the terminal device 2 is the sound output from the nearest speaker. This is because the sound of the speaker that arrives late is superimposed. That is, by performing cepstrum analysis, the difference in arrival time between the closest speaker sound and the delayed speaker sound is known. By calculating the product of this known arrival time difference and sound speed by multiplying, the difference in distance between the user position and these two speakers can be calculated.

  In addition, since the time for the artificial echo sound to reach the user is obtained by adding the delay time of the artificial echo to the arrival delay of the speaker, as shown in FIG. 5, the appearance position of the cepstrum peak in each artificial echo section is specified. By doing this, it is possible to acquire a distance order (that is, a distant or close order) indicating the magnitude relationship of the distances from the user position to the three speakers 1A to 1C. FIG. 5 is a diagram for explaining a method for determining a distance order from a sound source.

  The position estimation process is started at an arbitrary timing. In the example of FIG. 5, it is assumed that recording starts in a period in which no artificial echo is added, as shown in FIG. In this case, the processor 21 performs A / D conversion on the sound data detected by the microphone 26 in units of 50 msec, for example, and acquires a sample sequence of the sound source. As described above, the sampling period ts is set to 50 msec. When each artificial echo addition period (or a fixed period) t1 is 100 msec, if the sampling period ts is set to 100 msec, two sampling periods ts are set. This is because there is a possibility that normal cepstrum analysis may not be possible because the artificial echo is added. For this reason, by setting the sampling period ts to 50 msec (that is, ts = t1 / 2), the cepstrum analysis can be performed without straddling different artificial echo addition periods at least once.

  In this way, when cepstrum analysis is performed on the obtained sample sequence in units of ts = 50 msec, a period during which two peaks can be observed in the cepstrum analysis result is generated, and this is the result of cepstrum analysis without artificial echo addition. And In FIG. 5, such a peak is indicated by a circle. Since the sound source output from the other speaker reaches the microphone 26 with a delay corresponding to the sound speed, the three speakers 1A to 1C are used as in this embodiment. In this case, two echoes are superimposed on one sound source in the period without artificial echo. In Fig.5 (a), it shows by T1, T2 from the one where a position is near among two peaks.

  The cepstrum analysis is not performed during the 50 msec period after the period when two peaks can be observed in the cepstrum analysis result. This is because the artificial echo addition period (or fixed period) t1 is 100 msec, while the sampling period ts is 50 msec, so the next period is acquired 50 msec from the period in which two peaks are observed. This is because the artificial echo addition period can be acquired without straddling the period. Thus, in the next period after 50 msec, the original sound with the artificial echo added is output from the speaker 1A, so a cepstrum analysis is performed. In the cepstrum generated from the sampled sample sequence of 50 msec by this cepstrum analysis, as shown in FIG. 5B, one more peak is observed with respect to the previous two peaks. When this new peak appears at the position of the cepstrum origin + ta (for example, ta = 0.3 msec), the speaker 1A is closest to the user, and when it appears at the position of T1 + ta, the speaker 1A is second from the user. If it appears far away (or the second nearest) and appears at the position of T2 + ta, it can be seen that the speaker 1A is farthest from the user. This is because the echo output from the speaker 1A arrives at the user by the sum of the arrival delay of the original sound of the speaker 1A and the sum of ta, and therefore further to the cepstrum positions 0, T1, T2 indicating the position of the speaker. This is because a cepstrum peak should appear at a delayed position. In the example shown in FIG. 5B, since a new peak appears at the position of T1 + ta, the distance order of the speakers 1A, 1B, and 1C viewed from the user is B <A <C or C <A <. It turns out that it is B.

  Similarly, the cepstrum analysis is not performed for the next 50 msec period, the cepstrum analysis is performed for the next 50 msec period, and a cepstrum peak newly appears at the position of T2 + tb (eg, tb = 0.5 msec). Assuming that the speaker 1B is located farthest from the user by the same logic as that of the speaker 1A. Therefore, it can be seen that the distance order of the speakers 1A, 1B, and 1C viewed from the user is C <A <B.

  The processor 21 estimates the user position based on the distance order of the speakers 1A, 1B, and 1C obtained as described above and the coordinate values indicating the known positions of the speakers 1A, 1B, and 1C. From the above processing results, it can be seen that the difference in arrival time of the sound from the speakers 1C and 1A to the user is T1. Since the sound speed in the air is 345 m / sec, the distance difference Lca between the speaker 1A and the speaker 1C viewed from the user is obtained by multiplying T1 by the value of this sound speed. The distance difference Lcb between the speaker 1C and the speaker 1B as seen from the user can also be obtained by the same method.

  In general, it is known that when a point having a constant distance difference between two points is plotted, a hyperbolic function is obtained. Therefore, the focal points of the hyperbola are the speakers 1C and 1A, and a hyperbola α is drawn so that the distance difference is Lca. Similarly, the focal points of the hyperbola are the speakers 1C and 1B, and a hyperbola β whose distance difference is Lcb can be drawn. Since it is clear that the user is in a position satisfying both of these two hyperbolic curves α and β, the user position can be determined by obtaining the intersection of the two simultaneous hyperbolic equations. However, since a plurality of solutions are derived from the two simultaneous hyperbolic equations, the user position must be narrowed down from a plurality of solution candidates.

  In the present embodiment, since three speakers 1A to 1C are used, there may be three hyperbolic curves. However, two hyperbolic equations are selected and two simultaneous hyperbolic equations are constructed, and two hyperbolic equations are selected. A plurality of solution candidates are set as user position candidates.

  In the present embodiment, a plurality of solution candidates for the two simultaneous hyperbolic equations are narrowed down by the process described with reference to FIG. FIG. 6 is a diagram illustrating an example of a narrowing process of solution candidates for user positions. In FIG. 6, 1A to 1C indicate positions of the speakers 1A to 1C. First, each hyperbola α, β is composed of two curves as shown by a one-dot chain line and a two-dot chain line in FIG. 6, but the distance order of the speakers 1A to 1C as viewed from the user by the above processing. It can be seen that C <A <B. For this reason, based on the distance order of each speaker 1A-1C seen from the user, for example, on two hyperbola α focusing on the speakers 1A, 1C, the user exists on one hyperbola α on the side closer to the speaker 1C. I understand that Similarly, on the two hyperbola β focusing on the speakers 1B and 1C based on the distance order of the speakers 1A to 1C as seen from the user, the user exists on one hyperbola β closer to the speaker 1C. I understand that. Thereby, candidate points (or solution candidates) can be narrowed down to two points P1 and P2. Further, on the assumption that the user exists in an area surrounded by the speakers 1A to 1C, the user is finally limited to one candidate point, that is, P1 existing in a triangular area surrounded by the speakers 1A to 1C. . As a result, it is possible to estimate the coordinates of the user position from the candidate point P1.

  The method for deriving the solution of the two simultaneous hyperbolic equations is not limited to the above example, and other appropriate methods may be used. For example, in the case of a smart phone, it is possible to derive a solution of two simultaneous hyperbolic equations by using an operation library prepared in an application processor that incorporates even a basic application in the CPU.

  FIG. 7 is a flowchart for explaining an example of the position estimation process executed by the processor 21. When the position estimation process is started, in step S1, the sound data detected by the microphone 26 is A / D converted, for example, in units of ts = 50 msec to obtain a sample sequence of the sound source. In step S2, cepstrum analysis of the sample sequence acquired in step S1 is performed. Step S3 determines whether there are two peaks in the cepstrum analysis result. If the determination result is NO, the process returns to step S1. When two peak periods occur, the determination result in step S3 is YES, and the processing proceeds to step S4, interpreting that synchronization has been achieved.

  In step S4, the cepstrum analysis for this period is not performed by discarding the sample string for the next period of ts = 50 msec. In step S5, a sample sequence of the artificial echo addition period ts output from the speaker 1A is acquired, and in step S6, a cepstrum analysis is performed on the sample sequence of the artificial echo addition period ts output from the speaker 1A. In step S7, the cepstrum analysis for this period is not performed by discarding the sample string for the next period of ts = 50 msec. Step S8 obtains a sample string of the artificial echo addition period ts output from the speaker 1B, and step S9 performs cepstrum analysis of the sample string of the artificial echo addition period ts output from the speaker 1B. In step S10, the cepstrum analysis for this period is not performed by discarding the sample string for the next period of ts = 50 msec. Step S11 obtains a sample sequence of the artificial echo addition period ts output from the speaker 1C, and step S12 performs a cepstrum analysis of the sample sequence of the artificial echo addition period ts output from the speaker 1C.

  In step S13, the distance order from the three speakers 1A to 1C to the user position is determined using the distance order determination method described in conjunction with FIG. In the above example, the distance order of the speakers 1A, 1B, and 1C as viewed from the user is determined as C <A <B. In step S14, as described above, a plurality of candidate hyperbolic equations are calculated by drawing two hyperbolic curves having the focal point of the closest speaker and deriving the intersection of the two hyperbolic curves. In step S15, as described with reference to FIG. 6, a plurality of solution candidates are narrowed down from the distance order of the speakers 1A to 1C already determined in step S13, and the user is present in an area surrounded by the speakers 1A to 1C. The coordinates of the user position are determined by narrowing down the solution candidates to one using the premise. After step S15, the position estimation process ends.

  The position estimation program may be stored in an appropriate computer-readable storage medium such as a semiconductor storage device, a magnetic recording medium, an optical recording medium, or a magneto-optical recording medium.

  In the embodiment described above, it is assumed that the coordinate values of the positions of the speakers 1 </ b> A to 1 </ b> C are known and stored in the terminal device 2. The method of inputting the coordinate values of the positions of the speakers 1A to 1C to the terminal device 2 is not particularly limited. For example, even if the user inputs the position information of the speaker presented by the service provider (the coordinate values of the positions of the speakers 1A to 1C) from the operation unit 30 of the terminal device 2, the user presents it on the home page of the service provider Even if the position information of the speaker is input through the Internet access via the communication unit 29, the position information of the speaker presented by the service provider is captured by the camera 28 and input by character recognition of the captured image. Even if the user captures a code such as a QR (Quick Response) code (registered trademark) related to the position information of the speaker presented by the service provider with the camera 28 and inputs it by image recognition of the captured image. good.

  FIG. 8 is a flowchart for explaining an example of the speaker position acquisition process executed by the processor 21. The speaker position acquisition process is executed before the position estimation process described above. In this example, the service provider provides a code such as a QR code (registered trademark) relating to the coordinate values (speaker position information) of the positions of the speakers 1A to 1C in advance in the store. Assume that the camera 28 takes a QR code. In step S21, it is determined whether a QR code can be captured by the camera 28 or not. If the determination result in step S21 is YES, the process proceeds to step S22, and if the determination result is NO, the process proceeds to step S23. In step S22, the coordinate values of the positions of the speakers 1A to 1C in the store are acquired by performing image recognition of the captured image of the QR code, and stored in the memory 25 or the data memory 32. In step S23, the coordinate values of the positions of the speakers 1A to 1C previously presented by the service provider are acquired by a method other than the QR code. Specifically, the user inputs the coordinate values of the positions of the speakers 1A to 1C that the service provider presents in advance by printed matter or the like from the operation unit 30 of the terminal device 2, or is presented on the home page of the service provider. The coordinate value of the position of the speaker is input by Internet access via the communication unit 29, or the coordinate value of the position of the speaker presented by the service provider on the bulletin board in the store is photographed by the camera 28. It is possible to input by image character recognition, and the input coordinate value of the speaker position is stored in the memory 25 or the data memory 32. After step S22 or step S23, the speaker position acquisition process ends.

  The estimation accuracy of the user position depends on the sampling rate of the microphone 26 of the terminal device 2 and the ADC (Analog-to-Digital Converter) function of the processor 21 that samples the output of the microphone 26 and performs A / D conversion. For example, when the sampling rate is 48 kHz, the user position can be estimated with an accuracy of 1/48 (kHz) × 345 (m / sec) ≈0.72 cm. In the above embodiment, it is assumed that the sound speed is 345 m / sec. However, since the sound speed fluctuates according to the air temperature, the sound speed value is set according to the environmental temperature detected by the temperature sensor 27 of the terminal device 2. It may be corrected. An example of the sound velocity correction formula used in the correction process of the processor 21 when the temperature sensor 27 is used is as follows.

Speed of sound = 331 + 0.6Temp [m / sec]
Here, Temp is a Celsius temperature detected by the temperature sensor 27. Thus, by correcting the value of the sound speed according to the temperature, it becomes possible to estimate the user position with higher accuracy.

  According to the disclosed position estimation method, apparatus, and program, a local position can be estimated with high accuracy at low cost. That is, it is possible to construct a simple and low-cost local position estimation system that uses only legacy devices such as PCs, speakers, and terminal devices.

  Speakers required on the service provider side can be used at a relatively low cost, and a general-purpose computer such as a PC is used to add artificial echoes with different delays to the sound output from each speaker. It is sufficient to create a sound source in advance. Further, in order to output a sound source to which different artificial echoes are added from each speaker, it can be easily realized by using a multi-channel playback device such as an existing 5.1 channel AV amplifier, for example.

  A terminal device required on the user side is a microphone and a processor for executing position estimation processing, and these can be realized by a device equipped with an application processor such as a general smartphone or tablet PC.

  In addition, by embedding position information in the sound, for example, information can be concealed in in-store broadcasts, music flowing in the store, and the like, and there is no sense of incongruity in terms of compatibility with existing services and user-side usage.

  The position information acquired by the user through the position estimation process can be used as guide information to the user by, for example, showing it on a map or layout displayed on the display unit of the terminal device. On the other hand, when the service provider can acquire the position information acquired by the user, guidance according to the user position can be provided to the user. If the user position can be acquired, the service provider transmits guidance information such as explanation of products displayed on the product shelf installed near the user position in the store to the terminal device by voice or text information to the user. Voice guidance information can be provided from a speaker near the product shelf.

  Note that the signal processing is somewhat complicated as compared with the case where the number of speakers is three as in the above embodiment, but the disclosed position estimation method, apparatus and program are the same when the number of speakers is four or more. It is applicable to.

The following additional notes are further disclosed with respect to the embodiment including the above examples.
(Appendix 1)
The same original sound is output from three speakers provided on the same plane with artificial echoes having different delay times for each speaker only for a certain period, and the artificial echo is added to the output of each speaker. There is an exclusive relationship with a fixed period in which artificial echoes are added to the output of other speakers, and a fixed period in which artificial echoes are not added to the outputs of all three speakers is provided,
In the terminal device, the cepstrum is generated by the cepstrum analysis of the sample sequence obtained by detecting the output of the three speakers with the microphone,
In the terminal device, based on the cepstrum peak, obtain a distance order indicating a magnitude relationship of the distance from the position of the terminal device to each of the three speakers,
In the terminal device, the position estimation method estimates the position of the terminal device based on the distance order and the known positions of the three speakers.
(Appendix 2)
The position of the terminal device is estimated as follows:
The distance difference Lca between the pair of speakers viewed from the terminal device is determined from the product of the arrival time difference between the outputs of the pair of speakers among the three and the sound velocity,
The distance difference Lcb between the other pair of speakers viewed from the terminal device is obtained from the product of the arrival time difference between the output of the other pair of speakers among the three and the sound speed, and the sound speed,
A hyperbola α whose distance difference is Lca with the pair of speakers as the focal point, and a hyperbola β whose distance difference is Lcb with the other pair of speakers as the focal point,
The position estimation method according to claim 1, wherein the position of the terminal device is determined from an intersection of simultaneous hyperbolic equations satisfying both the hyperbola α and β.
(Appendix 3)
The position estimation method according to claim 2, wherein a plurality of solution candidates for the simultaneous hyperbolic equations are narrowed down based on a distance order of the three speakers as viewed from the terminal device.
(Appendix 4)
The plurality of solution candidates of the simultaneous hyperbolic equations are narrowed down based on the distance order, and then are narrowed down to solution candidates in a triangular region whose vertex is the position of the three speakers. Position estimation method.
(Appendix 5)
The position estimation method according to any one of appendices 1 to 4, wherein in the terminal device, the value of the sound velocity is corrected according to an environmental temperature detected by a temperature sensor.
(Appendix 6)
The cepstrum analysis
Cepstrum = IFFT (log (abs (FFT (X))))
X is a vector of digitally recorded samples, FFT is a fast Fourier transform, abs is an absolute value acquisition, log is a logarithm (the base of the logarithm is arbitrary), IFFT is a function that performs inverse fast Fourier transform The position estimation method according to any one of appendices 1 to 5, wherein the position estimation method is provided.
(Appendix 7)
In the terminal device, information indicating the known positions of the three speakers is photographed by a camera, and the coordinate values of the known positions recognized by image recognition or character recognition of the photographed information are stored in a memory. The position estimation method according to any one of appendices 1 to 6, wherein:
(Appendix 8)
The same original sound is output from three speakers provided on the same plane with artificial echoes having different delay times for each speaker only for a certain period, and the artificial echo is added to the output of each speaker. Position estimation apparatus used in a system that has an exclusive relationship with a fixed period in which artificial echo is added to the output of another speaker and has a fixed period in which no artificial echo is added to the outputs of all three speakers Because
A microphone for detecting the output of the three speakers;
A processor that executes position estimation processing for estimating the position of the position estimation device based on the detection output of the microphone;
The position estimation process includes:
A cepstrum is generated by cepstrum analysis of a sample sequence obtained by sampling the detection output,
Based on the cepstrum peak, obtain a distance order indicating the magnitude relationship of the distance from the position of the position estimation device to each of the three speakers,
A position estimation device that estimates the position of the position estimation device based on the distance order and the known positions of the three speakers.
(Appendix 9)
The position estimation process includes:
The distance difference Lca between the pair of speakers as viewed from the position estimation device is obtained from the product of the arrival time difference of the outputs of the pair of speakers to the microphone and the sound speed of the three speakers,
The distance difference Lcb between the other pair of speakers as viewed from the position estimation device is obtained from the product of the arrival time difference between the output of the other pair of speakers to the microphone and the sound speed of the three pairs,
A hyperbola α whose distance difference is Lca with the pair of speakers as the focal point, and a hyperbola β whose distance difference is Lcb with the other pair of speakers as the focal point,
The position estimation apparatus according to appendix 8, wherein the position estimation apparatus determines a position of the position estimation apparatus from an intersection of simultaneous hyperbolic equations satisfying both the hyperbola α and β.
(Appendix 10)
The position estimation apparatus according to appendix 9, wherein the position estimation process narrows down a plurality of solution candidates of the simultaneous hyperbolic equations based on a distance order of the three speakers as viewed from the position estimation apparatus.
(Appendix 11)
In the position estimation process, after narrowing down a plurality of solution candidates of the simultaneous hyperbolic equations based on the distance order, the solution is narrowed down to solution candidates in a triangular region having the positions of the three speakers as vertices. The position estimation apparatus according to appendix 10.
(Appendix 12)
A temperature sensor for detecting the environmental temperature;
The position estimation apparatus according to any one of appendices 8 to 11, wherein the position estimation process corrects the value of the sound velocity according to an environmental temperature detected by the temperature sensor.
(Appendix 13)
In the position estimation process, the cepstrum analysis is performed.
Cepstrum = IFFT (log (abs (FFT (X))))
X is a vector of digitally recorded samples, FFT is a fast Fourier transform, abs is an absolute value acquisition, log is a logarithm (the base of the logarithm is arbitrary), IFFT is a function that performs inverse fast Fourier transform The position estimation apparatus according to any one of appendices 8 to 12, wherein the position estimation apparatus is provided.
(Appendix 14)
A camera,
Further comprising a memory,
When the camera captures information indicating a known position of the three speakers, the processor stores the coordinate value of the known position recognized by image recognition or character recognition of the captured information in the memory. 14. The position estimation device according to any one of appendices 8 to 13, characterized in that:
(Appendix 15)
The same original sound is output from three speakers provided on the same plane with artificial echoes having different delay times for each speaker only for a certain period, and the artificial echo is added to the output of each speaker. In a system in which a fixed period in which artificial echoes are added to the outputs of other speakers is provided and a fixed period in which artificial echoes are not added to the outputs of all three speakers is provided, A program for causing a computer to execute position estimation processing,
A procedure for generating a cepstrum by cepstrum analysis of a sample sequence obtained by detecting outputs of the three speakers with a microphone of the terminal device;
A procedure for obtaining a distance order indicating a magnitude relationship of distances from the position of the terminal device to each of the three speakers based on the peak of the cepstrum;
A program for causing the computer to execute a procedure for estimating the position of the terminal device based on the distance order and the known positions of the three speakers.
(Appendix 16)
The procedure for estimating the position of the terminal device is as follows:
A procedure for obtaining a distance difference Lca between the pair of speakers as viewed from the position estimation device from a product of an arrival time difference of the output of the pair of speakers to the microphone and a sound speed among the three;
A procedure for obtaining a distance difference Lcb between the other pair of speakers as viewed from the position estimation device from a product of arrival time difference between the output of the other pair of speakers to the microphone and sound speed among the three pairs,
A procedure for obtaining a hyperbola α whose distance difference is Lca when the focal point is the pair of speakers, and a hyperbola β whose distance difference is Lcb when the focal point is the other pair of speakers,
The program according to claim 15, further comprising a procedure for determining a position of the position estimation device from an intersection of simultaneous hyperbolic equations satisfying both the hyperbolic α and β.
(Appendix 17)
The procedure of estimating the position of the terminal device further includes a procedure of narrowing down a plurality of solution candidates of the simultaneous hyperbolic equations based on a distance order of the three speakers viewed from the position estimation device. The program according to appendix 16.
(Appendix 18)
The procedure for estimating the position of the terminal device is to narrow down a plurality of solution candidates of the simultaneous hyperbolic equations based on the distance order, and then select a solution candidate in a triangular region having the positions of the three speakers as vertices. The program according to appendix 17, further comprising a narrowing-down procedure.
(Appendix 19)
The program according to any one of appendices 15 to 18, which causes the computer to execute a procedure of correcting the value of the sound velocity according to an environmental temperature detected by a temperature sensor of the terminal device.
(Appendix 20)
The procedure for generating the cepstrum includes the cepstrum analysis.
Cepstrum = IFFT (log (abs (FFT (X))))
X is a vector of digitally recorded samples, FFT is a fast Fourier transform, abs is an absolute value acquisition, log is a logarithm (the base of the logarithm is arbitrary), IFFT is a function that performs inverse fast Fourier transform 20. The program according to any one of appendices 15 to 19, which is characterized in that it exists.
(Appendix 21)
Storing the coordinate values of the known positions recognized by image recognition or character recognition of information indicating the known positions of the three speakers taken by the camera of the terminal device in the memory of the terminal device; 21. The program according to any one of appendices 15 to 20, wherein the program is executed by a computer.

  Although the disclosed position estimation method, apparatus, and program have been described in the above embodiments, the present invention is not limited to the above embodiments, and various modifications and improvements can be made within the scope of the present invention. Needless to say.

1A to 1C Speaker 2 Terminal device 11 PC
12A to 12C Adder circuit 13A to 13C Delay circuit 14A to 14C Selector 15A to 15C DAC
16A-16C Amplifier 21 Processor 22 DMAC
25 Memory 26 Microphone 30 Operation unit

Claims (6)

The same original sound is output from three speakers provided on the same plane with artificial echoes having different delay times for each speaker only for a certain period, and the artificial echo is added to the output of each speaker. Position estimation apparatus used in a system that has an exclusive relationship with a fixed period in which artificial echo is added to the output of another speaker and has a fixed period in which no artificial echo is added to the outputs of all three speakers Because
A microphone for detecting the output of the three speakers;
A processor that executes position estimation processing for estimating the position of the position estimation device based on the detection output of the microphone;
The position estimation process includes:
A cepstrum is generated by cepstrum analysis of a sample sequence obtained by sampling the detection output,
Based on the cepstrum peak, obtain a distance order indicating the magnitude relationship of the distance from the position of the position estimation device to each of the three speakers,
A position estimation device that estimates the position of the position estimation device based on the distance order and the known positions of the three speakers. The position estimation process includes:
The distance difference Lca between the pair of speakers as viewed from the position estimation device is obtained from the product of the arrival time difference of the outputs of the pair of speakers to the microphone and the sound speed of the three speakers,
The distance difference Lcb between the other pair of speakers as viewed from the position estimation device is obtained from the product of the arrival time difference between the output of the other pair of speakers to the microphone and the sound speed of the three pairs,
A hyperbola α whose distance difference is Lca with the pair of speakers as the focal point, and a hyperbola β whose distance difference is Lcb with the other pair of speakers as the focal point,
3. The position estimating apparatus according to claim 2, wherein the position of the position estimating apparatus is determined from an intersection of simultaneous hyperbolic equations satisfying both the hyperbola α and β. The position estimation device according to claim 2, wherein the position estimation process narrows down a plurality of solution candidates of the simultaneous hyperbolic equations based on a distance order of the three speakers viewed from the position estimation device. . In the position estimation process, after narrowing down a plurality of solution candidates of the simultaneous hyperbolic equations based on the distance order, the solution is narrowed down to solution candidates in a triangular region having the positions of the three speakers as vertices. The position estimation apparatus according to claim 3. The same original sound is output from three speakers provided on the same plane with artificial echoes having different delay times for each speaker only for a certain period, and the artificial echo is added to the output of each speaker. In a system in which a fixed period in which artificial echoes are added to the outputs of other speakers is provided and a fixed period in which artificial echoes are not added to the outputs of all three speakers is provided, A program for causing a computer to execute position estimation processing,
A procedure for generating a cepstrum by cepstrum analysis of a sample sequence obtained by detecting outputs of the three speakers with a microphone of the terminal device;
A procedure for obtaining a distance order indicating a magnitude relationship of distances from the position of the terminal device to each of the three speakers based on the peak of the cepstrum;
A program for causing the computer to execute a procedure for estimating the position of the terminal device based on the distance order and the known positions of the three speakers. The same original sound is output from three speakers provided on the same plane with artificial echoes having different delay times for each speaker only for a certain period, and the artificial echo is added to the output of each speaker. There is an exclusive relationship with a fixed period in which artificial echoes are added to the output of other speakers, and a fixed period in which artificial echoes are not added to the outputs of all three speakers is provided,
In the terminal device, the cepstrum is generated by the cepstrum analysis of the sample sequence obtained by detecting the output of the three speakers with the microphone,
In the terminal device, based on the cepstrum peak, obtain a distance order indicating a magnitude relationship of the distance from the position of the terminal device to each of the three speakers,
In the terminal device, the position estimation method estimates the position of the terminal device based on the distance order and the known positions of the three speakers.

Images