US20150177000A1 - Music-Based Positioning Aided By Dead Reckoning - Google Patents

Abstract

Music-based positioning (MP) provides positioning service only sporadically. To continuously provide positioning service, the present invention discloses music-based positioning aided by dead reckoning (MP-DR). At each signature burst (i.e., a highly unique short musical segment suitable for positioning), sounds of a music piece or a human speech are used for positioning. Between signature bursts, dead reckoning (DR) is used.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS
• This is a continuation of an application “Music-Based Positioning Aided By Dead Reckoning”, Ser. No. 14/088,437, filed Nov. 24, 2013, which claims benefits of a provisional application “Music-Based Positioning Aided By Dead Reckoning”, Ser. No. 61/835,527, filed Jun. 14, 2013.
BACKGROUND
• 1. Technical Field of the Invention
• The present invention relates to the field of mobile electronic system, and more particularly to indoor positioning using sounds of a music piece or a human speech.
• 2. Prior Arts
• Sound has been suggested as a medium for positioning. For positioning in a large venue like a shopping mall or a supermarket, an important metric is the range of the sound. To limit the number of sound sources (e.g., loudspeakers), the sound used for positioning preferably has a long range.
• Ultrasound, although widely used for ranging, fails in this aspect. Ultrasound suffers from severe attenuation when transmitted in air. For a distance of 100 meters, the transmission loss for a 40 kHz ultrasound is ˜140 dB (FIG. 1). In fact, ultrasound can only be practically projected to a range of ˜15 meters in air. As a result, ultrasound is not suitable for positioning in a large venue.
• On the other hand, audible sounds attenuate much less in air. For example, the transmission loss for a 1 kHz audible sound is only ˜40 dB for 100 meters (FIG. 1). To be projected to a long range, audible sounds are preferably the sounds of a music piece or a human speech, whose volume can be turned up without causing annoyance to humans in the immediately vicinity. Furthermore, large venues are generally equipped with public address (PA) systems, where loudspeakers are required to provide a good acoustic coverage. It would be very attractive to leverage the existing PA systems and use the sounds of a music piece or a human speech for positioning in a large venue. Hereinafter, music is used as a primary example for indoor positioning. This concept can be easily extended to human speech.
• Although it has many advantages, music-based positioning (MP) faces a difficult challenge. A large venue is filled with background noises and multi-path reflections. Apparently, not every portion of a music piece can be used for positioning. For example, the portion of the music piece that is barely distinguishable from background noise cannot be used. To be suitable for positioning, a musical segment (i.e., a burst) should possess enough uniquely identifiable properties. A figure of merit is its correlativity, which represents the relative strength of its auto-correlation vs. its correlation with other signals. A burst with a large correlativity is relatively un-correlated with its lagged replica or background noise. In a music piece or a human speech, a burst suitable for positioning is referred to as its signature burst. The auto-correlation function of the signature burst will exhibit a distinct central peak with quickly diminishing side lobe tails.
• Litwhiler et al. (“A simple method for evaluating audible signals for acoustic measurements”, the Technology Interface, Vol. 7, No. 2, Spring 2007) taught a music-based positioning (MP) method. A music file is first sliced into bursts of 1 s long each. The correlativity of each burst is then calculated as the ratio between the peak value and the root-mean-square (rms) value of its auto-correlation function. A burst with correlativity higher than a pre-determined threshold (e.g., 20) is a signature burst, while the interval between two successive signature bursts is a non-signature interval. In the example illustrated inFIG. 2, there are 19 signature bursts (shown as cross-hatched bars) among 60 bursts evaluated. Its temporal coverage (i.e., the percentage of bursts that are suitable for positioning within a period) is ˜30%. The longest non-signature interval is 10 s, which is substantially longer than any signature burst. During a non-signature interval, no positioning can be performed using the musical sounds. Because it only provides positioning service sporadically, music-based positioning (MP) was not suitable for indoor positioning.
OBJECTS AND ADVANTAGES
• It is a principle object of the present invention to provide a continuous positioning service in a large venue using the existing infra-structure (e.g., loudspeakers in the public address system and microphones in mobile devices).
• It is a further object of the present invention to improve the temporal coverage of music-based positioning (MP).
• It is a further object of the present invention to improve the accuracy of dead reckoning (DR).
• In accordance with these and other objects of the present invention, the present invention discloses music-based positioning aided by dead reckoning (MP-DR).
SUMMARY OF THE INVENTION
• A music piece includes at least a melody and/or a rhythm. Being pleasant to hear, the music volume can be turned up to cover a larger area. However, because it only provides positioning service sporadically, music-based positioning (MP) was not suitable for indoor positioning. To provide continuous positioning service, the present invention discloses music-based positioning aided by dead reckoning (MP-DR). Generally, a music piece comprises at least first and second signature bursts separated by a non-signature interval. At each signature burst, musical sounds are used for positioning. During the non-signature interval, positioning is performed by dead reckoning. It should be apparent to those skilled in the art that this concept can be easily extended to human speech.
• MP-DR is an ideal combination of music-based positioning (MP) and dead reckoning (DR), whose strengths and weaknesses exactly complement each other. MP provides accurate positioning with an error of ˜1%. However, because most musical segments (i.e., non-signature bursts) are not suitable for positioning, MP can be performed only sporadically. The time gap between two successive MP measurements ranges from a few seconds to tens of seconds, during which a person can move a substantial distance. On the other hand, DR suffers from accumulation of errors. After calibrating its location at a reference point, although DR initially has a small error, this error grows rapidly and becomes unacceptable after a mobile user walks for more than one minute or so. For MP-DR, DR extends temporal coverage for MP, while MP provides the much needed periodic calibrations for DR.
BRIEF DESCRIPTION OF THE DRAWINGS
• FIG. 1 displays the relationship of the sound absorption by air with the sound frequency;
• FIG. 2 displays the correlativity of the first 60 bursts (each burst is 1 s long) in “Margaritaville” by Jimmy Buffet (from Litwhiler);
• FIG. 3 is a flow chart of a preferred MP-DR method;
• FIG. 4 displays error vs. time for the MP-DR;
• FIGS. 5A-5B illustrate a way to perform MP-DR between the signature bursts S_iand S_i+1:FIG. 5A illustrates the positioning of the mobile device using the DR-enhanced acoustic positioning method at S_i;FIG. 5B illustrates the progression of location errors between S_iand S_i+1;
• FIG. 6 is a functional block diagram of a mobile device in a preferred MP-DR system;
• FIG. 7 is a functional block diagram of a preferred acoustic positioning (AP) module;
• FIGS. 8A-8C illustrate three preferred signal generators.
• It should be noted that all the drawings are schematic and not drawn to scale. Relative dimensions and proportions of parts of the device structures in the figures have been shown exaggerated or reduced in size for the sake of clarity and convenience in the drawings. The same reference symbols are generally used to refer to corresponding or similar features in the different embodiments.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
• Those of ordinary skills in the art will realize that the following description of the present invention is illustrative only and is not intended to be in any way limiting. Other embodiments of the invention will readily suggest themselves to such skilled persons from an examination of the within disclosure.
• As used herein, a “location” refers to a geographic area with a lower resolution. Because it has a low accuracy, a geographic area identified by dead reckoning (DR) is generally referred to as a location. Accordingly, the process of identifying a geographic area by DR is referred to as localization. A “position” refers to a geographic area with a higher resolution. Because it has a high accuracy, a geographic area identified by acoustic positioning (AP) or music-based positioning (MP) is generally referred to as a position. Accordingly, the process of identifying a geographic area by AP or MP is referred to as positioning.
• A music piece includes at least a melody and/or a rhythm. Being pleasant to hear, the music volume can be turned up to cover a larger area. As music is used as a primary example for positioning, the terms “music-based positioning (MP)” and “acoustic positioning (AP)” are used interchangeably. They both refer to a positioning method using sounds of a music piece.
• Referring now toFIGS. 3-5B, a general description for music-based positioning aided by dead reckoning (MP-DR) is provided. Generally, a music piece (or, a human speech) comprises at least first and second signature bursts (i.e., a highly unique short musical segment suitable for positioning) separated by a non-signature interval. At each signature burst, musical sounds are used for positioning. During a non-signature interval, positioning is performed by dead reckoning. More details of MP-DR are disclosed below.
• At a first signature burst S_i, amobile device100 is positioned using the signature burst S_i(step200 ofFIG. 3). This is performed using a DR-enhanced acoustic positioning method. It includes the following steps. At the starting point TS_iof the signature burst S_i, at least two sound-transmitting devices (e.g., speakers)10a,10bsimultaneously produce a signature burst S_i(FIG. 5A). The time-of-flight (TOF) between themobile device100 and each speaker is measured and converted to distances r_a, r_bby multiplying the speed of sound. The intersections X, Y ofcircles20a(with its center at10aand a radius of r_a) and20b(with its center at10band a radius of r_b) are two possible positions of themobile device100. This positioning is further narrowed down to30a* (i.e., intersection X) by selecting the intersection (X or Y) which is located within thepossible DR locations40a.Here,possible DR locations40ainclude all locations predicted by dead reckoning. In this example, it is represented by a circle whosecenter30ais the estimated DR location and whose radius is the DR error. Besides TOF, time-difference-of-flight (TDOF) may also be used for positioning.
• Although dead reckoning could have a large error (meters to tens of meters), this error is not passed on to theposition30a* because dead reckoning is only used to select one out of two intersections X, Y. The accuracy of theposition30a* is primarily determined by the accuracy of the radii r_a, r_b. Because acoustic ranging has a high accuracy (˜1%), theposition30a* can achieve a high accuracy, typically around tens of centimeters (FIG. 4).
• During a non-signature interval NS_i(e.g., including all non-signature bursts between two successive signature bursts S_iand S_i+1ofFIG. 4), themobile device100 is localized by dead reckoning (step300 ofFIG. 3). Dead reckoning (DR) is a process of calculating one's current location by using a previously determined location, and advancing that location based upon known or estimated speeds over elapsed time. Dead reckoning uses the accelerometer and the compass of the mobile device to track the mobile user. Based on the accelerometer reading, it is possible to tell whether the user has taken a step, and therefrom estimate the displacement. Based on the compass reading, the direction of each step can be tracked.
• Because of the noisy sensors, dead reckoning suffers from accumulation of errors, which can grow cubically with the total number of steps walked from the last reference point, where its location is calibrated. This is further illustrated inFIG. 5B. Atposition30a*, the location of themobile device100 is calibrated by acoustic positioning and has a small error. As the mobile user walks along thepath50, each step (e.g.,30b,30c,30d) increases the location error, as indicated by the increased size of these circles.
• At a second signature burst S₊₁, themobile device100 is again positioned using the signature burst S_i+1(step400 ofFIG. 3). Similar toFIG. 5A, themobile device100 has two possible positions, i.e., two intersections ofcircles20a′,20b′ (FIG. 5B). Becausepossible DR locations40ehave already been predicted from dead reckoning, themobile device100 can be easily positioned. Overall, the location error of themobile device100 exhibits a saw-tooth behavior: at the first signature burst S_i, the error is small; during the non-signature interval NS_i, the error grows and then sharply drops at the next signature burst S_i+1(FIG. 4).
• Referring now toFIG. 6, amobile device100 used in a preferred MP-DR system is illustrated. Themobile device100 can be a variety of different types of devices, with different devices being the same or different types of devices. For example, device can be a cellular or other wireless phone, a laptop or netbook computer, a tablet or notepad computer, a mobile station, an entertainment appliance, a game console, an automotive computer, and so forth. Thus, device may range from a full resource device with substantial memory and processor resources (e.g., personal computers, game consoles) to a low-resource device with limited memory and/or processing resources (e.g., traditional set-top boxes, hand-held game consoles).
• Themobile device100 comprises an acoustic positioning (AP)module60, a dead-reckoning (DR)module70, aprocessor80 and amemory90. In some embodiments, themobile device100 may include many more components than those shown inFIG. 6. However, it is not necessary that all of these generally conventional components be shown in order to disclose an illustrative embodiment.
• TheAP module60 is primarily a music-based positioning (MP) module. It measures the device position using sounds of a music piece or a human speech. The measurement principles range from signal strength measurement to time-of-flight (TOF) measurement. The TOF measurement further includes pulsed time measurement, continuous-wave measurement (e.g., pattern-matching or phase-shift measurement) and others. More details of theAP module60 are disclosed inFIGS. 7-8C.
• TheDR module70 receives sensor data and executes DR algorithm to determine the location of the mobile device based on vector analysis of changes in the sensor data. It comprises a plurality of inertial sensors that detect movement (e.g., rotation, motion, velocity, etc.), altitude, and/or direction. These inertial sensors can include an accelerometer, a compass, a gyroscope, and so forth. They collect data regarding the detected movement, position, and/or direction of the device.
• Theprocessor80 accepts digital data as input, processes it according to instructions stored in thememory90 and provides results as output. Thememory90 is adapted to store software. According to the teachings of the present invention, software is provided which includes a set of executable instructions, programs, and or program modules adapted to control theAP module60 and theDR module70. In some preferred embodiments, thememory90 also stores the content data of the music sounds and/or the human speech. It can also store the signature-burst meta-data associated with the music sounds (or, the human speech).
• Referring now toFIG. 7, a functional block diagram of apreferred MP module60 is disclosed. The preferredMP module60 comprises asound receiver62, asignal generator64 and acorrelator66. Thesound receiver62 is generally a microphone. It receives theacoustic signals61 from the speakers and converts them into electrical signals, i.e., the received audio signals (which lag behind the transmitted audio signals). Thesignal generator64 produces a replica of the transmitted audio signals. Thecorrelator66 has twoinputs65aand65b:thefirst input65aincludes the received audio signals, while the second input includes the replica of the transmitted audio signals. In order to provide real-time acoustic (or, music-based) positioning, a signature-burst meta-data94 is transmitted to themobile device100 and stored in itsmemory90. This signature-burst meta-data94 includes the timing information (TS_i, TS_i+1. . . ) of all signature bursts (S_i, S_i+1. . . ) of the audio contents to be played. At the expected occurrence of a signature burst (e.g., S_i), thecorrelator66 is turned on. By matching the pattern of the transmitted audio signals in the received audio signals, the delay of the received audio signals with respect to the transmitted signals can be measured. This delay is the time-of-flight (TOF) between the speaker and the microphone.
• Referring now toFIGS. 8A-8C, severalpreferred signal generators64 are disclosed. InFIG. 8A, all remote speakers simultaneously driven by an audio player and produce sounds from radio broadcast (e.g., AM/FM) signals. Thepreferred signal generator64 is aradio receiver64a.It receives the same radio broadcast signals as the audio player. Through mixing, filtering and amplifying, it converts the radio signals into base-band signals65b.Because the radio signals have a much faster speed (i.e., the speed of light) than the acoustic signals (i.e., the speed of sound), the base-band signals65bis considered as a replica of the transmitted audio signals. In this preferred embodiment, the signature-burst meta-data94 is preferably transmitted by radio broadcast signals, too.
• InFIG. 8B, all remote speakers are simultaneously driven by an audio player. The audio player plays a list of audio contents pre-defined by an operator. Thesignal generator64 comprises anRF receiver68aand anaudio synthesizer64b.At a pre-determined location (e.g., at the entrance) of the venue, the play-list (including the playback schedule) and the associated digital audio files92 are downloaded to the mobile device via RF signals67 (e.g., via WiFi and/or cellular signals) through theRF receiver68a.This does not require full WiFi coverage at the venue. The digital audio files92 further include the signature-burst meta-data94 and thecontent data96. Theaudio synthesizer64bconverts thecontent data96 into a replica of the transmitted signals65b.
• FIG. 8C is similar toFIG. 8B, except that thememory90 of themobile device100 already stores the digital audio files on the play-list. In this case, only the play-list (including playback schedule) are downloaded to themobile device100 via the RF signals67. This requires less download time and incurs less download cost.
• While illustrative embodiments have been shown and described, it would be apparent to those skilled in the art that may more modifications than that have been mentioned above are possible without departing from the inventive concepts set forth therein. Although music is used as a primary example to illustrate AP-MD, this concept can be easily extended to human speech. The invention, therefore, is not to be limited except in the spirit of the appended claims.

Claims (20)

What is claimed is:

1. An apparatus for continuously determining a geographic area thereof using sounds of a music piece comprising first and second signature bursts separated by a non-signature interval, comprising:

acoustic-positioning (AP) module for determining a position of said apparatus at said first and second signature bursts;

a dead-reckoning (DR) module for determining a location of said apparatus by using said position as an initial location for dead reckoning during said non-signature interval.

2. The mobile device according toclaim 1, wherein said first or second signature burst has a correlativity higher than a pre-determined value.

3. The mobile device according toclaim 2, wherein the correlativity of a signature burst is the ratio of the peak value and the root-mean-square (rms) value of the auto-correlation function of the signature burst.

4. The mobile device according toclaim 1, further comprising a memory for storing the content data for said music piece.

5. The mobile device according toclaim 4, wherein said memory stores a signature-burst meta-data associated with said music piece.

6. The mobile device according toclaim 4, wherein said AP module further comprises a correlating element for calculating a correlation function between the received audio signals and the transmitted audio signals.

7. The mobile device according toclaim 6, wherein said AP module measures the time-of-flight of said received audio signals with respect to said transmitted audio signals.

8. The mobile device according toclaim 1, wherein said sounds of said music piece are simultaneously generated by at least two sound-transmitting devices.

9. The mobile device according toclaim 8, wherein said AP module measures the time-difference-of-flight of said received audio signals.

10. The mobile device according toclaim 1, wherein said DR module comprises at least an accelerometer, a compass and/or a gyroscope.

11. An apparatus for continuously determining a geographic area thereof using sounds of a human speech comprising first and second signature bursts separated by a non-signature interval, comprising:

acoustic-positioning (AP) module for determining a position of said apparatus at said first and second signature bursts;

a dead-reckoning (DR) module for determining a location of said apparatus by using said position as an initial location for dead reckoning during said non-signature interval.

12. The mobile device according toclaim 11, wherein said first or second signature burst has a correlativity higher than a pre-determined value.

13. The mobile device according toclaim 12, wherein the correlativity of a signature burst is the ratio of the peak value and the root-mean-square (rms) value of the auto-correlation function of the signature burst.

14. The mobile device according toclaim 11, further comprising a memory for storing the content data for said human speech.

15. The mobile device according toclaim 14, wherein said memory stores a signature-burst meta-data associated with said human speech.

16. The mobile device according toclaim 14, wherein said AP module further comprises a correlating element for calculating a correlation function between the received audio signals and the transmitted audio signals.

17. The mobile device according toclaim 16, wherein said AP module measures the time-of-flight of said received audio signals with respect to said transmitted audio signals.

18. The mobile device according toclaim 11, wherein said sounds of said music piece are simultaneously generated by at least two sound-transmitting devices.

19. The mobile device according toclaim 18, wherein said AP module measures the time-difference-of-flight of said received audio signals.

20. The mobile device according toclaim 11, wherein said DR module comprises at least an accelerometer, a compass and/or a gyroscope.

Images