Music-Based Positioning Aided By Dead Reckoning

Abstract

Music-based positioning (MP) provides positioning service only sporadically and therefore is not suitable for indoor positioning. The present invention discloses system and method for music-based positioning aided by dead reckoning (MP-DR). At each signature burst (i.e. a highly unique short musical segment suitable for positioning), musical sounds are used for positioning. Between signature bursts, positioning is performed by dead reckoning (DR). MP-DR is an ideal combination of MP and DR: DR extends temporal coverage for MP, while MP provides the much needed periodic calibrations for DR.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims priority of a provisional application entitled “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 human consumable sounds such as music.

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 music or other human consumable sounds, 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 human consumable sounds produced thereby for positioning in a large venue. Hereinafter, music is used as a primarily example of human consumable sounds.

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 the music can be used for positioning. For example, the portion of the music 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 an audio content, 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. Any burst with correlativity higher than a pre-determined threshold (e.g. 20) is considered as a signature burst. In the example illustrated in FIG. 2, there are 19 signature bursts (shown as cross-hatched bars) among the 60 bursts evaluated. Its temporal coverage (i.e. the percentage of bursts that are suitable for positioning within a period) is ˜30%. The longest time interval without a signature burst (i.e. non-signature interval) is 10 s. During a non-signature interval, no positioning can be performed using musical sounds. Because it can be used for positioning only sporadically, music was considered un-suitable for indoor positioning in the past.

OBJECTS AND ADVANTAGES

It is a principle object of the present invention to provide a 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 system and method for music-based positioning aided by dead reckoning (MP-DR).

SUMMARY OF THE INVENTION

The present invention discloses system and method for music-based positioning aided by dead reckoning (MP-DR). At each signature burst (i.e. a highly unique short musical segment suitable for positioning), musical sounds are used for positioning. During a non-signature interval (i.e. between successive signature bursts), positioning is performed by dead reckoning.

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 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. 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 music-based position aided by dead reckoning (MP-DA);

FIGS. 5A-5B illustrate a way to perform MP-DA between the signature bursts S_i and 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_i and 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.

In this specification, “human consumable sounds” are sounds whose volume can be turned up to project the sounds to a long range without causing annoyance to humans in the immediately vicinity. Human consumable sounds are generally produced from an audio content, including music, human speech or others. As music is used as a primary example of human consumable sounds, the terms “music-based positioning (MP)” and “acoustic positioning (AP)” are used interchangeably. They both refer to a positioning method using human consumable sounds.

Referring now to FIGS. 3-5B, a general description for music-based positioning aided by dead reckoning (MP-DR) is provided. At each signature burst (i.e. a highly unique short musical segment suitable for positioning), musical sounds are used for positioning. During a non-signature interval (i.e. between successive signature bursts), positioning is performed by dead reckoning. More details of MP-DR are disclosed below.

At a first signature burst S_i, a mobile device 100 is positioned using the signature burst S_i (step 200 of FIG. 3). This is performed using a DR-enhanced acoustic positioning method. It includes the following steps. At the starting point TS_i of the signature burst S_i, at least two sound-transmitting devices (e.g. speakers) 10a, 10b simultaneously produce a signature burst S_i (FIG. 5A). The time-of-flight (TOF) between the mobile device 100 and each speaker is measured and converted to distances r_a, r_b by multiplying the speed of sound. The intersections X, Y of circles 20a (with its center at 10a and a radius of r_a) and 20b (with its center at 10b and a radius of r_b) are two possible positions of the mobile device 100. This positioning is further narrowed down to 30a* (i.e. intersection X) by selecting the intersection (X or Y) which is located within the possible DR locations 40a. Here, possible DR locations 40a include all locations predicted by dead reckoning. In this example, it is represented by a circle whose center 30a is 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 the position 30a* because dead reckoning is only used to select one out of two intersections X, Y. The accuracy of the position 30a* is primarily determined by the accuracy of the radii r_a, rb. Because acoustic ranging has a high accuracy (˜1%), the position 30a* can achieve a high accuracy, typically around tens of centimeters (FIG. 4).

During a non-signature interval NS_i (e.g. between two successive signature bursts S_i and S_i+1 of FIG. 4), the mobile device 100 is localized by dead reckoning (step 300 of FIG. 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 in FIG. 5B. At position 30a*, the location of the mobile device 100 is calibrated by acoustic positioning and has a small error. As the mobile user walks along the path 50, 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_i+1, the mobile device 100 is again positioned using the signature burst S_i+1 (step 400 of FIG. 3). Similar to FIG. 5A, the mobile device 100 has two possible positions, i.e. two intersections of circles 20a′, 20b′ (FIG. 5B). Because possible DR locations 40e have already been predicted from dead reckoning, the mobile device 100 can be easily positioned. Overall, the location error of the mobile device 100 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 to FIG. 6, a mobile device 100 used in a preferred MP-DR system is illustrated. The mobile device 100 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).

The mobile device 100 comprises an acoustic positioning (AP) module 60, a dead-reckoning (DR) module 70, a processor 80 and a memory 90. In some embodiments, the mobile device 100 may include many more components than those shown in FIG. 6. However, it is not necessary that all of these generally conventional components be shown in order to disclose an illustrative embodiment.

The AP module 60 is primarily a music-based positioning (MP) module. It measures the device position using human consumable sounds such as music. 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 the AP module 60 are disclosed in FIGS. 7-8C.

The DR module 70 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.

The processor 80 accepts digital data as input, processes it according to instructions stored in the memory 90 and provides results as output. The memory 90 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 the AP module 60 and the DR module 70. In some preferred embodiments, the memory 90 also stores audio contents and the associated signature-burst meta-data.

Referring now to FIG. 7, a functional block diagram of a preferred MP module 60 is disclosed. The preferred MP module 60 comprises a sound receiver 62, a signal generator 64 and a correlator 66. The sound receiver 62 is generally a microphone. It receives the acoustic signals 61 from the speakers and converts them into electrical signals, i.e. the received audio signals. The signal generator 64 produces a replica of the transmitted audio signals. The correlator 66 has two inputs 65a and 65b: the first input 65a includes 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-data 94 is transmitted to the mobile device 100 and stored in its memory 90. This signature-burst meta-data 94 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), the correlator 66 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 to FIGS. 8A-8C, several preferred signal generators 64 are disclosed. In FIG. 8A, all remote speakers simultaneously driven by an audio player and produce sounds from radio broadcast (e.g. AM/FM) signals. The preferred signal generator 64 is a radio receiver 64a. It receives the same radio broadcast signals as the audio player. Through mixing, filtering and amplifying, it converts the radio signals into base-band signals 65b. 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 signals 65b is considered as a replica of the transmitted audio signals. In this preferred embodiment, the signature-burst meta-data 94 is preferably transmitted by radio broadcast signals, too.

In FIG. 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. The signal generator 64 comprises an RF receiver 68a and an audio synthesizer 64b. 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 files 92 are downloaded to the mobile device via RF signals 67 (e.g. via WiFi and/or cellular signals) through the RF receiver 68a. This does not require full WiFi coverage at the venue. The digital audio files 92 further include the signature-burst meta-data 94 and the audio data 96. The audio synthesizer 64b converts the audio data 96 into a replica of the transmitted signals 65b.

FIG. 8C is similar to FIG. 8B, except that the memory 90 of the mobile device 100 already stores the digital audio files on the play-list. In this case, only the play-list (including playback schedule) are downloaded to the mobile device 100 via the RF signals 67. 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. The invention, therefore, is not to be limited except in the spirit of the appended claims.

Claims

1. A mobile device, comprising:

an acoustic positioning (AP) module for positioning said mobile device using a signature burst, wherein said signature burst includes human consumable sounds suitable for positioning;

a dead-reckoning (DR) module for localizing said mobile device during an interval without signature burst.

2. The mobile device according to claim 1, wherein said human consumable sounds are sounds of an audio content.

3. The mobile device according to claim 2, wherein said audio content is music or human speech.

4. The mobile device according to claim 2, further comprising a memory for storing said audio content.

5. The mobile device according to claim 2, further comprising a memory for storing a signature-burst meta-data associated with said audio content.

6. The mobile device according to claim 1, wherein said human consumable sounds are simultaneously generated by at least two sound-transmitting devices.

7. The mobile device according to claim 1, wherein said signature burst has a correlativity higher than a pre-determined value.

8. The mobile device according to claim 7, wherein the correlativity of a burst is the ratio of the peak value and the root-mean-square (rms) value of the auto-correlation function of said burst.

9. The mobile device according to claim 1, wherein said AP module further comprises a correlating element for calculating a correlation function between received audio signals and transmitted audio signals.

10. The mobile device according to claim 9, wherein said AP module further comprises a sound-receiving element for converting said human consumable sounds into said received audio signals.

11. The mobile device according to claim 9, wherein said AP module further comprises a radio-receiving element for converting a radio signal into a replica of said transmitted audio signals.

12. The mobile device according claim 9, wherein said AP module further comprises an audio-synthesizing element for converting an audio content into a replica of said transmitted audio signals.

13. The mobile device according claim 9, wherein said AP module measures the time-of-flight or time-difference-of-flight of said received audio signals with respect to said transmitted audio signals.

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

15. A mobile device, comprising:

a memory for storing a signature-burst meta-data associated with an audio content, wherein signature bursts include human consumable sounds suitable for positioning;

an acoustic positioning module for positioning said mobile device using a signature burst of said audio content.

16. The mobile device according to claim 15, wherein said audio content is music or human speech.

17. The mobile device according to claim 15, wherein said human consumable sounds are simultaneously generated by at least two sound-transmitting devices.

18. The mobile device according to claim 15, wherein said signature bursts have a correlativity higher than a pre-determined value.

19. The mobile device according to claim 18, wherein the correlativity of a burst is the ratio between the peak value and the root-mean-square (rms) value of the auto-correlation function of said burst.

20. The mobile device according to claim 15, further comprising a dead-reckoning module for localizing said mobile device during an interval without signature burst.