WO2009052490A1

WO2009052490A1 - Method and apparatus for soothing a baby

Info

Publication number: WO2009052490A1
Application number: PCT/US2008/080493
Authority: WO
Inventors: John B. Carnett; Meghanad D. Wagh
Original assignee: Carnett John B; Wagh Meghanad D
Priority date: 2007-10-18
Filing date: 2008-10-20
Publication date: 2009-04-23

Abstract

A method and apparatus for soothing a baby capture a set of sound samples from a surrounding area. The captured samples are passed through a filter bank and converted to a digital sound format and a transform is performed on the samples to compute a spectrum of the sample. Spectrums of consecutive samples are compared to identify a sound made by the baby. At least one pre-determined response sound is analyzed and a spectrum calculated and stored. The spectrum from a detected target sound is compared to available response sound spectrums and the response with closest matching spectrum is selected for playback. The selected response is played back at a volume substantially equal to the target sound. The area is periodically monitored for target sounds, or changes in existing target sounds and matching responses are selected and played back.

Description

[0001] METHOD AND APPARATUS FOR SOOTHING A BABY

[0002] FIELD OF INVENTION

[0003] The claimed invention relates generally to the field of childcare.

More specifically, the claimed invention relates to unattended soothing of babies.

[0004] BACKGROUND

[0005] A common problem for the parents of a newborn baby or young infant is comforting the baby when it cries. While all babies cry, the frequency or intensity at which an individual child cries may create an environment that is stressful for both the parent and the child. In extreme cases, the stress of hearing a crying infant can lead to detrimental psychological and physiological effects such as post-partum depression and situations where an exasperated parent may feel that the last and only alternative is to strike or shake a baby. [0006] A baby cries for many reasons. Being unable to communicate verbally, crying is the only means by which a baby can communicate its needs. Among the reasons a baby cries are, fatigue, hunger, discomfort, pain, illness, frustration, boredom, loneliness, worry and fear. Generally, when the needs of the baby are met, the crying impulse will wane and the baby will settle down and stop crying. It is instinctual for a parent to react to the cry of a baby and to attempt to soothe the baby by providing for the baby's needs. When the baby receives a response to his or her need, the baby will generally stop crying, thereby indicating that his/her needs have been satisfied. Over the years, parents have devised many ways of comforting crying babies, and many times, these methods are approved by pediatricians and become part of the collective wisdom passed on to new generations of parents.

[0007] Most calming techniques fall into one of four categories, physical- close contact, touching or sucking, rhythmic motion, visual stimulation or distractions, and soothing sounds. Sounds have long been recognized as an effective way to calm a baby as evidenced by the passing down of lullabies through the ages. Additionally, during the time a baby spends developing in the womb, the baby is constantly exposed to sounds such as blood rushing through vessels or the sound of the mother's heart. After birth, sounds that simulate those intrauterine sounds will frequently act to calm a baby by providing a sense of familiarity and safety. Sounds of a mother's voice or music the mother enjoyed listening to during pregnancy are also perceptible to the baby in-utero and can serve as a soothing impulse when heard after birth.

[0008] In view of the fact that the familiar sounds of the womb comfort babies, the most effective soothing sounds are rhythmic, monotonous and humming in quality and have gently rising crescendos and decrescendos as experienced during fetal development. Sounds with such qualities are frequently referred to as "white noise" as their tonal qualities are random without distinct features that distinguish them from the surroundings. Examples include, static between radio stations, running water, vacuum cleaners, sounds from other electric motors, like washing machines or bathroom fans, and the engine and road noise experienced while riding in a car. All of these sounds have been used to pacify crying babies with varying degrees of success.

[0009] Parents are not always in a position to respond to a crying baby immediately upon hearing the baby's cry. The parent may be sleeping, taking a shower, or attending to other personal business requiring their attention. Because not all cries are of an urgent nature, the cries can be ignored, but doing so may upset the child more, causing the crying to become more intense and of longer duration. This will cause more stress for both the parent and the child. Existing devices for soothing babies are not interactive, responding to the level or intensity of sound the baby is making. It would be beneficial to have a way to soothe a baby by providing interactive feedback through soothing sounds that would get the baby's attention and make the baby feel its cries are being responded to.

[0010] SUMMARY

[0011] A method and apparatus for soothing a crying baby are disclosed. A microphone is used to capture two types of sounds: from the surroundings or to record response sounds from a caregiver. The caregiver can record and store a number of different sounds with different tones and intensities may be recorded and stored. The process starts with a set of sound samples that are captured and stored. The samples are then analyzed to determine the spectrum of the sounds captured. Repetitive ambient noise is identified by a static spectrum, while target sounds like those of a crying baby are identified by perceived changes in the spectrum. In this analysis, the machine is able to differentiate between a baby's cries and a siren, for example. Based on an identified sound from a baby and its associated spectrum, a response sound is selected from at least one stored response sound. That response sound has a spectrum most closely matching the target sound identified as the baby.

[0012] The volume level of the baby's sound is matched by outputting the selected response sound from at least one speaker coupled with an amplifier. The amplifier is configured to control the volume of the response sound. The originally recorded response sound may be shortened or portions of the sound sample may be repeated to create a response sound that closely matches the length, intensity and volume of the sound being made by the baby. [0013] When a sound sample is taken and analysis detects the sound of a baby, a response selected as described above is played and emitted from at least one speaker. When the playback of the response sound is complete, another set of sound samples is captured by the microphone and analysis is performed again. If the sounds of a baby remain detectable in the sample, then a response sound is selected to match the sequence spectrum of the detected baby sound. If the sound spectrum of the baby has changed, the selected response sound is adapted to most closely match the characteristics of the newly detected sound. In this manner, as the intensity or length of the baby's crying changes, the response sound that is emitted changes in a similar manner. The cycle of sampling, analyzing, matching, playing and re-sampling repeats until the sound of a baby is no longer detected in the captured samples and an idle mode is entered to periodically take samples until sound from the baby is once again detected. [0014] BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The described embodiments can best be understood by reference to the drawings wherein:

[0016] Figure 1 is a front perspective view of one possible embodiment of the claimed invention.

[0017] Figure 2 is a rear perspective view Figure 1.

[0018] Figure 3 is a flow diagram for soothing a baby according to an embodiment of the claimed invention.

[0019] Figure 4 is a further flow diagram for soothing a baby according to an embodiment of the claimed invention.

[0020] Figure 5 is a block diagram depicting an apparatus for soothing a baby according to an embodiment of the claimed invention.

[0021] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [0022] Figure 1 is perspective view of the front of a baby soothing device

100 in accordance with the present invention. Ambient sounds are detected by sensor 102. Sounds picked up by sensor 102 include sounds coming from the vicinity of the device 100. Among those sounds are sounds generated by the baby, identified as target sounds, to which the device 100 may play a responsive sound.

[0023] In order to establish the target sound, a series of sound samples are taken and analyzed for changes in the sound spectrum based on multiple samples. When a desired sample is captured, it is analyzed by calculating a transform on the sample to compute a spectrum representing that sample. Consecutive sound sample spectrums are compared in order to identify a target sound. The device 100 then analyzes the target sound and determines the volume or intensity level of the target sound based on its computed spectrum. When the target sound is identified and its spectrum computed, an output signal is produced instructing the device 100 to playback a selected response sound. The response sound is selected from a number of predetermined response sounds.

Based on the analysis of the input samples, the output response having the closest spectrum match to the target sound is selected. Output source selection may be chosen by a user via controls 106 located on the device 100. For example, input sound sources may come from tapes, radio stations, compact discs (CD's), digital sound storage devices, such as memory cards or flash drives, or external media players connected to the device 100 by an appropriate interface. [0024] Once an output response sound has been selected, the device 100 plays the selected responsive sound through one or more speakers 104. The volume or intensity of the responsive sound is at a level substantially equal that of the target sound captured by the sensor 102.

[0025] Figure 2 depicts the rear of the device 200 where an additional sound detection sensor 206 is located. Those skilled in the art will recognize that any number of sound detection sensors may be used and placed advantageously, even remotely, to best detect target sounds. An external memory card slot 204 is provided to allow digitally stored sounds as a library of response sounds for the device. The memory card slot 204 comprises an interface with the control circuitry of the device 100. Such memory devices may store recorded music, the voice of the baby's parent, another person with whom the baby is familiar, or white noise such as static, ocean sounds, or even the sound of a vacuum cleaner that may be directed to the speakers 104 for playback by the control circuitry. These examples of recorded sounds are provided by way of example and are in no way intended to limit the scope of the invention, as any sound capable of reproduction could be used as the responding sound. While a battery compartment 202 is provided for portable power, alternating current may also be used.

[0026] Figure 3 is a block diagram illustrating an operational overview for the apparatus. The operation begins by detecting a target sound 301 such as the sound of a baby crying. A target sound is identified when the samples collected in a predetermined set indicate that the sound spectrum of the samples in the set changed by a predetermined amount from one sample to the next. When that change in the sound spectrum is detected, the target sound is identified and a response sound of similar characteristics, such as pitch and volume, is selected and played back to the baby. To do this, a response sound is selected from the available response sounds by matching the detected target sound to the selected response 303 by an analysis of the sound spectrum of the two sounds and the response sound with the closest matching sound spectrum is selected. Once selected, the response sound is configured to match the target sound 305. For example, the volume of the selected response sound may be modulated and amplified before playback in order to have a playback of the selected response at a volume level equal to the detected target sound. After the response sound is matched and configured to closely match the target sound, the selected response sound is played back based on its configured characteristics 307. The playback continues for a selected time period. At the end of the playback period, a monitoring phase 309 is entered and another set of sound samples is captured. The captured set of sound samples is then analyzed to determine if there is a need or further soothing. The apparatus may cycle through multiple samples, an automated timing cycle or user intervention, such as by a remote control. [0027] With reference to Figure 4, an area is monitored and sound samples are collected by a microphone (block 401) as the detector 200. The microphone collects a plurality of samples from the surrounding area and the samples are converted to a digital format by an analog to digital converter, stored in a memory, and analyzed (block 403) and the spectrum of each sample is determined by techniques known in the art. For example, a Hagamard transform may be used to compute a sequency spectrum for a sample. Spectrums of the collected samples are compared. When spectrums remain constant over a number of consecutive samples, they are analyzed and identified as ambient noises, such as a fan or other background noise while a change in spectrum may be baby noises, such as crying. Characteristics of the spectrum that are indicative of a baby crying as opposed to other noises that may produce a change in spectrum over a certain threshold may be analyzed to distinguish that a detected sound is that of a baby and not some other sound that does not require a response. When a change in spectrum is determined to be a baby crying, a response procedure, explained in greater detail below, is initiated. If no significant change in spectrums is present, indicating there is no response is needed, then monitoring phase (block 401) is re-entered as shown by decision block 405.

[0028] The response procedure begins by comparing the spectrum of the identified target sound (i.e. the sound being made by the baby) and comparing it to responses stored in a memory. The response sounds may be recorded by the user, may be pre-recorded sounds provided with a baby soothing unit or may be audio signals from an external source such as a radio, tape player, CD player or MP3 player. The spectrums of the stored response sounds are compared to the spectrum of the target sound and the response sound with the closest matching spectrum is selected (block 407). Other characteristics of the target sound, such as volume or duration are determined by the sampling analysis (block 403) and the selected response sound is configured to correlate it to the target sound (block 409). If a stored response sound is longer than the target sound, it may be truncated. Similarly, if the stored response sound is of a shorter duration than the target sound, the response sound, or a portion of the response sound may be repeated and concatenated to the response sound to make it of a duration that corresponds to the target sound.

[0029] The configured response sound is then played back (block 411) to the crying baby who receives adaptive feedback to its crying. After the selected and formatted response sound is played back, the monitoring phase (401) is re- entered and sampling of the surrounding sounds is again initiated as described above.

[0030] With reference to Figure 5 a microphone 501 provides an input for sounds to the device 500. While one microphone 501 is shown for simplicity, any number of sound detection sensors may be used. They may be advantageously positioned to best detect target sounds. The target sound may be ambient sounds which the device 500 is monitoring, such as a crying baby. Alternatively, the microphone 501 may be utilized to capture user generated response sounds that will be stored and played back in response to a detected target sound. The microphone is coupled to a voice recording/playback module 517 for recording responsive sounds created by the user. Additionally other responsive sounds may be selected from sources other than the voice recording/playback module 517. An audio input interface may be used to receive audio signals for use as responsive sounds. The audio input interface may be a memory card slot configured to receive an external memory card that contains digital music files, or an analog input interface for receiving an audio signal from an external device such as a radio, a CD player, or a cassette tape player.

[0031] When the microphone 501 is utilized to monitor ambient sounds to identify target sounds, the sounds sensed by the microphone 501 are directed to the filter bank 503, which takes an input sample and filters out higher frequencies using low pass filters to remove brief oscillations in the captured input signal and provide a smoother input signal more representative of a target sound, such as a baby crying. Filter bank 503 provides a long term input sound without temporary interference in the signal to the sample and hold circuit 505. [0032] A sample and hold circuit 505 receives periodic samples from the filter bank 503 and stores them while sample processing is performed. The sample and hold circuit 505 holds a sample until the previous sample is converted from an analog to a digital sound signal by the associated analog to digital (AJO) converter. The sample and hold circuit 505 and the A/D converter 507 are controlled by microcontroller which preferably includes a processor 511 and a memory 513. Microcontroller 509 provides control functions to the A/D converter 507, the voice recording/playback module 517 and amplifier 529. The microcontroller 509 also preferably comprises a transform entity 509 which computes a transform on the signal output from the A/D converter 507. The transform entity 509, which may be implemented in software, computes a spectrum for the received digital signal. For example, a fast Hadamard transform may be implemented to calculate a sequency spectrum; however, any transform may be performed to compute a spectrum for comparison and selection of a response. The computed spectrum is then used to select a pre-recorded response, the process of which will be explained in greater detail hereinafter. [0033] The spectrums of the previously processed samples are compared to decide if their spectrums have changed substantially. If the input spectrums of two consecutive samples change substantially, it is an indication that the target sound such as a baby crying has been detected. Further analysis of the characteristics relating to the change in spectrum may be performed to more closely approximate the changes in a baby's voice versus other substantial changes, for example sirens or thunder. In addition to being used for detection of a target sound, the spectrum is also used to select a response sound for playback. [0034] The microcontroller 509 is coupled to and controls the voice recording/playback module 517. The voice recording/playback module 517 is coupled to the microphone 501, the microcontroller 509, and the speaker 531 through amplifier 529. In a recording mode, user created response sounds are sensed by the microphone 501 when the user speaks or makes some other sound directed at the microphone and the created sounds are directed to the voice recording/playback module 517. The voice recording/playback module 517 comprises a memory 519, an A/D converter 521, and digital to analog (D/A) converter 523 and a processor 525 for storing recorded samples connected by a bus 527. The processor 525 is configured to interpret commands from the microcontroller 509 and may control the storage of a recorded response at a specified address, or may initiate the playback of a selected response sound from a specified address in memory 519. The recorded samples may also be stored in any suitable memory available to the microcontroller 509. User controls enable a user to record a set of response sounds. The response sounds may be reassuring comments, cooing or any soothing sound the user deems appropriate for soothing the baby. The response sounds may be of different lengths or volumes. The microcontroller 509 stores the recorded response samples into memory 519 in sequential memory addresses. When completing the storing of a recorded response sound, the microcontroller 509 remembers the next available memory address. When the next response sound is recorded, the microcontroller 509 stores the recorded response in memory 519 at the address where the previous recorded response had ended. The microcontroller 509 maintains a table of the memory addresses of each recorded response. In this manner, when a playback response is selected, the microcontroller 509 may use the memory address of the selected response and process the address to retrieve and playback the selected response from memory 519. The microcontroller 509 also maintains the spectrum information for each stored response sound stored in memory 519. [0035] As was discussed earlier, after the transform entity 515 has determined the spectrum of a detected target sound, the spectrum is used to select a response recording from memory 519 for playback. The recorded response sound whose spectrum most closely matches the spectrum of the detected target sound is selected for playback as a response. The spectrum of the target sound also indicates the volume at which the playback should occur. The playback volume is controlled by an amplifier 529 coupled to the microcontroller 509 which controls the volume at which the selected recorded response sound is played back through the speaker 531 based on the sound spectrum of the detected target sound.

[0036] When the response sound is selected, then the memory address of the selected response is sent to the processor 525 which directs the voice recorder/playback module 517 to send the selected response from memory 519 from the memory address provided by the processor 509 and to direct the selected response to a digital to analog (D/ 'A) converter 523 in the voice recorder/playback module 517. The D/A converter output is directed to the amplifier 529, whose output is controlled by the microcontroller 509 based on the spectrum of the target sound as determined by the transport entity 515. The amplifier 529 then plays back the selected response through the speaker 531. [0037] When a target sound has been identified by the analysis of the spectrum of a captured sample by the transport entity 515, the selected response sound is played back for a pre- determined period of time. For example, the playback may last ten seconds. After the pre-determined time period has ended, the apparatus 500 takes another input sample set through the microphone 501 to determine if the target sound still exists, or whether the spectrum of the target sound has changed. If a target sound still exists, and the spectrum is similar to the previous target sound, the same selected response is played back again. If the spectrum of the detected target sound has changed, then a new response sound is selected from the response sounds stored in memory 519. The response sound with the spectrum most closely matching the target sound is selected for playback and the newly selected response sound is played back through the speaker 531 as described above.

[0038] The embodiments described herein are provided by way of example and in no way restrict the scope of the claimed invention.

Claims

CLAIMS What is claimed is:

1. A method of soothing a baby comprising the steps of: monitoring at least one input sound sample; analyzing the at least one input sound sample; determining sound characteristics of the at least one input sound sample based on the analysis; matching a pre-recorded response sound to the at least one input sound sample based on the determined sound characteristics of the at least one input sound; playing back the matched pre-recorded response sound.

2. The method of claim 1 wherein matching a pre-recorded response sound to the at least one input sound sample further comprises: computing a spectrum for the at least one input sound sample and computing a spectrum for at least one pre-recorded response sound; comparing the spectrum of the at least one input sound to the spectrum for the at least one pre-recorded response sound; selecting one of the at least one pre-recorded response sounds that has the closest matching spectrum to the spectrum of the at least one input sound sample.

3. The method of claim 1, wherein analyzing the at least one sound sample further comprises: calculating a spectrum on a plurality of consecutive input sound samples; comparing the spectrums of two consecutive input sound samples; identifying a target sound in one of the two consecutive input sound samples, based on variation of spectrums of the two consecutive input sound samples exceeding a pre-determined threshold and analyzing the spectrum to determine that the target sound is a sound that requires a response.

4. The method of claim 1, wherein monitoring at least one input sound sample further comprises: capturing a plurality of sound samples; converting the plurality of sound samples from analog sound signals to digital sound signals; calculating a transform based on the converted digital sound signals.

5. The method of claim 4, wherein the capturing of a plurality of sound samples is performed once per second.

6. The method of claim 4, wherein 64 sound samples are captured at a rate of 20 Kilosamples per second.

7. The method of claim 1, wherein the analysis of the at least one input sound sample further comprises: computing a Hadamard transform on the at least one input sound sample to calculate a sequency spectrum.

8. The method of claim 1, wherein the analysis of the at least one input sound sample further comprises: computing a Fourier transform on the at least one input sound sample to calculate a frequency spectrum.

9. The method of claim 1 further comprising: recording at least one response sound; converting the at least one response sound to a digital sound signal; storing the converted at least one digital sound signal.

10. The method of claim 9, further comprising: performing a transform on the stored at least one digital sound signal; storing the spectrum of the at least one digital sound signal based on the transform.

11. An apparatus for soothing a baby comprising: a microphone configured to capture input sounds and output a sound signal; a filter bank coupled to the microphone configured to provide low pass filtering to a captured input sound; a sample and hold circuit coupled to the filter bank and configured to take periodic samples of an analog input sound signal and hold the periodic sample until the periodic sample is processed by an analog to digital converter; a microcontroller comprising a processor, a memory and a transform entity, the microprocessor coupled to the analog to digital converter; a voice recording/playback module coupled to the microphone and the microcontroller, the voice recording module comprising a analog to digital converter , a digital to analog converter, a second processor and a second memory; an amplifier coupled to the voice recording/playback module and the microcontroller configured to control the volume level of a selected sound output; a speaker coupled to the amplifier for outputting the selected sound output.

12. The apparatus of claim 11, wherein the processor is further configured to: compute a transform of an input sound sample and at least one recorded response sound sample; store the computed results of the transform performed on the at least one recorded response sound sample; compare the computed results of the input sound sample to each of the stored results based on the at least one stored response sound sample; select one stored response sample that most closely matches the computed result of the transform of the input.

13. The apparatus of claim 11, wherein the transform entity is configured to perform a Hadamard transform on the digital input sound signal and calculate a sequency spectrum.

14. The apparatus of claim 11, wherein the transform entity is configured to perform a Fourier transform on the digital input sound signal and calculate a frequency spectrum.

15. The apparatus of claim 11, wherein the second memory is configured to store at least one pre- determined response sound for playback in response to a target sound.

16. The apparatus of claim 15, wherein the at least one pre-determined response sound is an audio recording of a human voice.

17. The apparatus of claim 16, wherein the human voice is the voice of a parent of a baby, wherein the crying of the baby is the target sound.

18. The apparatus of claim 15, wherein the at least one pre-determined response sound is an external audio signal.

19. The apparatus of claim 17, wherein the external audio signal is a radio broadcast.

20. The apparatus of claim 17, wherein the external audio signal is the output of a cassette tape player.

21. The apparatus of claim 17, wherein the external audio signal is the output of a digital music player, wherein the digital music player is playing a digital music file stored on an external memory device.

22. A baby soothing apparatus comprising: a detector that detects sounds from a baby; a filter and format unit that receives a detected sound and formats it to a digital signal for subsequent processing; a microprocessor that receives the previously formatted signal and outputs a transformed signal; a compare and playback module that analyzes the transformed signal to determine if soothing is required, and, upon detecting that soothing is required, outputs the required soothing signal; and, an amplifier and speaker for outputting the selected soothing sound at a audible level that is responsive to the baby.

23. The apparatus of claim 22 wherein the filter and format unit further comprises a low pass filter, a sample and hold circuit and an analog to digital converter.

24. The apparatus of claim 23 wherein the microprocessor further comprises a processor, a memory and a transform entity.

25. The apparatus of claim 24 wherein the filter and format unit further comprises a low pass filter, a sample and hold circuit and an analog to digital converter.

26. The apparatus of claim 22 wherein the compare and playback module further comprises a analog to digital converter, a digital to analog converter, a processor and a memory of pre-recorded sounds.