Pure tone combinations with balanced interference

Abstract

A technique is presented in which signals are constructed from multiple tone frequencies, combined in such a way as to minimise fluctuations in the combined signal amplitude. Fluctuations in acoustic pressure, and stimulation effects, are reduced. This is achieved by configuring combinations of tones such that there is a characteristically flat signal envelope in the time domain. Presented is an apparatus for estimating motion of a target from reflections of such a combined signal.

Description

[0001] Pure Tone Combinations with Balanced Interference

[0002] Field of Invention

[0003] The present disclosure relates to motion estimation, and more particularly to an apparatus for estimating motion of a target from reflection of acoustic pure tones with balanced interference patterns.

[0004] Technical Background

[0005] Acoustic pure tones can be utilized on devices such as smartphones to enable contactless motion detection functionality for a myriad of useful applications, including sleep tracking, breathing detection, gaming, musical instruments, and user interface control, as described in earlier patent by the Inventor (British patent 2609061), and the present invention is made in this context. Ultrasonic pure tones with frequencies above 18 kHz are of particular utility as they are inaudible to most adults. Motion can be inferred via analysis of reflected signals received by a microphone, which contain information about fast moving reflective surfaces such as a hand moving towards the phone, through Doppler shifts, and about slower changes in the distance of reflective surfaces such as chest movements that cause spectral shifts over time, which is useful for estimating breathing during sleep tracking.

[0006] It is possible to increase the accuracy of ultrasonic motion detection systems by combining multiple pure tones of different frequencies. Information from multiple frequency ranges around the constituent tones can be combined, such as by averaging, to mitigate the impact of factors such as external noises and interference effects related to the distance of the reflective surface with respect to the tone wavelength, to enable more accurate motion estimates.

[0007] Although multi-tone systems provide greater information than single tone systems, deploying them does create complications that require consideration. This is because when multiple tones are combined, interference patterns occur at the frequency interval between the tones (the beat frequency). Such interference patterns can cause fluctuations in acoustic pressure that can induce physiological effects such as brainwave entrainment for lower frequency beat intervals (e.g. less than 100Hz) and nerve and muscular responses at higher frequency beat intervals, such as tingling sensations in the skin. Brainwave entrainment is a phenomenon whereby brain activity synchronizes with the frequency of external stimuli, such as acoustic patterns, which can lead to effects such as heightened alertness. The Inventor has previously explored methods for deploying such effects beneficially. However, in many instances, such stimulation effects are undesirable, for example, in a context such as sleep-tracking. Such effects do not typically occur with single frequency systems, making them generally superior for applications such as sleep tracking where it is desirable for stimulation effects to be minimised. Summary of Invention

[0008] The present disclosure presents methods to construct signals that combine multiple tone frequencies in such a way as to minimise fluctuations in the combined signal amplitude, and hence acoustic pressure, and reduce stimulation effects. This is achieved by configuring combinations of tones such that there is a characteristically flat signal envelope in the time domain.

[0009] Specifically, tones are combined with relative amplitudes and phase such that there exist periodic instants in time at which a majority of the tones are in phase alignment with each other, whilst other tones are in antiphase alignment with the majority, creating an algebraic cancellation of beat effects.

[0010] In the present disclosure, ‘phase alignment’ describes two signals which have a phase offset of zero, or multiples of 2TT radians at an instant in time, which combine constructively at that instant. ‘Antiphase alignment’ describes two signals which have a phase offset of TT radians at an instant in time, and which combine destructively at that instant.

[0011] When tone combinations with such opposing interference patterns are combined, the points of constructive interference from one set of tones align with points of destructive interference from the other set of tones. This has the effect of suppressing constructive and destructive interference peaks in the combination and flattening the envelope of the combined signal. The frequency offset and tone frequencies can be optimized for compatibility with other motion detection techniques from the Inventor, notably those disclosed in British patent 2609061 and international patent application WO 2024 / 175934. A short and precisely-defined period of alignment, referred to herein as a ‘loop cycle’, can be obtained if tone frequencies and frequency offset are a function of the sampling rate, which is used to advantageous effect in combination with the Doppler motion systems described in British patent 2609061 and international patent application WO 2024 / 175934.

[0012] Algebraic solutions are presented for configuring five and seven tone combinations so that amplitude fluctuations are minimised in their combination.

[0013] The algorithms have another useful property of compressing the maximal amplitude of the tone combination compared to the sum of the constituent tone amplitudes. For example, some 7-tone combinations (referred to herein as EQUINOX combinations) have maximal amplitude of less than 45% of the sum of the amplitudes of the constituent tones, whilst exhibiting a flat and consistent signal envelope within a recurrent loop optimized for motion tracking.

[0014] The optimised multitone signals enable precise and superior measurement of both Doppler shifts and breathing compared to single tone systems, and the flat output signals have the intended effect of substantially reducing stimulation effects compared to nonoptimised combinations that exhibit greater amplitude variation.

[0015] Despite these superior properties, however, these flattened tone combinations are not able to match single frequency systems in terms of stimulation effect minimisation. This is because when their precisely balanced output signals interfere with time delayed reverberations (e.g. reflections from walls and ceilings), the delicate phase and amplitude structure becomes unbalanced, leading to increased pressure fluctuations.

[0016] Consequently, these optimised multitone configurations are best suited for short term applications that benefit from increased precision such as gaming or short term breathing rate measurement (e.g. over a 20-minute meditation session), but are less well suited to prolonged use over multiple hours for sleep tracking.

[0017] According to an aspect of the present invention, there is provided an apparatus for estimating motion of a target, comprising: a speaker, configured to output a combination of three or more acoustic pure tones having frequencies and relative amplitudes symmetrically distributed about a central pilot frequency, wherein the spacing between each tone frequency is an integer multiple of an offset frequency unit; a microphone for receiving and sampling reflections, by the target, of the three or more tones; and a motion estimator for estimating motion of the target from frequency analysis of the reflected three or more tones; wherein periodically, the tones have a phase alignment such that a majority of the tones align in phase, while all other tones have a phase offset of TT radians from the majority of the tones, and the relative amplitudes are such that all interference patterns between constituent tones in the combination are counterbalanced, with the exception of interference patterns between the two tones with greatest frequency spacing. In embodiments, the speaker is configured to output a first tone and a third tone having respective first and third frequencies which are offset respectively below and above a central second frequency of a second tone by one offset frequency unit, and the majority of the tones comprises either: the first tone and the second tone, or the second tone and the third tone.

[0018] In embodiments, the first, second and third tones have relative amplitudes in the ratio of 1 : x : 1 , for any value of x.

[0019] In embodiments, the speaker is configured to output: a first tone and a fifth tone having respective first and fifth frequencies which are offset respectively below and above a third central frequency of a third tone by two offset frequency units, and a second tone and a fourth tone having respective second and fourth frequencies which are offset respectively below and above the third central frequency of the third tone by one offset frequency unit, and the majority of the tones comprises either: the first tone, the second tone, the third tone and the fifth tone, or the first tone, the third tone, the fourth tone and the fifth tone.

[0020] In embodiments, the first, second, third, fourth and fifth tones have relative amplitudes in the ratio of 1 : x : (x2 / 2) : x : 1 for any value of x.

[0021] In embodiments, the speaker is configured to output: a first tone and a seventh tone having respective first and seventh frequencies which are offset respectively below and above a fourth central frequency of a fourth tone by three offset frequency units, and a second tone and a sixth tone having respective second and sixth frequencies which are offset respectively below and above the fourth central frequency of the fourth tone by two offset frequency units, and a third tone and a fifth tone having respective second and sixth frequencies which are offset respectively below and above the fourth central frequency of the fourth tone by one offset frequency unit, and the majority of the tones comprises either: the first tone, the second tone, the third tone, the fourth tone, and the sixth tone, or the second tone, the fourth tone, the fifth tone, the sixth tone and the seventh tone.

[0022] In embodiments, the first, second, third, fourth, fifth, sixth and seventh tones have relative amplitudes in the ratio of 1 : x : (x2 / 2) : ((z2- 2z) / 2x) : (x2 / 2) : x : 1 , with z = (x2 / 2) for any value of x.

[0023] According to a further aspect of the present invention, there is provided a method of estimating motion of a target comprising outputting three or more acoustic pure tones having frequencies and relative amplitudes symmetrically distributed about a central pilot frequency, wherein the spacing between each tone frequency is an integer multiple of an offset frequency unit; receiving and sampling reflections, by the target, of the three or more tones; estimating motion of the target from frequency analysis of the reflected three or more tones; wherein periodically, the tones have a phase alignment such that a majority of the tones align in phase, while all other tones have a phase offset of TT radians from the majority of the tones, and the relative amplitudes are such that all interference patterns between constituent tones in the combination are counterbalanced, with the exception of interference patterns between the two tones with greatest frequency spacing; and outputting the estimated motion. According to a further aspect of the present invention, there is provided a computer program which, when executed by one or more processors, causes the above method to be performed.

[0024] Brief Description of Drawings Embodiments of the present invention are disclosed by way of example only, with reference to Figure 1 , which shows an apparatus for estimating motion according to embodiments of the present invention. Detailed Description

[0025] According to an embodiment, illustrated in Figure 1 , an apparatus 10 for estimating motion of a target 20 comprises a speaker 11 , a microphone 12, and a motion estimator 13 that work together to provide contactless motion detection functionality. The apparatus 10 may be implemented on smartphones or other mobile devices to enable various applications including sleep tracking, breathing detection, gaming, musical instruments, and user interface control. The apparatus 10 comprises an output 14 for outputting motion estimated by the motion estimator 13 and each component of the apparatus 10 may be controlled by a central controller 15 or processing unit.

[0026] The speaker 11 is configured by the controller 15 to output a combination of three or more acoustic pure tones having frequencies and relative amplitudes symmetrically distributed about a central pilot frequency. The acoustic pure tones may have ultrasonic frequencies above 18 kHz to make the tones inaudible to most adults. The spacing between each tone frequency is an integer multiple of an offset frequency unit, providing a structured arrangement of frequencies that enables balanced interference patterns.

[0027] The microphone 12 receives and samples reflections, by the target 20, of the three or more tones. The microphone 12 captures acoustic signals that have been reflected from moving objects or surfaces, such as a hand moving towards the apparatus or chest movements during breathing. The reflected signals contain information about the motion of reflective surfaces through various acoustic phenomena.

[0028] The motion estimator 13 estimates motion of the target from frequency analysis of the reflected three or more tones. The motion estimator 13 processes the reflected signals to extract motion information by analyzing frequency characteristics of the received acoustic data. Motion may be inferred through analysis of Doppler shifts for fast moving reflective surfaces and through spectral shifts over time for slower changes in the distance of reflective surfaces. The estimated motion is output, via the output means 14, on a display or user interface, or is transmitted to one or more other devices for further processing, use in a further application, or presentation to the user. In embodiments, the controller 15 may perform at least some or all of the functions of the motion estimator 13.

[0029] The combination of tones is configured with specific phase alignment characteristics where periodically, the tones have a phase alignment such that a majority of the tones align in phase, while all other tones have a phase offset of TT radians from the majority of the tones. The relative amplitudes are such that all interference patterns between constituent tones in the combination are counterbalanced, with the exception of interference patterns between the two tones with greatest frequency spacing. This configuration minimizes fluctuations in the combined signal amplitude and reduces acoustic pressure variations.

[0030] Worked Examples

[0031] An example of a five-tone combination using this principle has the following characteristics, with respect to first (Ti), second (T2), third (T3), fourth (T4) and fifth (T5) tones, having respective amplitudes A1, A2, A3, A4, As distributed around the central frequency of the third tone, and with a constant frequency spacing fstePbetween adjacent tones in the combination: Relative amplitudes, A: 1 :4:8:4:1 ;

[0032] Phase alignment condition, <j>: 0:0:0:TT:0.

[0033] The magnitude of the interference effects between two tones is determined by the product of their respective amplitudes. Additionally, a scalar of 1 is applied when the two tones have the same phase alignment condition, whilst a scalar of -1 is applied when the two tones have an opposite phase alignment condition.

[0034] Four tone pairs have a frequency separation of a single unit of fsteP. The products of their amplitudes are:

[0035] A1 x A2= 1 X 4 = 4

[0036] A2 x A3 = 4 x 8 = 32

[0037] A3 x A4 = 8 x -4 = -32

[0038] T4x T5= -4 x 1 = -4

[0039] The sum of their interference patterns is 4 + 32 - 32 - 4 = 0, meaning that the interference at fstePis balanced.

[0040] Three tone pairs have a frequency separation of a two offset units, 2fsteP- The products of their amplitudes are: Ai x As = 1 x 8 = 8

[0041] A2x A4 = 4 x -4 = -16

[0042] A3 X A5 = 8 x 1 = 8

[0043] The sum of their interference patterns is 8 - 16 + 8 = 0, meaning that the interference pattern at 2fstePis balanced.

[0044] Two tone pairs have a frequency separation of three offset units, 3fsteP. The products of their amplitudes are:

[0045] A1 x A4= 1 x -4 = -4

[0046] A2x A5= 4 x 1 = 4 The sum of their interference patterns is 4 - 4 = 0, meaning that the interference pattern at 3fstep is balanced.

[0047] The interference between the outer pair of tones, T1 and T5, cannot be cancelled in this way as there is no other pair with the same frequency spacing. In practice, in order to produce a flat signal, the outer pair of tones are configured to have a low amplitude relative to other tones in the combination.

[0048] The explanation set out above applies equivalently to a five-tone combination in which the phase offset of TT is applied to the second tone, rather than the fourth tone.

[0049] More generally, for the five-tone configuration set out above, the tone combination is the sum of the following products:

[0050] (A1 x A2) + (A2x A3) + (A1 x A3) + (A3 x As) + (A2x As) + (A1 x As ) — (A3 x A4) — (A4x As) — (A2X A4) - (AI X A4).

[0051] Disregarding the product of the outer tones, which do not cancel, and with T1 and Ts each having a weighting of 1 , the summation becomes: A2+ (A2X A3) + A3 + A3 + A2— (A3X A4) - A4— (A2X A4) - A4. With a symmetrically distributed set of relative amplitudes, such that (i.e. Ai = As, and A2 = A4), the summation becomes: A3 - A22.

[0052] For complete interference cancellation of these terms, the expression sums to zero, such that A22= 2A3. As such, if an amplitude weighting, x, is applied to A2 and A4, the required weighting of T3is (x2 / 2). This condition is satisfied in the example above, in which x is 4. The relative amplitudes of the combination are thus 1 : x : (x2 / 2) : x : 1 .

[0053] A similar algebraic solution for cancelling inter-tone interference with seven tones has amplitudes 1 : x : (x2 / 2) : ((z2- 2z) / 2x) : (x2 / 2) : x : 1 , with z = (x2 / 2), and phase conditions of either 0 : 0 : 0 : 0 : TT : 0 : 77, or TT : 0 : TT : 0 : 0 : 0 : 0. The resultant combinations (referred to previously as EQUINOX combinations) can exhibit a high level of amplitude compression, with peak amplitudes of less than 45% of the sum of the individual tone amplitudes.

[0054] In the case where x = 2, an interesting six-tone solution is yielded, with relative tone amplitudes of 1 : 2 : 2 : 0 : 2 : 2 : 1. ln this combination, each non-zero tone is over 10% of the total amplitude, and four of the tones have equal amplitude, making the combination particularly suitable for averaging. This combination is used effectively in high-precision, low-latency motion tracking applications such as musical instrument simulation or ‘air guitar’ application.

[0055] Further examples are illustrated in Table 1 below, with tone frequencies distributed symmetrically around a central frequency of 18,750 Hz, and with a frequency offset unit of 234.375 Hz. Such frequencies and spacings enable seven tones to be used which all have a frequency of more than 18 kHz, with an upper limit of less than 19,500 Hz, which is within the output capabilities of current-generation mobile phone speakers and microphone sensitivities. The precise frequencies and tone spacings are such that the tone combination loops precisely in 1024 frames of a 48 kHz sampling rate. Furthermore, the tone frequencies correspond precisely to bin frequencies of a Radix-2 FFT used in embodiments of the present invention to estimate motion.

[0056] Table 1 : Examples for tone frequencies 18,750 + offset * 234.375 Hz; <|) refers to a phase alignment condition; Tone amplitudes A are normalised by their sum SA; Max refers to the maximum value of the combined signal The motion detection apparatus may be implemented across various applications that benefit from contactless motion sensing capabilities. Different tone configurations provide advantages for specific use cases based on their precision requirements and duration of operation. In embodiments configured for sleep tracking applications, the apparatus monitors breathing patterns and chest movements over extended periods. The motion estimator analyzes spectral shifts over time to detect changes in the distance of reflective surfaces, such as chest movements during respiration. In embodiments configured for gaming applications requiring motion detection, the apparatus detects objects such as hand gestures or body movements. In further embodiments, user interface control through contactless motion detection enables hands-free interaction with devices. The motion estimator analyzes reflected acoustic signals to translate detected movements into control commands for device operation.

[0057] The apparatus and methods described herein may be implemented in various combinations of hardware, software, firmware, or any combination thereof. The implementation may be tailored to specific applications, performance requirements, and device constraints.

[0058] In some embodiments, the motion estimator may be implemented as software executing on a general-purpose processor. The processor may be a central processing unit, digital signal processor, or other suitable processing unit capable of executing instructions stored in memory. The motion estimator may alternatively be implemented in dedicated hardware, such as an application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other specialized processing circuitry. In some implementations, the motion estimator may comprise a hybrid approach combining both hardware and software components. For example, signal acquisition and preprocessing may be performed by dedicated hardware, while higher-level motion analysis and interpretation may be implemented in software. This approach may balance performance requirements with implementation flexibility.

[0059] Features of any of the examples or embodiments outlined above may be combined to create additional examples or embodiments without deviation from the scope of the claims. Those skilled in the art will recognise that numerous further modifications and combinations of various aspects are possible. Accordingly, the described aspects are intended to encompass all such alterations, modifications, and variations that fall within the scope of the appended claims.

Claims

Claims1 . An apparatus for estimating motion of a target, comprising: a speaker, configured to output a combination of three or more acoustic pure tones having frequencies and relative amplitudes symmetrically distributed about a central pilot frequency, wherein the spacing between each tone frequency is an integer multiple of an offset frequency unit; a microphone for receiving and sampling reflections, by the target, of the three or more tones; and a motion estimator for estimating motion of the target from frequency analysis of the reflected three or more tones; wherein periodically, the tones have a phase alignment such that a majority of the tones align in phase, while all other tones have a phase offset of TT radians from the majority of the tones, and the relative amplitudes are such that all interference patterns between constituent tones in the combination are counterbalanced, with the exception of interference patterns between the two tones with greatest frequency spacing.

2. An apparatus according to claim 1 , wherein the speaker is configured to output a first tone and a third tone having respective first and third frequencies which are offset respectively below and above a central second frequency of a second tone by one offset frequency unit, and the majority of the tones comprises either: the first tone and the second tone, or the second tone and the third tone.

3. An apparatus according to claim 2, wherein the first, second and third tones have relative amplitudes in the ratio of 1 : x :

1.

4. An apparatus according to claim 1 , wherein the speaker is configured to output: a first tone and a fifth tone having respective first and fifth frequencies which are offset respectively below and above a third central frequency of a third tone by two offset frequency units, anda second tone and a fourth tone having respective second and fourth frequencies which are offset respectively below and above the third central frequency of the third tone by one offset frequency unit, and the majority of the tones comprises either: the first tone, the second tone, the third tone and the fifth tone, or the first tone, the third tone, the fourth tone and the fifth tone.

5. An apparatus according to claim 4, wherein the first, second, third, fourth and fifth tones have relative amplitudes in the ratio of 1 : x : (x2 / 2) : x : 1.

6. An apparatus according to claim 1 , wherein the speaker is configured to output: a first tone and a seventh tone having respective first and seventh frequencies which are offset respectively below and above a fourth central frequency of a fourth tone by three offset frequency units, and a second tone and a sixth tone having respective second and sixth frequencies which are offset respectively below and above the fourth central frequency of the fourth tone by two offset frequency units, and a third tone and a fifth tone having respective second and sixth frequencies which are offset respectively below and above the fourth central frequency of the fourth tone by one offset frequency unit, and the majority of the tones comprises either: the first tone, the second tone, the third tone, the fourth tone, and the sixth tone, or the second tone, the fourth tone, the fifth tone, the sixth tone and the seventh tone.

7. An apparatus according to claim 6, wherein the first, second, third, fourth, fifth, sixth and seventh tones have relative amplitudes in the ratio of 1 : x : (x2 / 2) : ((z2- 2z) I 2x) : (x2 / 2) : x : 1 , with z = (x2 / 2).

Images