# US 2007/0165881 A1 · System and method for acquiring acoustic signals using doppler techniques

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2007/0165881 A1
- Document ID: US20070165881A1
- Status: Published application
- Kind: A1
- Application number: 11/110,396
- Filed: 2005-04-20
- Published: 2007-07-19
- Earliest claimed: 2005-04-20
- Assignees: Mitsubishi Electric Research Laboratories, Inc.
- Inventors: Bhiksha Ramakrishnan; Paul Dietz; Bent Schmidt-Nielsen
- Web page: https://patentgrep.com/patent/US20070165881A1

## CPC

- G10L 21/0208: Noise filtering
- H04B 7/01: Reducing phase shift
- H04B 17/391: Modelling the propagation channel
- G10L 15/20: Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech (G10L21/02 takes precedence)
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- H04B: TRANSMISSION

## Abstract

A method and system enhances an acoustic signal acquired by a microphone from an acoustic source while concurrently acquiring a Doppler signal from moving parts of the acoustic source. The acoustic signal and the Doppler signal are then analyzed according to a model to generate an enhanced acoustic signal.

## Claims

### Claim 1

A method for enhancing a noisy acoustic signal, comprising:
acquiring an acoustic signal from an acoustic source;
acquiring concurrently an ultra-sonic Doppler signal reflected from moving parts of the acoustic source; and
analyzing jointly the acoustic signal and the ultra-sonic Doppler signal according to a model to generate an enhanced acoustic signal.

### Claim 2 (depends on 1)

The method of claim 1, in which the acquired acoustic signal includes noise, and the enhanced acoustic signal has a reduced amount of noise.

### Claim 3 (depends on 1)

The method of claim 1, further comprising:
generating the ultra-sonic Doppler signal.

### Claim 4 (depends on 1)

The method of claim 1, in which the ultra-sonic Doppler signal has a frequency greater than the acoustic signal.

### Claim 5 (depends on 1)

The method of claim 1, in which the acoustic signal is acquired by a microphone and the ultra-sonic Doppler signal is acquired by an ultra-sonic sensor.

### Claim 6 (depends on 1)

The method of claim 1, in which the acoustic source is a human speaker.

### Claim 7 (depends on 1)

The method of claim 1, in which the acoustic source is a machine with moving parts.

### Claim 8 (depends on 1)

The method of claim 1, in which the acoustic source is a loudspeaker.

### Claim 9 (depends on 6)

The method of claim 6, further comprising:
performing speech recognition on the enhanced acoustic signal.

### Claim 10 (depends on 6)

The method of claim 6, further comprising:
enhancing noisy speech in automobiles for hands-free telephony.

### Claim 11 (depends on 6)

The method of claim 6, further comprising:
enhancing noisy speech on cellular phones operating in noisy environments.

### Claim 12 (depends on 7)

The method of claim 7, further comprising:
diagnosing an operation of the machine.

### Claim 13 (depends on 1)

The method of claim 1, further comprising:
training the model with a plurality of clean acoustic spectra and a plurality of corresponding Doppler spectra.

### Claim 14 (depends on 1)

The method of claim 1, in which the model is in a form of a mixture of distributions.

### Claim 15 (depends on 14)

The method of claim 14, in which the distributions are Gaussian distributions.

### Claim 16 (depends on 14)

The method of claim 14, in which the model includes a plurality of states, each state modeling one acoustic spectrum and one corresponding Doppler spectrum.

### Claim 17 (depends on 16)

The method of claim 16, in which the model includes a priori probabilities of the states, state dependent distributions of the Doppler spectra, and state dependent distributions of the acoustic spectra.

### Claim 18 (depends on 17)

The method of claim 17, further comprising:
combining acoustic spectra and the Doppler spectra using a posteriori probabilities of the states of the model as weights to generate the enhanced acoustic signal.

### Claim 19 (depends on 1)

The method of claim 1, in which the model is a hidden Markov model.

### Claim 20 (depends on 1)

The method of claim 1, in which the model is a Bayesian network.

### Claim 21

A system for acquiring and enhancing an acoustic signal, comprising:
a microphone configured to acquire an acoustic signal from an acoustic source;
a sensor configured to acquire concurrently an ultra-sonic Doppler signal reflected from the acoustic source; and
means for analyzing jointly the acoustic signal and the Doppler signal according to a model to generate an enhanced acoustic signal.
