close

Вход

Забыли?

вход по аккаунту

?

JP2007290691

код для вставкиСкачать
Patent Translate
Powered by EPO and Google
Notice
This translation is machine-generated. It cannot be guaranteed that it is intelligible, accurate,
complete, reliable or fit for specific purposes. Critical decisions, such as commercially relevant or
financial decisions, should not be based on machine-translation output.
DESCRIPTION JP2007290691
A vehicle communication system and a method of controlling speech output of the vehicle
communication system. A plurality of microphones (110, 111, 112) adapted to detect speech
signals of different vehicle passengers, and a mixer (140) for combining audio signal components
of different microphones into a resulting speech output signal And a weighting unit (130) for
determining the weighting of the audio signal component to the resulting speech output signal,
the weighting unit taking into account non-acoustic information about the presence of a vehicle
passenger. Determining a weight of the vehicle. [Selected figure] Figure 1
Vehicle communication system
[0001]
The present invention relates to a vehicle communication system and a method for controlling
speech output of said vehicle communication system.
[0002]
In vehicles, the use of communication systems has increased dramatically over the last few years.
In current vehicles, communication modules are often incorporated. These communication
systems usually include a number of microphones positioned at predetermined locations in the
vehicle. It may happen that one microphone is used for one vehicle seat position to receive the
04-05-2019
1
voice signal of the user. Furthermore, it is possible to use an array of microphones for one
vehicle seat position. Such an array of microphones can be used to receive vehicle passenger
voice signals for a given direction. The reception of these direction limited signals is also known
as the expression "beamforming". The most common application is hands-free telephony, where
the other party is outside the vehicle.
[0003]
Communication systems in vehicles may be used for different purposes. First of all, it is possible
to use human speech to control certain electronic devices with speech commands. In addition,
telephone calls in conference calls can be with more than one subscriber in the vehicle. In this
example, the person sitting in the front seat and the person sitting in one of the back seats use a
hands-free communication system in the vehicle with a third party at the other end of the line. I
can talk. Furthermore, it is possible to use a communication system in the vehicle for the
communication of different vehicle passengers to each other. If several people sitting in different
seats in the vehicle use the same communication system, different microphones provided in the
vehicle receive voice signals of different passengers. The mixer combines different audio signal
components of different microphones into the resulting speech output signal. To that end, the
communication system needs to weight different signal components detected by different
microphones to generate the resulting speech output signal. The weighting of the signal
components is usually by time and by the passenger's speech activity of the recorded voice
signal. In systems known in the art, speech activity for each vehicle seat is determined by
processing an audio signal detected by the microphone or array of microphones for a given
vehicle seat. In these systems, speech activity for a given vehicle seat needs to be controlled
during use of the communication system. It is possible that the passenger does not speak at the
beginning of the conversation, but communication with others starts during the conversation. In
all these cases, the system needs to verify during the entire conversation whether a speech signal
can be detected for a given vehicle seat.
[0004]
This can lead to false detection of the speech signal, so that the speech signals of other vehicle
passengers can be detected by a predetermined microphone. This false detection usually reduces
the weight of the passengers actually talking, resulting in speech output signals of low signal
quality. Furthermore, the weights for different vehicle seats are not set to zero to avoid that the
natural speech of the passengers is also transmitted via the communication system. If the weight
for a given vehicle seat is set to zero, the beginning of the utterance is not transmitted. This is
04-05-2019
2
why vehicle seat speech signals do not have zero weighting in known communication systems.
[0005]
Therefore, there is a need to improve the vehicle communication system such that the resulting
speech output signal reflects the actual presence of the passenger in the vehicle. This need is met
by the features of the independent terms. In the dependent claims, preferred embodiments of the
invention are described.
[0006]
According to a first aspect of the present invention, a vehicle communication system includes a
plurality of microphones adapted to detect speech signals of different vehicle passengers. In
addition, a mixer is provided that combines audio signal components of different microphones
into the resulting speech output signal. A weighting unit determines the weighting of the audio
signal component to the resulting speech output signal. According to the invention, the weighting
unit determines the weighting of the signal component taking into account non-acoustic
information about the presence of a vehicle passenger. This non-acoustic information can
determine with high probability whether a vehicle passenger is present on a given vehicle seat. In
prior art systems, only acoustic information was used to determine the weighting of the different
signal components. However, these acoustic information do not provide, with high probability,
information as to whether a predetermined acoustic signal is coming from a predetermined
vehicle seat. By taking into account non-acoustic information, it can be determined whether the
vehicle seat is not occupied. Due to the presence of passengers on the seat, the resulting speech
output signal can be generated taking into account the signal component from the vehicle seat or
without taking into account the signal component of the vehicle seat.
[0007]
According to one embodiment of the invention, a vehicle seat pressure sensor is used and the
weighting unit takes into account the signal from the pressure sensor to determine the weighting
of the audio signal component. The pressure sensor can determine with high accuracy whether
the passenger is seated in the vehicle seat or not. If the pressure sensor of a given vehicle seat
determines that no one is seated in the seat, the weighting for the signal component for the seat
may be set to zero. In this embodiment, the system determines the vacant seat and controls the
04-05-2019
3
weighting unit such that the weighting of the vacant seat signal component is set to zero.
[0008]
Furthermore, it is possible to use non-acoustic information provided by the image sensor.
According to this embodiment of the invention, an image sensor is used and the weighting unit
determines the weighting of the audio signal component taking into account the signal from the
image sensor. For example, the camera can take pictures of different vehicle seats. If no
passenger is detected on the vehicle seat, the weighting for the vehicle seat microphone may be
set to zero. The weighting of the other vehicle seats in which the passenger is detected can now
be adjusted according to the acoustic information detected. Thus, signal components in the
absence of a speech signal are not used to generate the resulting output signal, as the quality of
the resulting speech output signal may be improved. If a camera is used, it is also possible to
control the camera so that moving pictures can be generated. Hereby, the speech activity of a
person can be determined from the picture of the camera, for example by detecting lip
movement. This information can be used to detect the passenger's speech activity for which the
picture was taken. If it is determined that the passenger on the seat does not speak using the
image signal, the noise can be suppressed in a more effective manner, thus distinguishing
between the moment when the passenger is speaking and the moment when the passenger is not
speaking Things are easier.
[0009]
According to another aspect of the invention, the loudspeakers of the vehicle communication
system can also be adapted to the presence of vehicle passengers. If several loudspeakers are
used, it is possible to direct the emitted sound in a given direction by introducing a time delay in
the signals emitted by the different loudspeakers. If the system knows that a given vehicle seat is
not occupied, the loudspeakers of the vehicle communication system may be optimized for the
person or persons present in the vehicle. The beamforming of this audio signal is performed
using any audio signal emitted by the loudspeaker so that it becomes a music or voice signal for
communication with a third party.
[0010]
In the in-vehicle communication system, a loudspeaker located near the passenger is used for the
04-05-2019
4
playback signal to this passenger. If the seat is not occupied by a person, the in-vehicle
communication system does not have to play back the signal via the corresponding loudspeaker.
Thus, the risk of howling is significantly reduced and system stability is increased.
[0011]
Surround sound systems aim to optimize sound quality and sound effects for different seats.
There is always a compromise on quality as the effect needs to be present for all seats. If the seat
is not occupied by a person, there is no need to optimize the sound quality for this position. Thus,
the quality for other seat positions can be further optimized.
[0012]
Often, microphone arrays are used to detect audio signals of different vehicle seats. For example,
a microphone array of two to four microphones may be configured in the front portion of the
vehicle, receiving speech signals from the driver's seat and from the front seats of other
passengers. These different signals detected by one array of microphones may then be provided
to a beamformer that combines the signals detected from the array of microphones to produce a
vehicle seat speech signal for each of the vehicle seats.
[0013]
According to another aspect of the present invention, there is provided a method of controlling
speech output of a vehicle communication system comprising the following steps. Speech signals
of at least one vehicle passenger are detected using a plurality of microphones. Furthermore, the
different speech signal components detected by the different microphones are weighted, and the
weighted speech component signals are combined into the resulting speech output signal.
According to the invention, the weighting of the different speech signal components is
determined taking into account non-acoustic information about the presence of a vehicle
passenger. For example, the seat pressure sensor signal may be used to detect the presence of a
passenger and / or the image sensor signal may be used to detect the presence of a passenger on
a different vehicle seat position. If it is detected that a passenger can not be detected for a given
vehicle seat position, the signal component for the vehicle seat position is set to zero. Since this
serves to improve the signal quality of the speech output signal generated, it is possible to take
into account only the signal component of the vehicle seat position at which the passenger is
04-05-2019
5
present.
[0014]
Preferably, when an array of microphones is used, an audio signal component for the vehicle seat
position is determined, a weighting of the signal component is determined, and a speech signal
beam is formed for the vehicle position. Beamforming is then performed for different seat
positions. If it is detected that there are no passengers for a given vehicle seat position, the
weighting for the unoccupied vehicle seat position is set to zero.
[0015]
These and other aspects of the invention will be apparent from the embodiments described
below.
[0016]
The present invention further provides the following means.
(Item 1) A plurality of microphones (110, 111, 112) adapted to detect speech signals of different
vehicle passengers, and a mixer (140) for combining audio signal components of different
microphones into the resulting speech output signal And a weighting unit (130) for determining
the weighting of the audio signal component to the resulting speech output signal, the weighting
unit taking into account non-acoustic information about the presence of a vehicle passenger.
Determining a weighting of the at least one weighting unit. (Item 2) The vehicle seat pressure
sensor (160) is further provided, and the weighting unit determines the weighting of the audio
signal component in consideration of the signal from the pressure sensor, in item 1 Vehicle
communication system as described. (Item 3) The vehicle according to item 1 or 2, further
comprising an image sensor, wherein the weighting unit determines the weighting of the audio
signal component in consideration of a signal from the image sensor. Communications system.
(Item 4) Any of items 1 to 3, further comprising a plurality of loudspeakers for outputting an
audio signal, the use of different loudspeakers being according to the above information about
the presence of the vehicle passenger. Vehicle communication system according to one of the
claims. (Item 5) The vehicle communication system according to any one of items 1 to 4, wherein
the image sensor detects speech activity of a passenger vehicle. (Item 6) The apparatus further
includes a beam forming unit that generates a vehicle seat speech signal combining the above
04-05-2019
6
signals detected from a plurality of microphones that receive speech signals from one passenger
sitting on the vehicle seat. Vehicle communication system according to any one of 1 to 5. (Item 7)
When the presence of the passenger at a predetermined vehicle seat position can not be
detected, the weighting unit sets the weighting of the signal component of the vehicle seat
position to zero. The vehicle communication system according to any one of items 1 to 6. 8. A
plurality of microphones adapted to detect speech signals of different vehicle passengers, a mixer
combining audio signal components of different microphones into the resulting speech output
signal, and the presence of an unoccupied vehicle seat A seat occupancy detection unit for
detecting the weight, and a weighting unit for determining the weighting of the audio signal
component to the resulting speech output signal, the weighting unit determining the weighting of
the audio signal component of the unoccupied seat A weighting unit, which is set to zero.
9. A method of controlling speech output in a vehicle communication system, comprising:
detecting speech signals of at least one vehicle passenger using a plurality of microphones (110,
111, 112); Weighting the different speech signal components and combining the weighted
speech component signals into the resulting speech output signal, wherein the weighting of the
different speech signal components represents non-acoustic information about the presence of a
vehicle passenger. The way it is determined taking into account. 10. The method of claim 9,
further comprising the step of determining audio signal components for different vehicle seat
passenger positions, wherein the weighting of the signal components is determined for the
different vehicle seat positions. the method of. 11. The method of claim 10, wherein the
weighting of the signal component for a given vehicle seat position is set to zero if no passenger
is detected on the vehicle seat position. 12. The resulting speech signal is for: voice-controlled
operation of vehicle components, conference call with an external subscriber and at least two
vehicle passengers and / or communication of different vehicle passengers with one another 12.
The method according to any one of items 9-11, used for 13. A method according to any one of
claims 9 to 12, further comprising the step of adding different weighted signal components
detected by the microphone. 14. The method of claim 9, further comprising the step of
controlling the output of audio signals via a plurality of loudspeakers according to the nonacoustic information about the presence of the vehicle passenger for a predetermined vehicle
position. The method according to any one of 13. 15. The method of claim 14, wherein the
output of the audio signal through the loudspeaker is optimized for the vehicle seat position
where it is determined that a passenger is present on the seat. 16. The method of any of claims 915, wherein the signal of a seat pressure sensor is detected to detect the presence of the
passenger. 17. The method according to any one of claims 9 to 16, wherein the signal of an
image sensor is detected to detect the presence of the passenger. (Item 18) The method
according to any one of items 9 to 17, further comprising the step of detecting speech activity of
a vehicle passenger taking into account the signal from the image sensor. Method.
04-05-2019
7
19. A method of controlling the speech output of a vehicle communication system, comprising:
detecting speech signals of at least one vehicle passenger using a plurality of microphones (110,
111, 112); Weighting the speech signal component, combining the weighted speech component
signal into the resulting speech output signal, and detecting the presence of an unoccupied seat,
the unoccupied seat And wherein the weighting of the signal components of is set to zero.
[0017]
The present invention comprises a plurality of microphones (110, 111, 112) adapted to detect
speech signals of different vehicle passengers and a mixer for combining audio signal
components of different microphones into the resulting speech output signal (140) and a
weighting unit (130) for determining the weighting of the audio signal component to the
resulting speech output signal, the weighting unit taking into account non-acoustic information
about the presence of a vehicle passenger And a weighting unit that determines the weighting of
the signal component.
[0018]
In FIG. 1, a vehicle communication system is shown in which speech output signals are generated
using non-acoustic information about the presence of a passenger.
The communication system includes several microphones to receive the passenger's audio signal.
For example, in the present case, four microphones are positioned in the front microphone array
110 of the vehicle to detect speech signals in the driver's seat and the other front passenger's
seat. Furthermore, a microphone 111 is provided for detecting the speech signal of the passenger
sitting in the rear on the left side, and the microphone 112 is configured to receive the speech
signal of a person sitting in the rear on the right side. The four microphones of array 110 form a
speech signal x 1 (t) for the driver at beamformer 120. In the beamformer 120, the four
microphone signals are processed to use signals originating primarily from the driver's direction.
The same is done for the other front seat. Here, the signals of the four microphones of the array
110 are used to determine the signal x2 (t) at the beamformer 121 of the front seat.
Microphones 111 and 112 receive the speech signals of the seats at the rear on the left and right
sides respectively. Here, in the present case, only one microphone is used at the rear so that no
beamformer is required. In the noise reduction unit 122, the right rear speech signal is processed
by a noise reduction algorithm. The same is done for the speech signal detected by the
microphone 111 in the noise reduction unit 123.
04-05-2019
8
[0019]
Organizing the seat-related speech signals is determined for each different vehicle position. In
the embodiment shown, four different positions are possible for speech to be detected. For each
passenger, a signal xp (t) is calculated. From the different passenger signals xp (t), the resulting
speech output signal y (t) is calculated using the following equation:
[0020]
The largest number of passengers participating in communication is P, and the factor ap (t) is a
weighting for different users of the communication system. As can be seen from the above
equation, the weighting is time dependent, so the resulting output signal should be generated
from the signal component of the passenger who is actually talking. The weighting of different
signal components is determined by the weighting unit 130. In weighting unit 130, different
weights ap (t) are calculated to produce the resulting speech output signal y (t) and provided to
mixer 140 which mixes different vehicle seat speech signals. Additionally, a passenger detector
150 is provided that uses non-acoustic information about the presence of a vehicle passenger for
different vehicle seat positions. The passenger detector may use different sensors 160 to detect
the presence of passengers on different vehicle seats. For example, sensor 160 may be a pressure
sensor that detects the presence of a passenger on a seat by the weight provided on the seat.
Furthermore, it is possible to use cameras that take pictures from different vehicle seats. If a
camera is used, video information may also be used to detect the passenger's speech activity by
detecting lip movement. If the lips are moving and the passenger is talking, the weighting of the
signal to the passenger may be increased, but the weighting may be reduced to the passenger not
actually talking. As shown in the figure, when it is detected that there are no passengers on the
vehicle seat for the right front seat and the left rear seat, the weighting factor for the seat related
speech signal xp (t) is: May be set to zero. In the embodiment shown, this means that the weights
for the signals x2 (t) and x4 (t) are set to zero so that the signal components from these vehicle
seats do not contribute to the output signal y (t) means. The weighting of the vacant seat signal
component is set to zero.
[0021]
The different steps required to calculate the improved output signal y (t) are described with
reference to FIG. The process starts at step 210. At step 220, the speech signal is detected using
different microphones located in the vehicle. At step 230, the audio signals detected by the
04-05-2019
9
microphones 110 through 112 are combined to generate a seat-related speech signal xp (t) for
each vehicle seat (step 230). Furthermore, occupancy of different vehicle seats is detected at step
240. At step 250, the occupancy condition sets the signal component for which no passenger
was detected to zero. This helps to avoid using signal components detected by the microphone
where the passenger is not sitting. After setting the seat related speech signal components to
zero, at step 260, the remaining seat related speech signals are combined and weighted by the
speech activity of different passengers. The process ends at step 270.
[0022]
In connection with FIG. 3, it is explained how information about the presence of passengers on
the vehicle seat can be used by a loudspeaker (not shown in FIG. 1) to improve the audio signal
output . After starting the process at step 310, at 320, occupancy of different vehicle seats may
be detected using either pressure sensors or using a camera or both sensors. For vehicle seats
where there are no passengers, there is no need for the loudspeaker to produce an audio output.
If several loudspeakers are used, it is possible to form a sound beam directed to one of the
vehicle seats. If it is known that one of the vehicle seats, for example the seat next to the driver's
seat, is not occupied, the loudspeaker can be controlled so that the sound beam is directed to the
occupied vehicle seat (step 330). With this beamforming, the sound can be concentrated on the
person sitting on the vehicle seat. The weighting for the formed sound beam formed for the
vacant seat may be set to zero. This is further because the microphone for receiving the
passenger's speech signal, which is usually positioned at a distance of about 30 cm to 70 cm,
receives the smaller sound emitted from the loudspeaker so that the sound is directed more
accurately to the passenger's head Have an advantage. The process ends at step 340.
[0023]
Summarizing the present invention, the present invention provides a system and method having
an enhanced speech output signal so that signal components that are free of passengers can be
reliably eliminated. There is no need to corrupt the signal components detected by the
microphones located in the vicinity of the vehicle seat where the passengers are not sitting.
Consequently, less signal needs to be considered to calculate the output signal, and the use of
signal components detected from the microphones from other passengers can be avoided.
[0024]
04-05-2019
10
FIG. 1 shows a schematic view of a vehicle communication system taking into account nonacoustic information. FIG. 2 is a flow chart showing different steps for determining an enhanced
output signal. FIG. 3 shows a flow chart of the method in which the audio output is optimized for
the actual existing vehicle passengers.
Explanation of sign
[0025]
DESCRIPTION OF SYMBOLS 110 Microphone array 111, 112 Microphone 120, 121 Beam
former 122, 123 Noise reduction part 130 Weighting part 140 Mixer 150 Passenger detection
part 160 Sensor
04-05-2019
11
Документ
Категория
Без категории
Просмотров
0
Размер файла
22 Кб
Теги
jp2007290691
1/--страниц
Пожаловаться на содержимое документа