Patent Application
20230351996
Not Applicable.
Not Applicable.
The present invention relates in general to transportation vehicles, and, more specifically, to providing a motor vehicle having a passenger cabin adapted for performing as an office or other work environment which can maintain privacy of conversations/teleconferences in the passenger cabin.
In addition to traveling from place to place, vehicle consumers are becoming increasingly interested in engaging in diverse activities within their vehicles. For example, a vehicle and its electronic accessories may be utilized to institute an office-type work environment including work surfaces, business machines, and/or telecommunications services. For example, construction personnel at a construction site or other job site may need a protected space in which to review or update documents or to participate in teleconference calls with remote parties. An interior passenger cabin of a vehicle can be advantageously configured to provide such a space. Unlike traditional offices in a fixed building or home, however, a vehicle work environment may be subject to public perception when pedestrians are justifiably close to the vehicle. This creates a need for privacy measures in order to meet the usual expectations for standard workplaces.
In some instances, two or more persons may wish to share the space which can lead to potential disruption between their separate activities. For example, one or more occupants of the vehicle may conduct a teleconference with a remote party using an open microphone(s) and speakers wherein the conversation is distracting to others in the vehicle, or the contents of the conversation are meant to be kept private from others inside or outside the vehicle. It would be especially desirable to prevent the hearing of private conversations by persons outside the vehicle. In addition, multiple occupants in the vehicle may attempt to conduct separate teleconferences (or a teleconference and a separate face-to-face conversation) simultaneously so that the audio content of one discussion may impinge on the other. It would also be desirable to filter out the effects of any crossover between conversations.
The present invention helps ensure that pedestrians (e.g., people passing nearby a vehicle) do not hear private conversations occurring inside the vehicle. For example, microphones inside the vehicle can be used to measure the loudness of the private sounds, and then an evaluation is made whether it is possible for a pedestrian to hear the private conversation. If so, then a masking noise is projected outside the vehicle to the pedestrian which renders the private conversation undiscernible.
For purposes of having multiple, independent conversations in the cabin of the vehicle, a controller may determine how many occupants are present (e.g., using an interior camera, seat sensors, or other means) and tracks the positions and orientations of their head/ear/mouth while they are participating on a teleconference call. The positions/orientations are used to adjust speaker and/or microphone performance to optimize each call to the corresponding occupant(s). Furthermore, the vehicle controller may also employ audio signatures of each occupant to process audio signals from the microphones to isolate audio content of one particular occupant from the other occupants. Separating the audio signals according to each speaking person can be achieved using artificial intelligence (AI) or Machine Learning (ML) de-mixing technology as known in the art. Spoken audio corresponding to each individual occupant can then be steered to the appropriate remote party with less interference whenever other conversations are occurring in the vehicle.
In one aspect of the invention, a transportation vehicle comprises a passenger cabin defining an interior, an exterior, and at least one seating location in the interior for conducting a teleconference with a remote party. At least one internal microphone receives a first portion of an audio content of the teleconference spoken by a vehicle occupant. At least one internal speaker generates a second portion of the audio content of the teleconference spoken by a remote party. External sensors are configured to scan the exterior of the vehicle and configured to detect a pedestrian in the exterior. A sound exciter is configured to generate a masking noise directed to the exterior. A control circuit which enhances privacy of the teleconference is configured to A) detect the audio content of the teleconference being present in the interior of the vehicle, B) quantify an interior sound level of the audio content, C) estimate a discernability of the audio content at an external location corresponding to the detected pedestrian, and D) activate the sound exciter such that the masking noise is adapted to mask the audio content at the external location.
In addition to desk-like surfaces and large built-in display screens for enabling office productivity applications, a vehicle passenger cabin can be configured to provide personal sound zones adapted to provide some audio isolation between zones during teleconferencing calls. For example, interior cameras may track ear/head locations and orientations in different occupant zones to adjust audio sound production and/or microphone sensitivity for targeting respective occupants of the zones. The microphones or loudspeakers may be directional. Several microphones may be deployed in the walls, ceiling, instrument panel, or other structures of the passenger cabin, and loudspeakers may be deployed in those locations and in the passenger headrests. Personal mobile devices may also be used (e.g., an occupant's smartphone linked by a Bluetooth® connection).
A vehicle controller may be used to filter, combine, isolate, or pass-through various audio signals as needed to facilitate an optimized teleconferencing experience for each occupant in the vehicle to ensure other participants on a call only hear audio from the correct vehicle occupant as well as to ensure the occupants of the vehicle do not hear the conversations from others in the vehicle. The vehicle controller may control microphone settings, speaker volume and cancelation properties, for example. The vehicle controller may access interior cameras to determine which occupants are speaking in order to steer the audio signals accordingly.
Referring to
A control module 18 is configured to provide audio processing associated with one or more teleconferences being conducted within cabin 11 using corresponding ones of microphones 15 and loudspeakers 16/17. Control module 18 may include, or is connected to, a wireless transceiver such as a cellular transceiver for communicating with a remote site 20 (such as a cellular telephone base station). Site 20 is coupled to a telephone network 21 for completing a teleconference call with a remote party 22.
A human-machine interface (HMI) 39 is coupled to control module 18 for enabling vehicle occupants to perform setup and operations commands which control one or more teleconferences. HMI 39 may include a touchscreen display panel, a voice command interface, or other interfaces as known in the art. Among the commands that may be initiated via HMI 39 are commands for designating whether a teleconference is of a sensitive nature and should be kept private from passersby.
Control module 18 includes a communications controller 36 coupled to antenna 19. Communications controller 36 functions as a wireless transceiver (e.g., a cellular transceiver for carrying out cellular phone calls or a Bluetooth® node for exchanging audio signals with an occupant's mobile phone which completes a call). A first call interface 37 processes two-way audio for a first call conducted by an occupant in the first zone, and a second call interface 38 processes two-way audio for a second call conducted by an occupant in the second zone. Due to the relatively small size of a vehicle passenger cabin, some amount of crosstalk of sounds 29 between the first and second zones is likely to occur so that audio exchanged in one call includes an added signal from other sound in the cabin (e.g., another call in the other zone). In order to isolate the desired voice for a particular call from other calls or extraneous sounds in the vehicle, a de-mixer 40 receives all the microphone signals from microphones 27 and 30. As known in the art, an audio signature of each occupant in the vehicle can be obtained using de-mixing techniques employing artificial intelligence and/or machine learning. After filtering out audio content not to be included in the teleconference for Call 1, de-mixer 40 sends an audio signal to call interface 37 which extracts the speaking of the occupant in seat 25. Likewise, after filtering out audio content not to be included in the teleconference for Call 2, de-mixer 40 sends an audio signal to call interface 38 which extracts the speaking of the occupant in seat 26.
To ensure pedestrians (i.e., passersby or any other bystanders located within a close proximity to the vehicle) do not hear private conversations from the inside of the vehicle, a masking noise may be distributed to the exterior in appropriate circumstances to interfere with the ability of the pedestrians to discern the contents of the conversations. In particular, microphones inside the vehicle can determine the loudness of the conversation and whether it is possible for a pedestrian to perceive the conversation occurring in the vehicle. The invention may only take action when it is determined that the content of the conversation is discernible to the pedestrian (i.e., that the speech would be intelligible based on typical loudness thresholds and acoustic properties of the sounds such as frequency spectrum. While a private conversation is on-going and the vehicle is not moving (or moving at very low speed), exterior monitoring sensors such as radars and cameras can scan the environment to determine the distance/direction of pedestrians relative to the vehicle. Conversation loudness derived by an audio system for the teleconference can be compared to a threshold (e.g., a loudness threshold). The threshold may be determined during vehicle development to correlate a pedestrian's distance to the discernibility of the conversations. In some embodiments, a distance threshold based on the loudness of the conversation is compared to a measured distance of the pedestrian(s). Thus, the threshold may characterize discernibility based on the pedestrian location and the actual loudness of the conversation. When the conversation is discernible, sound exciters (e.g., loudspeakers) on the vehicle exterior can be used to play masking noise which renders the conversation undiscernible. Activation of the masking noise can have time hysteresis such that once it is enabled, it does not disable for a set period of time or until the pedestrian moves away to a distance greater than a predetermined threshold distance.
As used herein, masking noise means any sounds perceptible to the pedestrian which reduce the signal to noise ratio with reference to the private conversation sounds as evaluated at the location of the pedestrian. The masking noise can include music, predetermined tones, or random noise. Random noise may include white noise (i.e., broadband randomized signals). However, a random noise with a frequency spectrum which emphasizes the frequencies corresponding to the spoken conversation to be masked (e.g., 250 Hz to 4 kHz) is more preferable. The random noise can be derived from any uncorrelated signals that when added to the sound of the private conversation at the location of the pedestrian results in a reduction of the signal to noise ratio which is sufficient to make the conversation indiscernible. This may, for example, necessitate that a loudness of the masking noise be greater than a loudness of the conversation by a predetermined margin (e.g., measured in dB at the pedestrian location).
In some embodiments, the measured interior sound level of the audio content can be used to determine a threshold distance measured from vehicle 55 at which loudness of the spoken conversation becomes discernible to a typical pedestrian. A region 62 in
Controller 70 includes a privacy detection block 74 coupled to control logic 71. Preferably, the exterior sound exciter can be used to generate masking noise only when a privacy mode has activated. Privacy detection block 74 can initiate the privacy mode according to a manual activation by a vehicle occupant (e.g., via the HMI), automatically according to a scheduled time for a teleconference (e.g., identified via an occupant's smartphone calendar), automatically according to a predetermined list of call participants, or by detecting a keyword spoken by a vehicle occupant matching a predetermined list of keywords which identify private subject matter.
For purposes of estimating a discernability of the audio content at an external location corresponding to the detected pedestrian, controller 70 includes a sound analyzer block 75 which determines a sound attenuation between the vehicle interior and the external location (or a range of locations when determining a region such as region 62 in
Based on the evaluated sound levels, a check is performed in step 83 to determine whether private audio may be discernible (i.e., intelligible) at the listener's location. If not discernible, then steps 80-82 may be repeated as necessary. If it is determined that the private audio may be discernible, then a level and/or content of masking noise which can make the private audio undiscernible is determined in step 84. The content of masking sounds can include music or other prerecorded media (tones or speech) or can be comprised of random noise (preferably having a frequency spectrum spanning the frequency content of the spoken conversation). The frequency range of the conversation can be either measured for the actual spoken conversation to be masked or estimated according to a typical voice (e.g., 150 Hz to 10 kHz). In some embodiments, the frequency of random masking noise may dither around an estimated center frequency of the voice.
In order to determine an appropriate level for generating the masking noise, the distance from the sound source (exciter or loudspeaker) to the pedestrian may be determined. Based on the distance, a corresponding attenuation of the masking noise is determined. Using an estimate of the remote sound level of the audio content of the private conversation when it reaches the pedestrian, a target sound level of the masking noise can be determined which produces a signal-to-noise ratio such that the conversation becomes unintelligible. Some estimates of speech intelligibility suggest that a signal-to-noise ratio of about 12 dB is needed for understanding speech (i.e., the masking noise does not need to be as loud as the sound to be masked). Thus, a target sound level for the masking noise can be selected which degrades the remote signal-to-noise ratio below 12 dB. To determine the masking level to be generated at the vehicle, the target level of the masking noise is increased by the attenuation between the masking noise source and the pedestrian. In some circumstances, it may be desirable to impose a limit on the target sound level. For example, there may be a noise ordinance in effect, or the vehicle controlled may determine that the vehicle is in a location where any sound generation is undesirable such as near a hospital, school, or place of worship. When such a limit is imposed, the vehicle occupants can be informed that privacy cannot be obtained using masking noise and that it is recommended to switch a teleconference to use of a headset or a private mode on their smartphone.
In step 85, the masking noise is generated. Once activated, the masking noise may be maintained for a predetermined time in order to avoid annoying switching on and off of sounds. In step 86, the presence/locations of pedestrians and the sound levels of any private conversations continue to be monitored so that masking noise can be adjusted for changing conditions until masking is no longer needed (e.g., the private conversation ends or no pedestrians are present).
In a first embodiment, a sound attenuation is estimated from the vehicle interior to the pedestrian location in step 92. The attenuation estimate may be derived from a lookup table or model compiled from testing data. The table or model may take into account a position of a door, window, or moonroof. In step 93, an interior sound level threshold is determined based on a discernibility level or threshold at the pedestrian at which speech is intelligible (e.g., a signal-to-noise ratio of 12 dB or other parameter such as an absolute sound pressure level). In particular, the discernibility level at the pedestrian is increased by the inverse of the attenuation estimate between the vehicle interior and the pedestrian location to set the interior sound level threshold (i.e., the controller calculates the sound level needed in the interior in order to produce a discernible sound at the location of the pedestrian). A check is performed in step 94 to determine whether the actual (measured) sound level in the vehicle interior is greater than the threshold. If so, then masking noise is activated in step 91. Otherwise, a return is made to step 90 (via point A) for repeated monitoring of the interior sound level and the pedestrian distance.
In a second embodiment, a sound attenuation is estimated from the vehicle interior to the pedestrian location in step 95. The attenuation estimate may be derived from a lookup table or model compiled from testing data. The table or model may take into account a position of a door, window, or moonroof. In step 96, the measured interior sound level and the estimated attenuation between the vehicle interior and the pedestrian location are used to estimate an actual remote sound level of the private conversation at the remote location of the pedestrian. A check is performed in step 97 to determine whether the actual remote sound level of the audio content is greater than a discernible level for the pedestrian. The discernible level may correspond to a signal-to-noise ratio absent any competing noise (e.g., an absolute sound level of 12 dB or any other selected level). If greater, then masking noise is activated in step 91.
Otherwise, a return is made to step 90 (via point A) for repeated monitoring of the interior sound level and the pedestrian distance.
In a third embodiment, the level of audio content in the vehicle is used to set a trigger boundary around the vehicle. Thus, in step 100, an estimated attenuation from the vehicle interior along various outward directed vectors is determined (e.g., using a lookup table or model and the positions of windows and doors). Based on the attenuation, a threshold boundary is mapped around the vehicle according to a location in each direction where audio content at the currently measured sound level will attenuate below a discernible level for any pedestrian who may be present. Once the threshold boundary is set (see, e.g., a boundary around region 62 in
