Patent Application
20240212663
The present disclosure relates generally to electronic circuits, and, more particularly, to a system and a method for controlling noise cancellation systems in vehicles.
Vehicles are widely used in day-to-day life for commuting, transporting cargo, or the like. Various types of noises are present in a cabin of a vehicle. Examples of such noises include a noise of an engine of the vehicle, a road-tire interaction noise, a wind noise, or the like. Such noises may cause discomfort to the driver and passengers of the vehicle.
The following detailed description of the embodiments of the present disclosure will be better understood when read in conjunction with the appended drawings. The present disclosure is illustrated by way of example, and not limited by the accompanying figures, in which like references indicate similar elements.
The detailed description of the appended drawings is intended as a description of the embodiments of the present disclosure and is not intended to represent the only form in which the present disclosure may be practiced. It is to be understood that the same or equivalent functions may be accomplished by different embodiments that are intended to be encompassed within the spirit and scope of the present disclosure.
In an embodiment of the present disclosure, a circuit is disclosed. The circuit may include an adaptive noise classifier and a controller coupled to the adaptive noise classifier. The adaptive noise classifier may be configured to receive a plurality of audio signals from a plurality of audio-capturing devices of a vehicle and determine a first acoustic state of the vehicle based on the plurality of audio signals. The controller may be configured to receive status data indicative of one or more functional attributes of the vehicle and determine a second acoustic state of the vehicle based on the first acoustic state and the status data. The second acoustic state may be an updated version of the first acoustic state. The controller may be further configured to control an operational state of each of a plurality of noise cancellation systems of the vehicle based on the second acoustic state.
In another embodiment of the present disclosure, a vehicle control method is disclosed. The vehicle control method may include receiving a plurality of audio signals from a plurality of audio-capturing devices of a vehicle and determining a first acoustic state of the vehicle based on the plurality of audio signals, by an adaptive noise classifier. Further, the vehicle control method may include receiving status data indicative of one or more functional attributes of the vehicle by a controller. The vehicle control method may further include determining, based on the first acoustic state and the status data, a second acoustic state of the vehicle and controlling an operational state of each of a plurality of noise cancellation systems of the vehicle based on the second acoustic state, by the controller. The second acoustic state may be an updated version of the first acoustic state.
In some embodiments, the one or more functional attributes of the vehicle may include at least one of a group consisting of a speed of the vehicle, a position of a plurality of windows of the vehicle, a status of a voice detector of the vehicle, a tire pressure of a plurality of tires of the vehicle, a revolution count of an engine of the vehicle, and a status of a heating, ventilation, and air conditioning (HVAC) system of the vehicle.
In some embodiments, the first acoustic state may be indicative of a dominant noise associated with the vehicle. The dominant noise may correspond to one of a group consisting of an engine noise of the vehicle, a road-tire interaction noise of the vehicle, and a wind noise in the vehicle.
In some embodiments, the adaptive noise classifier may be further configured to generate classification data that is indicative of the first acoustic state and provide the classification data to the controller. The controller may determine the second acoustic state based on the classification data and the one or more functional attributes of the vehicle.
In some embodiments, the controller may be further configured to validate the first acoustic state based on the status data. The second acoustic state may be further determined based on the validation of the first acoustic state.
In some embodiments, to control the operational state of each of the plurality of noise cancellation systems, the controller may be further configured to generate a plurality of control signals and provide the plurality of control signals to the plurality of noise cancellation systems. Each control signal controls activation and deactivation of a corresponding noise cancellation system.
In some embodiments, the adaptive noise classifier has a training phase and a deployment phase associated therewith. During the training phase, the adaptive noise classifier may be further configured to generate a plurality of model parameters based on a plurality of training signals. During the deployment phase, the adaptive noise classifier may determine the first acoustic state based on the plurality of model parameters and the plurality of audio signals.
In some embodiments, during the training phase, the adaptive noise classifier may be further configured to receive the plurality of training signals, with each training signal being indicative of an audio associated with the vehicle, and extract a first plurality of features from each training signal. The first plurality of features may include at least one of a group consisting of a time period, a measure of periodicity, a spectral envelope, one or more cepstral attributes, and one or more spectral band energy levels associated with the corresponding training signal. The adaptive noise classifier may be further configured to process the first plurality of features of each training signal to generate the plurality of model parameters. In some embodiments, during the training phase, the adaptive noise classifier may be further configured to initialize the plurality of model parameters to a plurality of predefined values and update the plurality of model parameters in an iterative manner. For each iteration, the adaptive noise classifier may be further configured to generate output data based on the plurality of model parameters and the first plurality of features of each training signal and compare the output data with target data to determine an error associated with the plurality of model parameters. The plurality of model parameters may be updated based on the determined error. The training phase of the adaptive noise classifier is completed when the determined error is below a tolerance limit.
In some embodiments, during the deployment phase, the adaptive noise classifier may be further configured to filter the plurality of audio signals received from the plurality of audio-capturing devices and extract a second plurality of features from each audio signal. The second plurality of features may include at least one of a group consisting of a time period, a measure of periodicity, a spectral envelope, one or more cepstral attributes, and one or more spectral band energy levels associated with the corresponding audio signal. The first acoustic state may be determined based on the second plurality of features of each audio signal and the plurality of model parameters.
In some embodiments, during the deployment phase, the plurality of model parameters may be updated based on the one or more functional attributes of the vehicle.
In some embodiments, the controller may be further configured to generate, based on the first acoustic state and the one or more functional attributes of the vehicle, feedback data indicative of the update of the plurality of model parameters and provide the feedback data to the adaptive noise classifier. The adaptive noise classifier may be further configured to update the plurality of model parameters based on the feedback data.
In some embodiments, the first acoustic state may correspond to one of a group consisting of the vehicle moving with a plurality of windows thereof closed and the vehicle moving with one or more windows, of the plurality of windows, open. The second acoustic state may correspond to one of a group consisting of the vehicle moving at a reference speed with the plurality of windows closed and presence of a voice activity therein, the vehicle moving at the reference speed with the plurality of windows closed and absence of the voice activity therein, the vehicle moving at the reference speed with the one or more windows open and the presence of the voice activity therein, and the vehicle moving at the reference speed with the one or more windows open and the absence of the voice activity therein.
In some embodiments, the plurality of noise cancellation systems may include an engine noise cancellation system for an engine noise of the vehicle, a road noise cancellation system for a road-tire interaction noise of the vehicle, and an active noise cancellation system for one or more active noises in the vehicle that hamper the voice activity therein. Each of the engine, road, and active noise cancellation systems is controlled based on the second acoustic state.
In some embodiments, when the second acoustic state corresponds to the vehicle moving at the reference speed with the plurality of windows closed and the presence of the voice activity therein, the engine noise cancellation system is deactivated and the road and active noise cancellation systems are activated. When the second acoustic state corresponds to the vehicle moving at the reference speed with the plurality of windows closed and the absence of the voice activity therein, the engine and active noise cancellation systems are deactivated and the road noise cancellation system is activated. When the second acoustic state corresponds to the vehicle moving at the reference speed with the one or more windows open and the presence of the voice activity therein, the engine and active noise cancellation systems are deactivated and the road noise cancellation system is one of a group consisting of activated and deactivated. When the second acoustic state corresponds to the vehicle moving at the reference speed with the one or more windows open and the absence of the voice activity therein, the engine and active noise cancellation systems are deactivated and the road noise cancellation system is one of a group consisting of activated and deactivated.
Conventionally, to reduce noises inside a cabin of a vehicle, various noise cancellation systems are utilized. Examples of the noise cancellation systems may include an engine noise cancellation system for an engine noise, a road noise cancellation system for a road-tire interaction noise, and an active noise cancellation system for various active noises that hamper a voice activity in the vehicle. Such noise cancellation systems improve the experience of the driver and passengers of the vehicle by detecting the associated noise and generating an anti-noise that dampens the effect of the detected noise. However, the operations performed by each noise cancellation system are resource-intensive tasks and require significant computational power. Typically, each noise cancellation system is activated and deactivated in an independent manner. Consequently, there may be situations where all the noise cancellation systems in the vehicle are activated, thereby leading to consumption of significant power and resources (e.g., signal processors, math accelerators, or the like).
Various embodiments of the present disclosure disclose a circuit that includes an adaptive noise classifier and a controller for controlling various noise cancellation systems of a vehicle. The noise cancellation systems may include an engine noise cancellation system for an engine noise of the vehicle, a road noise cancellation system for a road-tire interaction noise of the vehicle, and an active noise cancellation system for active noises in the vehicle that hamper a voice activity therein. The adaptive noise classifier may receive audio signals from various audio-capturing devices of the vehicle and determine a first acoustic state of the vehicle based on the received audio signals. Further, the controller may determine a second acoustic state of the vehicle based on the first acoustic state and one or more functional attributes of the vehicle. The second acoustic state is an updated version of the first acoustic state. Each acoustic state may be indicative of a noise profile in the vehicle. For example, the first acoustic state may indicate that the vehicle is moving with windows closed or with windows open, whereas the second acoustic state may indicate that the vehicle is moving at a reference speed with windows closed or open and presence or absence of the voice activity therein. Based on the second acoustic state, the controller may control an operational state of each noise cancellation system of the vehicle.
In the vehicle, the engine noise is less than a reduced road-tire interaction noise (e.g., after the reduction implemented by the road noise cancellation system). Similarly, the wind noise is greater than the reduced road-tire interaction noise and a reduced active noise (e.g., after the reduction implemented by the active noise cancellation system). The controller may utilize the second acoustic state to efficiently control the noise cancellation systems. For example, when the second acoustic state indicates that the vehicle is moving with windows closed and the presence/absence of the voice activity therein, the effect of the engine noise cancellation system may be negated (e.g., may not be experienced in the cabin of the vehicle) as the reduced road-tire interaction noise may be greater. Hence, the controller may deactivate the engine noise cancellation system. Similarly, when the second acoustic state indicates that the vehicle is moving at the reference speed with windows open and the presence of the voice activity therein, the wind noise may negate the effect of all three noise cancellation systems. In such cases, the controller may deactivate all three noise cancellation systems.
Thus, in some embodiments of the present disclosure, the efficient control of the noise cancellation systems results in a significantly reduced utilization of vehicular resources as compared to some conventional techniques where noise cancellation systems are controlled in an independent manner. Additionally, the power consumption is significantly less in some embodiments of the present disclosure as compared to some conventional techniques.
The plurality of noise cancellation systems 102 may include an engine noise cancellation system 102a for the engine noise, a road noise cancellation system 102b for the road-tire interaction noise, and an active noise cancellation system 102c for the one or more active noises. In other words, the engine noise cancellation system 102a may reduce the engine noise of the vehicle 100, whereas, the road noise cancellation system 102b may reduce the road-tire interaction noise of the vehicle 100. Further, the active noise cancellation system 102c may reduce the one or more active noises in the vehicle 100 that hamper the voice activity therein. The engine noise and the road-tire interaction noise may also be classified as active noises in the vehicle 100, however, the active noises that hamper the voice activity are different from the engine noise and the road-tire interaction noise.
Each noise cancellation system of the plurality of noise cancellation systems 102 may be configured to execute various operations to reduce the effect of an associated noise. For example, the engine noise cancellation system 102a may be configured to detect the engine noise and generate an engine anti-noise based on the detected engine noise. The engine anti-noise may be complementary to the detected engine noise. For example, the engine anti-noise is a phase-reversed version of the detected engine noise. The engine noise cancellation system 102a may be further configured to drive a speaker (shown later in
Although it is described that the anti-noise generated by each noise cancellation system is played on the same speaker, the scope of the present disclosure is not limited to it. In other embodiments, different speakers (not shown) of the vehicle 100 may be utilized for playing one or more anti-noises.
The operations performed by each noise cancellation system are resource-intensive tasks and require significant computational power. If each of the plurality of noise cancellation systems 102 are simultaneously activated, the strain on vehicular resources (e.g., signal processors, math accelerators, or the like) may be significant and the overall power consumption may shoot up. Hence, efficient management of the plurality of noise cancellation systems 102 is paramount. The vehicle 100 may thus include a management system 104 to efficiently manage the plurality of noise cancellation systems 102.
The management system 104 may be coupled to the plurality of noise cancellation systems 102. The management system 104 may be configured to determine a noise profile of the vehicle 100 based on various audios present in the vehicle 100 and one or more functional attributes of the vehicle 100. The one or more functional attributes of the vehicle 100 may include a speed of the vehicle 100, a position of a plurality of windows of the vehicle 100, a status of a voice detector of the vehicle 100, a tire pressure of a plurality of tires of the vehicle 100, a revolution count of an engine of the vehicle 100, a status of a heating, ventilation, and air conditioning (HVAC) system of the vehicle 100, or a combination thereof. The aforementioned functional attributes are exemplary and may be different in other embodiments. Various components (e.g., the plurality of windows, the voice detector, the plurality of tires, the engine, and the HVAC system) of the vehicle 100 and the associated positioning are shown later in
The adaptive noise classifier 106 may include suitable circuitry that may be configured to perform one or more operations. For example, the adaptive noise classifier 106 may be configured to determine a first acoustic state of the vehicle 100. The first acoustic state may be indicative of a dominant noise associated with the vehicle 100. The dominant noise may correspond to a noise of the engine (e.g., the engine noise) of the vehicle 100, a noise associated with an interaction of the plurality of tires of the vehicle 100 with a road (shown later in
The training of the adaptive noise classifier 106 may be implemented in an offline environment (e.g., not when the management system 104 is in-field for its designated application). During the training phase, the adaptive noise classifier 106 may be configured to receive the plurality of training signals from a training circuit (not shown). The plurality of training signals correspond to an offline dataset that resembles real-world (e.g., in-field) use cases utilized to train the adaptive noise classifier 106. Thus, each training signal may be indicative of an audio associated with the vehicle 100.
The adaptive noise classifier 106 may be further configured to extract a first plurality of features from each training signal. The first plurality of features may include a time period, a measure of periodicity, a spectral envelope, one or more cepstral attributes, and one or more spectral band energy levels associated with the corresponding training signal. The aforementioned features are exemplary and may be different in other embodiments. Each feature may correspond to a numerical, logical, or Boolean representation of information extracted from the training signal. Additionally, each feature may be a scalar quantity or a vector quantity. For example, the time period and the measure of periodicity may be scalar quantities, whereas, the spectral envelope, the one or more cepstral attributes, and the one or more spectral band energy levels may be vector quantities. Further, in an embodiment, one set of features may be extracted based on the analysis of a predefined portion of each training signal and may be arranged in the form of a feature vector. Thus, the first plurality of features extracted for each training signal may include multiple feature vectors. The adaptive noise classifier 106 may be further configured to process the first plurality of features of each training signal to generate the plurality of model parameters. In other words, the first plurality of features extracted from each training signal are modeled (e.g., are generalized for inference).
During the training phase, the adaptive noise classifier 106 may be further configured to update the plurality of model parameters in an iterative manner. In some embodiments of the present disclosure, the plurality of model parameters are updated by way of supervised learning. For example, initially, the adaptive noise classifier 106 may be configured to initialize the plurality of model parameters to a plurality of predefined values. Further, for each iteration, the adaptive noise classifier 106 may be configured to generate output data (not shown) based on the plurality of model parameters and the first plurality of features of each training signal. The adaptive noise classifier 106 may be further configured to compare the output data with target data (not shown) to determine an error associated with the adaptive noise classifier 106 (e.g., the plurality of model parameters). The target data may correspond to a desired output of supervised learning. The plurality of model parameters for each iteration are updated based on the corresponding determined error (e.g., the determined error at the end of the corresponding iteration). In an embodiment, the adaptive noise classifier 106 may be further configured to determine a cost function from the determined error, and the plurality of model parameters may be updated based on the determined cost function. The cost function may correspond to a squared error, a squared error averaged over a predefined number of frames, or the like. The update of the plurality of model parameters may be executed iteratively until the error is below a tolerance limit. The training phase of the adaptive noise classifier 106 is thus completed when the determined error is below the tolerance limit.
The trained adaptive noise classifier 106 may be utilized in-field to determine the noise profile of the vehicle 100. The adaptive noise classifier 106 may be configured to receive various audio signals (e.g., first through sixth audio signals AS1-AS6) indicative of various audios (e.g., noises) associated with the vehicle 100. The vehicle 100 may further include first through sixth audio-capturing devices 110a-110f that may be configured to generate the first through sixth audio signals AS1-AS6, respectively. The adaptive noise classifier 106 may thus be coupled to the first through sixth audio-capturing devices 110a-110f, and receive the first through sixth audio signals AS1-AS6 from the first through sixth audio-capturing devices 110a-110f, respectively. In an embodiment, each of the first through sixth audio-capturing devices 110a-110f corresponds to a microphone. The first through sixth audio signals AS1-AS6 may be collectively referred to as a “plurality of audio signals AS1-AS6”, and the first through sixth audio-capturing devices 110a-110f may be collectively referred to as a “plurality of audio-capturing devices 110”. The plurality of audio-capturing devices 110 may be located at different parts of the vehicle 100. For example, the first through fourth audio-capturing devices 110a-110d may be located near the plurality of tires of the vehicle 100, the fifth audio-capturing device 110e may be located near the engine of the vehicle 100, and the sixth audio-capturing device 110f may be located in the cabin of the vehicle 100.
Six audio-capturing devices are illustrated in
The adaptive noise classifier 106 may be further configured to filter the plurality of audio signals AS1-AS6 and extract a second plurality of features from each audio signal of the plurality of audio signals AS1-AS6. The second plurality of features may be similar to the first plurality of features. For example, the second plurality of features may include a time period, a measure of periodicity, a spectral envelope, one or more cepstral attributes, and one or more spectral band energy levels associated with the corresponding audio signal. The first acoustic state of the vehicle 100 is determined based on the second plurality of features of cach audio signal of the plurality of audio signals AS1-AS6 and the plurality of model parameters.
The adaptive noise classifier 106 may be further configured to generate classification data CLF indicative of the first acoustic state and provide the classification data CLF to the controller 108. The first acoustic state is indicative of a partial noise profile of the vehicle 100. The one or more functional attributes of the vehicle 100 may be utilized by the controller 108 to validate the first acoustic state as well as determine a refined (e.g., a complete) noise profile of the vehicle 100.
The controller 108 may include suitable circuitry that may be configured to perform one or more operations. For example, the controller 108 may be configured to receive the classification data CLF from the adaptive noise classifier 106. Further, the controller 108 may be configured to receive status data SD indicative of the one or more functional attributes of the vehicle 100. The vehicle 100 may further include a sensor circuit 112 that may be configured to generate the status data SD. The controller 108 may thus be coupled to the sensor circuit 112, and receive the status data SD from the sensor circuit 112.
The status data SD may include a first set of values indicative of the speed of the vehicle 100, a second set of values indicative of the position of the plurality of windows of the vehicle 100, and a third set of values indicative of the status of the voice detector of the vehicle 100. Additionally, the status data SD may include a fourth set of values indicative of the tire pressure of the plurality of tires of the vehicle 100, a fifth set of values indicative of the revolution count of the engine of the vehicle 100, and a sixth set of values indicative of the status of the HVAC system of the vehicle 100. Each of the first through sixth sets of values may be generated by a set of sensors (not shown) of the vehicle 100. For example, a set of speed sensors may generate the first set of values, a set of position sensors may generate the second set of values, a set of voice sensors may generate the third set of values, a set of pressure sensors may generate the fourth set of values, a set of revolution counters may generate the fifth set of values, and a set of HVAC sensors may generate the sixth set of values. The sensor circuit 112 may be configured to receive the output of each set of sensors, generate the status data SD, and provide the status data SD to the controller 108.
Based on the first acoustic state of the vehicle 100 and the status data SD, the controller 108 may be further configured to determine a second acoustic state of the vehicle 100. In other words, the controller 108 determines the second acoustic state based on the classification data CLF and the one or more functional attributes of the vehicle 100. The second acoustic state is an updated version of the first acoustic state.
The controller 108 may be further configured to validate the first acoustic state based on the status data SD. As the first acoustic state is indicative of the dominant noise in the vehicle 100, the controller 108 may utilize the one or more functional attributes of the vehicle 100 to validate the first acoustic state and determine the second acoustic state exclusively after successful validation of the first acoustic state. Thus, the controller 108 determines the second acoustic state further based on the validation of the first acoustic state. Although it is described that the first acoustic state is validated prior to the determination of the second acoustic state, the scope of the present disclosure is not limited to it. In other embodiments, the validation of the first acoustic state may not be performed.
The second acoustic state thus corresponds to the refined (e.g., the complete) noise profile of the vehicle 100. In one example, the first acoustic state corresponds to the vehicle 100 moving with the plurality of windows thereof closed or open. The vehicle 100 moving with the plurality of windows thereof open may indicate that the wind noise is the dominant noise in the vehicle 100. Similarly, the vehicle 100 moving with the plurality of windows thereof closed may indicate that the road-tire interaction noise is the dominant noise in the vehicle 100. In such a scenario, the controller 108 may validate and update the first acoustic state based on the status data SD. The second acoustic state may correspond to the vehicle 100 moving at a reference speed with the plurality of windows closed or open and presence or absence of the voice activity therein. The reference speed may correspond to a speed greater than 20 kilometers/hour. However, in other embodiments, the reference speed may have other values. The position of the plurality of windows and the speed of the vehicle 100 may be utilized to validate the first acoustic state. Further, the speed of the vehicle 100, the position of the plurality of windows, and the status of the voice detector may be utilized for determining the second acoustic state.
The scope of the present disclosure is not limited to the aforementioned exemplary acoustic states. In other embodiments, different acoustic states, indicative of the corresponding noise profiles, may be determined, without deviating from the scope of the present disclosure. In such cases, some other functional attributes of the vehicle 100 (such as the tire pressure of the plurality of tires, the revolution count of the engine, the status of the HVAC system, or the like) may be utilized to validate and update the first acoustic state.
The controller 108 may be further configured to control the operational state of each of the plurality of noise cancellation systems 102 based on the second acoustic state of the vehicle 100. To control the operational state of each of the plurality of noise cancellation systems 102, the controller 108 may be further configured to generate various control signals (e.g., first through third control signals CS1-CS3) and provide the first through third control signals CS1-CS3 to the engine, road, and active noise cancellation systems 102a, 102b, and 102c, respectively. The first through third control signals CS1-CS3 are collectively referred to as a “plurality of control signals CS1-CS3”. Each control signal controls activation and deactivation of a corresponding noise cancellation system. In an embodiment, a logic low state of a control signal deactivates the corresponding noise cancellation system and a logic high state of the control signal activates the corresponding noise cancellation system. The control of the plurality of noise cancellation systems 102 based on the second acoustic state results in the efficient management of the plurality of noise cancellation systems 102. The control of the plurality of noise cancellation systems 102 is explained in detail in conjunction with
The controller 108 may be further configured to generate, based on the first acoustic state and the one or more functional attributes of the vehicle 100, feedback data FB indicative of the update of the plurality of model parameters. The feedback data FB may include label information utilized for updating the plurality of model parameters. Further, the controller 108 may be configured to provide the feedback data FB to the adaptive noise classifier 106. The adaptive noise classifier 106 may be further configured to update the plurality of model parameters based on the feedback data FB (e.g., the one or more functional attributes of the vehicle 100). The plurality of model parameters are thus updated even during the deployment phase of the adaptive noise classifier 106. As a result, the number and variety of the plurality of training signals required for training the adaptive noise classifier 106 in the offline environment are reduced. In an embodiment, the controller 108 is implemented as a processor core executing code. However, the controller 108 may be implemented in different ways including with different types of circuitry in other embodiments.
The scope of the present disclosure is not limited to the controller 108 enabling the model parameter update. In other embodiments, any other circuitry of the vehicle 100 may be configured to generate the feedback data FB, without deviating from the scope of the present disclosure.
The positions of various components of the vehicle 100 illustrated in
In a first example, the second acoustic state corresponds to the vehicle 100 moving at the reference speed with the plurality of windows closed and the presence of the voice activity therein. The movement of the vehicle 100 at the reference speed indicates the presence of the engine noise and the road-tire interaction noise. Further, as the voice activity is present in the vehicle 100, reduction of various active noises is necessitated to enhance the voice activity. In some conventional techniques, each of the engine, road, and active noise cancellation systems 102a, 102b, and 102c is controlled independently, and hence, the above-mentioned example of the second acoustic state may result in the activation of all three noise cancellation systems. In the present disclosure, however, the noise profile of the vehicle 100 is utilized to control the engine, road, and active noise cancellation systems 102a, 102b, and 102c in an efficient manner. For example, the road-tire interaction noise is greater than the engine noise by 20-30 decibels (dB), and the road noise cancellation system 102b may achieve a reduction of 3-5 dB. Thus, the engine noise is still less than the reduced road-tire interaction noise. In such cases, the effect of the engine noise cancellation system 102a may be negated (e.g., not be experienced in the cabin 306 of the vehicle 100) by the reduced road-tire interaction noise. Hence, in the present disclosure, when the second acoustic state corresponds to the vehicle 100 moving at the reference speed with the plurality of windows closed and the presence of the voice activity therein, the engine noise cancellation system 102a is deactivated and exclusively the road and active noise cancellation systems 102b and 102c are activated. In other words, the first control signal CS1 is at a logic low state and the second and third control signals CS2 and CS3 are at a logic high state. If it is assumed that the computational power required by each of the engine, road, and active noise cancellation systems 102a, 102b, and 102c is ‘X’ units, the computational power consumption in the present disclosure is ‘2X’ units, whereas, that in some conventional techniques is ‘3X’ units. The efficient control of the plurality of noise cancellation systems 102 thus achieves a power reduction of ‘X’ units (e.g., 33.3%).
In a second example, the second acoustic state corresponds to the vehicle 100 moving at the reference speed with the plurality of windows closed and the absence of the voice activity therein. In some conventional techniques, the engine and road noise cancellation systems 102a and 102b may be activated and the active noise cancellation system 102c may be deactivated. In the present disclosure, however, the engine and active noise cancellation systems 102a and 102c are deactivated, and exclusively the road noise cancellation system 102b is activated for the same reason as described above in the first example. Thus, the computational power consumption in the present disclosure is ‘X’ units, whereas, that in some conventional techniques is ‘2X’ units. The efficient control of the plurality of noise cancellation systems 102 thus achieves a power reduction of ‘X’ units (e.g., 50%).
In a third example, the second acoustic state corresponds to the vehicle 100 moving at the reference speed with the one or more windows open and the presence of the voice activity therein. The movement of the vehicle 100 at the reference speed with the one or more windows open indicates the presence of the engine noise, the road-tire interaction noise, and the wind noise. In some conventional techniques, each of the engine, road, and active noise cancellation systems 102a, 102b, and 102c may be activated. In the present disclosure, however, the noise profile of the vehicle 100 is utilized to control the engine, road, and active noise cancellation systems 102a, 102b, and 102c in an efficient manner. For example, the wind noise is greater than the road-tire interaction noise, the one or more active noises, and the engine noise by 5-6 dB, 10-15 dB, and 10-12 dB, respectively, and each noise cancellation system may achieve a reduction of 3-5 dB. Thus, the effect of each of the engine, road, and active noise cancellation systems 102a, 102b, and 102c may be negated by the wind noise.
Hence, in the present disclosure, when the second acoustic state corresponds to the vehicle 100 moving at the reference speed with the one or more windows open and the presence of the voice activity therein, the engine, road, and active noise cancellation systems 102a, 102b, and 102c are deactivated. Thus, the computational power consumption in the present disclosure is null, whereas, that in some conventional techniques is ‘3X’ units. The efficient control of the plurality of noise cancellation systems 102 thus achieves a power reduction of ‘3X’ units (e.g., 100%).
In a fourth example, the second acoustic state corresponds to the vehicle 100 moving at the reference speed with the one or more windows open and the absence of the voice activity therein. In some conventional techniques, the engine and road noise cancellation systems 102a and 102b may be activated and the active noise cancellation system 102c may be deactivated. In the present disclosure, however, each of the engine, road, and active noise cancellation systems 102a, 102b, and 102c is deactivated for the same reason as described in the third example. Thus, the computational power consumption in the present disclosure is null, whereas, that in some conventional techniques is ‘2X’ units. The efficient management of the plurality of noise cancellation systems 102 thus achieves a power reduction of ‘2X’ units (e.g., 100%).
Although for the third and fourth examples it is described that the road noise cancellation system 102b is deactivated, the scope of the present disclosure is not limited to it. In other embodiments, the road noise cancellation system 102b may be activated for the third and fourth examples. In an exemplary embodiment, whether to activate or deactivate the road noise cancellation system 102b for the third and fourth examples may be determined based on the value of the reference speed. For example, if the reference speed corresponds to speeds of 50-60 kilometers/hour, the wind noise may be significant, and hence, the road noise cancellation system 102b may be deactivated. Conversely, if the reference speed corresponds to speeds of 20-25 kilometers/hour, the difference between the wind noise and the road-tire interaction noise may not be significant, and hence, the road noise cancellation system 102b may be activated to reduce the effect of the road-tire interaction noise.
In each of the above-mentioned examples, in addition to the reduced computational power consumption, the utilization of vehicular resources as well as the power required to drive the speaker 308 are also significantly reduced as compared to some conventional techniques. The available vehicular resources may be utilized for executing various other critical operations associated with the vehicle 100.
The scope of the present disclosure is not limited to examples of the second acoustic state described in
Thus, the overall noise profile of the vehicle 100 is utilized to control the plurality of noise cancellation systems 102 in a co-dependent manner. The control technique implemented in the present disclosure is thus more efficient than some conventional techniques that control the noise cancellation systems in an independent manner.
At step 502, the adaptive noise classifier 106 may receive the plurality of training signals. At step 504, the adaptive noise classifier 106 may extract the first plurality of features from each training signal. At step 506, the adaptive noise classifier 106 may process the first plurality of features of each training signal. At step 508, the adaptive noise classifier 106 may generate the plurality of model parameters. The plurality of model parameters may be generated based on the processing of the first plurality of features of each training signal.
The generation of the plurality of model parameters is described by way of steps 508a-508c. At step 508a, the adaptive noise classifier 106 may initialize the plurality of model parameters to the plurality of predefined values. At step 508b, the adaptive noise classifier 106 may generate the output data based on the plurality of model parameters and the first plurality of features of each training signal. At step 508c, the adaptive noise classifier 106 may compare the output data with the target data to determine the error associated with the plurality of model parameters. At step 508d, the adaptive noise classifier 106 may determine whether the error is below the tolerance limit. If at step 508d, it is determined that the error is above the tolerance limit, step 508e is performed. At step 508e, the adaptive noise classifier 106 may update the plurality of model parameters based on the determined error. Step 508b may be performed after step 508e. If at step 508d, it is determined that the error is below the tolerance limit, the training of the adaptive noise classifier 106 is completed. The plurality of model parameters of the last iteration are utilized in the deployment phase for generating the first acoustic state.
Although it is described that the adaptive noise classifier 106 is trained in the offline environment, the scope of the present disclosure is not limited to it. In other embodiments, the adaptive noise classifier 106 may be trained in-field, without deviating from the scope of the present disclosure.
Referring to
Referring to
Thus, in some embodiments of the present disclosure, the efficient control of the plurality of noise cancellation systems 102 results in a significantly reduced utilization of the vehicular resources as compared to some conventional techniques where noise cancellation systems are controlled in an independent manner. The available vehicular resources may be utilized for executing various other critical operations associated with the vehicle 100. Additionally, the power consumption is significantly less in some embodiments of the present disclosure as compared to some conventional techniques.
While various embodiments of the present disclosure have been illustrated and described, it will be clear that the present disclosure is not limited to these embodiments only. Numerous modifications, changes, variations, substitutions, and equivalents will be apparent to those skilled in the art, without departing from the spirit and scope of the present disclosure, as described in the claims. Further, unless stated otherwise, terms such as “first” and “second” are used to arbitrarily distinguish between the elements such terms describe. Thus, these terms are not necessarily intended to indicate temporal or other prioritization of such elements.
| Number | Date | Country | Kind |
|---|---|---|---|
| 202221072339 | Dec 2022 | IN | national |
