Vehicle vibration transmission using machine learning

Abstract

The present disclosure provides "Vehicle Vibration Transfer with Machine Learning". A first vehicle includes a coupled dynamic system including a first component and a second component. A computer is programmed to receive first vibration data detected by a first vibration sensor mounted to the first component; implement a machine learning model to determine an amount of vibration energy transferred to the second component; and actuate a third component of the first vehicle based on the amount of vibration energy. The first vibration sensor is of a first type. The second component of the first vehicle has no vibration sensor mounted thereto. The first vibration data is an input to the machine learning model. The machine learning model is trained to match ground truth vibration data generated by a second vibration sensor mounted to the second component of a second vehicle. The second vibration sensor is of a second type different from the first type.

Description

Technical Field

[0001] This disclosure relates to actuating a vehicle based on the vibrational energy of components of a dynamic system that tracks the connection transmitted to the vehicle. Background Technology

[0002] When in motion, a vehicle experiences vibrations from the road. These vibrations can have numerous effects on the vehicle, such as real-time effects on handling and long-term effects on component lifespan. Summary of the Invention

[0003] The system is "connected" because it comprises multiple components linked together, and is "dynamic" because the state variables of the components change over time relative to each other or to the rest of the vehicle or a fixed frame of reference. The connected dynamic system can be, for example, a suspension system for the vehicle's wheels or another underbody system experiencing road vibrations. A computer is programmed to receive first vibration data detected by a first vibration sensor mounted to a first component of the connected dynamic system and input this first vibration data into a machine learning model. The machine learning model then outputs the amount of vibration energy (or multiple amounts from multiple second components) transmitted to a second component of the connected dynamic system. Vibration sensors may be missing on the second components of the vehicle. Finally, the computer actuates a third component of the vehicle based on the amount of vibration energy. The machine learning model is trained to match ground-based vibration data generated by second vibration sensors mounted to the second components in the test vehicle. Training the machine learning model allows for vibration energy estimation of components without vibration sensors by modeling the transmission of vibration energy through the connected dynamic system. This disclosure provides techniques for modeling energy transfer in dynamic systems with multiple interconnected components using a combination of data-driven techniques (including machine learning models) and physics-based modeling (e.g., modal analysis, coherence, etc.). A second vibration sensor used to collect ground-based data in a test vehicle is a different type from a first vibration sensor mounted on a first component of the vehicle. This difference in type allows the machine learning model to mimic a more accurate sensor type.

[0004] A computer includes a processor and a memory, the memory storing instructions executable by the processor to: receive first vibration data detected by a first vibration sensor; implement a machine learning model to determine the amount of vibration energy transmitted to a second component; and actuate a third component of a first vehicle based on the amount of vibration energy. The first vibration sensor is of a first type. The first vibration sensor is mounted to a first component of the first vehicle. The first vehicle includes a coupled dynamic system comprising the first component and the second component. The second component of the first vehicle does not have any vibration sensors mounted thereto. The first vibration data is input to the machine learning model. The machine learning model is trained to match ground vibration data generated by a second vibration sensor. The second vibration sensor is mounted to a second component of a second vehicle. The second vibration sensor is of a second type, different from the first type.

[0005] In the example, the first type could be a road noise cancellation (RNC) accelerometer.

[0006] In the example, the second type could be an internal circuit piezoelectric (ICP) sensor.

[0007] In one example, the connected dynamic system may include a plurality of second components, each of which has no vibration sensor mounted thereon, and the machine learning model may be trained to match ground vibration data generated by the plurality of second vibration sensors respectively mounted to the second components. In another example, the machine learning model may be a one-to-many regression model.

[0008] In another example, each second vibration sensor can be of a second type.

[0009] In one example, the dynamic system of the connection could be the suspension system of the wheels of the first vehicle. In another example, the first component could be a control arm.

[0010] In one example, ground vibration data can be processed by applying a bandpass filter before training a machine learning model. In another example, the bandpass filter's frequency band can be based on coherence analysis of the ground vibration data.

[0011] In the example, the instructions may also include instructions to actuate a third component in response to the amount of vibrational energy exceeding a threshold.

[0012] In one example, the instruction for receiving the first vibration data may include instructions for receiving the first vibration data from a buffer on the first vehicle. In another example, the instruction may also include instructions for storing the first vibration data in the buffer in response to a triggering condition being met. In yet another example, the triggering condition may be based on the speed of the first vehicle.

[0013] A method includes: training a machine learning model to match ground vibration data generated by a second vibration sensor; receiving first vibration data detected by a first vibration sensor; implementing the machine learning model to determine the amount of vibration energy transmitted to a second component; and actuating a third component of a first vehicle based on the amount of vibration energy. The second vibration sensor is mounted to the second component of the second vehicle. The second vibration sensor is of a second type. The first vibration sensor is of a first type, different from the second type. The first vibration sensor is mounted to a first component of the first vehicle. The first vehicle includes a coupled dynamic system comprising the first component and the second component. The second component of the first vehicle does not have any vibration sensors mounted thereon. The first vibration data is the input to the machine learning model.

[0014] In the example, the first type could be a road noise cancellation (RNC) accelerometer, and the second type could be an internal circuit piezoelectric (ICP) sensor.

[0015] In one example, the connected dynamic system may include a plurality of second components, each of which has no vibration sensor mounted thereon; the machine learning model may be trained to match ground vibration data generated by a plurality of second vibration sensors respectively mounted to the second components; and each second vibration sensor may be of a second type. In another example, the machine learning model may be a one-to-many regression model.

[0016] In the example, the dynamic system of the connection may be the suspension system of the wheels of the first vehicle, and the first component may be a control arm.

[0017] In the example, the method may further include processing the ground vibration data by applying a bandpass filter before training the machine learning model, and the frequency band of the bandpass filter may be based on the coherence analysis of the ground vibration data. Attached Figure Description

[0018] Figure 1 This is a perspective view of the example mechanical system of the example vehicle.

[0019] Figure 2 This is a block diagram of an example vehicle control system.

[0020] Figure 3 This is a flowchart of an example process used to train a machine learning model to track the vibrational energy transfer to vehicle components.

[0021] Figure 4 This is a flowchart of an example process for tracking the transfer of vibrational energy to components of a vehicle. Detailed Implementation

[0022] Referring to the accompanying drawings, where the same numerals denote the same components in several views, a first vehicle 100a includes a coupled dynamic system 115, which includes a first component 120 and a second component 125. A first vibration sensor 105 is mounted to the first component 120 of the first vehicle 100a. No vibration sensor is mounted to the second component 125 of the first vehicle 100a. A computer 205, 210 includes a processor and a memory, and the memory stores instructions executable by the processor to: receive first vibration data detected by the first vibration sensor 105; implement a machine learning model to determine the amount of vibration energy transmitted to the second component 125; and actuate a third component of the first vehicle 100a based on the amount of vibration energy. The first vibration sensor 105 is of a first type. The first vibration data is the input to the machine learning model. The machine learning model is trained to match ground vibration data generated by a second vibration sensor 110. The second vibration sensor 110 is mounted to the second component 125 of a second vehicle 100b. The second vibration sensor 110 is of a second type, different from the first type.

[0023] Figure 1 A portion of a second vehicle 100b is shown. The second vehicle 100b can be any passenger or commercial vehicle, such as a sedan, truck, SUV, crossover, van, minivan, taxi, bus, etc. The second vehicle 100b includes a coupled dynamic system 115, a first vibration sensor 105, and a second vibration sensor 110. The second vehicle 100b (or multiple second vehicles 100b) is used to collect data for training a machine learning model, as described below.

[0024] The first vehicle 100a is a vehicle implementing a trained machine learning model. The first vehicle 100a can be any passenger or commercial vehicle, such as a sedan, truck, SUV, crossover, van, minivan, taxi, bus, etc. The first vehicle 100a can be the same model as the second vehicle 100b. Except that the first vehicle 100a does not have the second vibration sensor 110, the first vehicle 100a can include the same components as the first vehicle 100a. Except for the second vibration sensor 110, the following description of components applies to both the first vehicle 100a and the second vehicle 100b. The reference to "vehicle" applies to both the first vehicle 100a and the second vehicle 100b.

[0025] The coupled dynamic system 115 includes a first component 120 and one or more second components 125. The components 120, 125 of the coupled dynamic system 115 are mechanically coupled together, and at least some of the components 120, 125 are movable relative to the other components 120, 125, thereby making the coupled dynamic system 115 dynamic. Each pair of components 120, 125 may be mechanically coupled together directly or indirectly via other components 120, 125 of the coupled dynamic system 115. The coupled dynamic system 115 may be a mechanical system of a vehicle 100. Vibrations from sources such as the road over which the vehicle 100 travels may be transmitted between the components 120, 125 of the coupled dynamic system 115. The coupled dynamic system 115 may be part of the underbody of the vehicle 100, which experiences a relatively large amount of road vibration. The components 120, 125 may be interconnected in different ways and at different points, and may have different shapes, thereby providing multiple paths and resonant frequencies through which input vibrations can travel. How vibrations travel through the connected dynamic system 115 can be a complex nonlinear system.

[0026] For example, the connected dynamic system 115 may be the suspension system of the wheels of the vehicle 100. Components 120 and 125 of the suspension system may include wheel hubs, wheel bearings, upper control arms, lower control arms, steering gears, ball joints, anti-roll bars, tie rods, shock absorbers, spring pads, electric motors, motor mounts, etc. The first component 120 may be a control arm that connects the wheel assembly to the frame of the vehicle 100 and thereby transmits road vibrations to the frame. The second component 125 may be the remaining components of the suspension system.

[0027] The first vibration sensor 105 is of a first type, and the second vibration sensor 110 is of a second type, different from the first type. In terms of the generated data, the second type may have greater resolution, lower noise, a faster refresh rate, or some other advantage compared to the first type. In terms of installation on production vehicles, the first type may have wider availability, greater durability, or some other advantage compared to the second type. Therefore, the first type may be a road noise cancellation (RNC) accelerometer, and the second type may be an internal circuit piezoelectric (ICP) sensor.

[0028] A first vibration sensor 105 is mounted to a first component 120 of vehicle 100, and a second vibration sensor 110 is mounted to a second component 125 of second vehicle 100b (and possibly to the first component 120). The term "first component" herein refers to the component on which the first vibration sensor 105 is mounted, and the term "second component" herein refers to a component of dynamic system 115 that is not connected to the first vibration sensor 105. The first vibration sensor 105 may be placed on a component (such as one of the control arms) experiencing relatively high or undispersed amounts of road vibration to collect an accurate assessment of the input road vibration. The second vibration sensor 110 may be placed on components 120, 125 to collect vibration energy transmitted to these components 120, 125. Vehicle 100 may include multiple first vibration sensors 105, such as one first vibration sensor 105 for each associated dynamic system 115 in vehicle 100.

[0029] On the first vehicle 100a, a first vibration sensor 105 is mounted to a first component 120, and none of the second components 125 have any vibration sensors mounted thereto. The first vehicle 100a may not have any second vibration sensor 110 (i.e., any vibration sensor of the second type). As described below, a machine learning model can be trained to determine the amount of vibration energy transmitted to the second component 125 based on input from the first vibration sensor on the first component 120. Therefore, even without a vibration sensor on the second component 125, computers 205, 210 can track the vibration energy applied to the second component 125, thereby simplifying the construction of the first vehicle 100a.

[0030] refer to Figure 2 The first vehicle 100a includes a vehicle computer 205, a communication network 215, a first vibration sensor 105, a user interface 220, and a transceiver 225. The transceiver 225 can communicate with a remote computer 210. The term "computer" can refer to either the vehicle computer 205 or the remote computer 210.

[0031] Vehicle computer 205 is a microprocessor-based computing device, such as a general-purpose computing device (including a processor and memory, electronic controller, etc.), a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), or a combination thereof. Typically, hardware description languages ​​such as VHDL (VHSIC (Very High Speed ​​Integrated Circuit) hardware description language) are used in electronic design to describe digital and mixed-signal systems such as FPGAs and ASICs. For example, an ASIC is manufactured based on VHDL programming provided before manufacturing, while the logic components inside an FPGA can be configured based on VHDL programming (e.g., stored in memory electrically connected to the FPGA circuitry). Therefore, vehicle computer 205 may include a processor, memory, etc. The memory of vehicle computer 205 may include media for storing instructions executable by the processor and for electronically storing data and / or databases, and / or vehicle computer 205 may include structures such as those providing programming. Vehicle computer 205 may be multiple computers connected together on the first vehicle 100a.

[0032] Vehicle computer 205 can transmit and receive data via communication network 215. Communication network 215 can be a controller area network (CAN) bus, Ethernet, WiFi, local area network (LIN), on-board diagnostic connector (OBD-II), and / or any other wired or wireless communication network. Vehicle computer 205 can be communicatively connected to first vibration sensor 105, user interface 220, transceiver 225, and other electronic components via communication network 215.

[0033] User interface 220 presents and receives information to and from the operator of vehicle 100. User interface 220 may be located on a dashboard in the passenger compartment of the first vehicle 100a, and / or anywhere easily visible to the operator. User interface 220 may include dials, digital readout devices, screens, speakers, etc., for providing information to the operator, such as known human-machine interface (HMI) elements. User interface 220 may include buttons, knobs, keypads, microphones, etc., for receiving information from the operator.

[0034] Transceiver 225 can be adapted to wirelessly transmit signals via any suitable wireless communication protocol, such as cellular, Bluetooth®, Bluetooth® Low Energy (BLE), Ultra Wideband (UWB), WiFi, IEEE 802.11a / b / g / p, Cellular-V2X (CV2X), Dedicated Short Range Communication (DSRC), other RF (radio frequency) communications, etc. Transceiver 225 can be adapted to communicate with a remote server (i.e., a server different from and spaced apart from vehicle 100). The remote server can be located outside vehicle 100. For example, the remote server can be associated with another vehicle (e.g., V2V communication), with infrastructure components (e.g., V2I communication), with a first responder, with a mobile device associated with the operator of vehicle 100, etc. Transceiver 225 can be a single device or may include separate transmitters and receivers.

[0035] The remote computer 210 is a microprocessor-based computing device, such as a general-purpose computing device including a processor and memory. The memory of the remote computer 210 may include media for storing instructions executable by the processor and for electronically storing data and / or databases, and / or the remote computer 210 may include structures such as those providing programming capabilities. The remote computer 210 may be multiple computers connected together. The remote computer 210 is different from and spaced apart from the vehicle 100.

[0036] refer to Figure 1 The machine learning model is trained to match ground-based vibration data generated by one or more second vibration sensors 110 respectively mounted to the second component 125. As a general overview, one or more second vehicles 100b are subjected to a set of test conditions during which vibration data is received from the first vibration sensor 105 and the second vibration sensor 110. The vibration data from the first vibration sensor 105 is the input to the machine learning model, and the vibration data from the second vibration sensor 110 is the ground-based vibration data for the machine learning model. Before training the machine learning model, the ground-based vibration data is processed in one or more processing steps. The vibration data is preprocessed. Coherence analysis is performed on the preprocessed data to generate local modes for each second component 125. A bandpass filter is applied to the vibration data from each second vibration sensor 110 to isolate the corresponding local modes. The corresponding first vibration data from the first vibration sensor 105 is then used as input to train the machine learning model to match the isolated local vibration modes of each second component 125 (or statistical features extracted from the vibration data bandpassed by the local modes). After training is complete, the machine learning model is validated and then installed on computers 205 and 210 for use with the first vehicle 100a.

[0037] The second vehicle 100b can be placed in a test setup (such as a dynamometer) for simulating road travel. A second vibration sensor 110 can be attached to the second vehicle 100b and remain on it for the duration of the test. The second vehicle 100b is subjected to various test conditions. For example, the second vehicle 100b can run at each of a set of constant speeds for a preset duration. The constant speed can be, for example, an increment of 5 miles per hour (mph) from 10 mph to 60 mph. During each test run, first vibration data and raw second vibration data are received and recorded from the first vibration sensor 105 and the second vibration sensor 110, respectively. To increase the dataset, the test conditions can be repeated for multiple second vehicles 100b.

[0038] Each of the first vibration sensor 105 and the second vibration sensor 110 can generate one or more acceleration dimensions at regular time intervals. For example, vibration data from each vibration sensor 105, 110 can be three linear dimensions of acceleration at each time interval, such as (a) at t1. x1 ,a y1 , a z1 ), at t2 (a x2 , a y2 , a z2 Vibration data streams from the first vibration sensor 105 and the second vibration sensor 110 can be synchronized in time (e.g., data points generated at the same time intervals t1, t2, etc.).

[0039] The raw second vibration data is preprocessed through one or more preprocessing steps. For example, downsampling and / or low-pass filtering (e.g., recursive low-pass filtering) are performed to remove noise from the vibration data. The removed frequencies can be selected as known frequencies of noise in the road vibration data. A bandpass filter can be applied to remove the DC bias. A finite impulse response (FIR) bandpass filter can be used to remove global modes of the coupled dynamic system 115. A global mode is a mode in which the system as a whole exhibits coherent temporal oscillations.

[0040] After preprocessing, coherence analysis was performed on the second vibration data. The term "coherence" is used herein in its statistical signal processing sense, measuring the relationship between two signals or datasets. The coherence between two signals is the ratio of the square of the cross spectral density of the two signals to the product of their autospectral densities, as shown in the following expression:

[0041]

[0042] Where a(t) and b(t) are two signals, f is the frequency, and C is the frequency. ab It is the coherence between a and b, Gab G is the cross-spectral density of a and b. aa It is the autospectral density of a, and G bb This is the autospectral density of b. In this case, a and b can each be a dimension of the second vibration data from the same or different second vibration sensors 110 (e.g., a can be the lateral acceleration detected by the second vibration sensor 110 on the control arm, and b can be the vertical acceleration detected by the second vibration sensor 110 on the damper). Coherence can be determined for each pair of dimensions based on the second vibration data.

[0043] Before training the machine learning model, ground-based vibration data is processed by applying a bandpass filter based on coherence analysis. The bandpass filter allows frequencies within its band to pass through and rejects or attenuates frequencies outside the band. The band of the bandpass filter is based on coherence analysis of the ground-based vibration data. The band encompasses the local patterns of the second component 125 relative to the first component 120, as indicated by the coherence between the dimensions of the second component 125 and the dimensions of the first component 120 (e.g., for all...). In the format where the subscript "2nd" indicates the second component 125, the subscript "1st" indicates the first component 120, the subscripts i and j are indices, and x, y, and z are spatial dimensions. After coherence analysis, a Fast Fourier Transform (FFT) is used for the second component 125 to bandpass the second vibration data from the second component 125 to the frequency range with the highest coherence with the first component 120. Then, an Inverse Fast Fourier Transform (iFFT) is used to return the second vibration data and reconstruct the time response of each second component 125, which represents the time-domain behavior of the second component 125 under the condition of highest coherence with the first component 120. When transformed to the frequency domain by FFT and back to the time domain by iFFT, the second vibration data will include the time response of the second component 125 in its first local mode. The aforementioned steps can be performed independently for each second component 125.

[0044] The machine learning model receives first vibration data or statistical features extracted from the first vibration data as input and determines the amount of multiple corresponding vibration energies transmitted to the corresponding second component 125, as shown in the following expression:

[0045]

[0046] Where y est,t It is a vector of the vibrational energy of the second component 125 at time step t, f() is the machine learning model, x t It is a vector of dimensions representing the statistical features extracted from the first vibration or the first vibration data at time step t, and These are the parameters of the machine learning model. For example, yest,t It could be a vector whose dimensions are concatenated together, representing the acceleration at 125 of each second component (e.g., (a x2nd_1 , a y2nd_1 , a z2nd_1 , a x2nd_2 , a y2nd_2 , a z2nd_2 , … a x2nd_i , a y2nd_i , a z2nd_i ,…), where the subscript “2nd_i” indicates the i-th second component 125).

[0047] A machine learning model can be a one-to-many regression model; in other words, it takes data about a single entity as input (e.g., a single vector x of the first component 120). t It outputs data about multiple entities (e.g., multiple vectors y of the second component 125). est,t The machine learning model can be any suitable type for one-to-many regression, such as a neural network. The neural network is used to model the nonlinear relationships involved in the vibration of a dynamic system 115 with motion through connections. For the neural network, the parameters... It can be a set of connection weights between nodes in a neural network.

[0048] The machine learning model is trained to match ground vibration data generated by one or more second vibration sensors 110 respectively mounted on the second component 125. For example, parameters can be adjusted. To minimize the loss function. The loss function can be the mean squared error, as given in the following expression:

[0049]

[0050] in It is a loss, n 训练 It is the number of data points in the training data, and y t It is the ground-based vector of the vibrational energy of the second component 125 at time step t. Training is an optimization scenario where parameters are chosen to minimize the loss, as given in the following expression:

[0051]

[0052] in It is an optimization parameter set. Any suitable optimization algorithm (such as gradient descent, stochastic gradient descent, Adam, etc.) can be used to solve the optimization scenario.

[0053] Before deploying a machine learning model, it can be validated. For example, only a portion (e.g., 70%) of the first and second vibration data can be used for training as described above, and the remaining portion (e.g., 30%) can be used for validation. The same loss function as described above can be used to validate the machine learning model, as given in the following expression:

[0054]

[0055] Only when the loss function is minimized or the test loss is performed A machine learning model can only be deployed when the value is below a pre-selected threshold. Once trained and validated, the machine learning model can be integrated and deployed on either the vehicle computer 205 or the remote computer 210.

[0056] As a general overview of tracking vibrations in the first vehicle 100a, the first vibration sensor 105 detects first vibration data. The vehicle computer 205 can store the first vibration data in a buffer on the first vehicle 100a in response to a startup condition being met. Computers 205 and 210 receive the first vibration data from the buffer. In response to an analysis condition being met, computers 205 and 210 implement a machine learning model. Computers 205 and 210 actuate a third component based on the amount of vibration energy output by the machine learning model (e.g., in response to one of the amounts exceeding a threshold).

[0057] The following steps can be performed by vehicle computer 205, by remote computer 210, or partly by vehicle computer 205 and partly by remote computer 210. For example, if the machine learning model is installed on remote computer 210, vehicle computer 205 can perform initial steps (including storing first vibration data in a buffer and testing whether analysis conditions are met), and then actuate transceiver 225 to transmit the data generated in the initial steps to remote computer 210 (e.g., in response to meeting analysis conditions). Remote computer 210 can then perform intermediate steps (including implementing the machine learning model) and transmit the results of the intermediate steps to vehicle computer 205. Finally, vehicle computer 205 can perform final steps (including actuating a third component). Alternatively, if the machine learning model is installed on vehicle computer 205, vehicle computer 205 can perform all steps.

[0058] The vehicle computer 205 can determine whether the activation conditions are met. The activation conditions can be selected to indicate conditions similar to those used for training machine learning models, and / or to minimize noise in the first vibration data. For example, the activation conditions can be based on the speed of the first vehicle 100a (e.g., the speed remains constant). Activation conditions can also be based on ambient temperature, tire pressure, steering wheel angle, train weight (i.e., the combined weight on the cargo box of the first vehicle 100a and the weight being towed), etc.

[0059] In response to the fulfillment of a startup condition, vehicle computer 205 may store first vibration data in a buffer on the first vehicle 100a. The buffer may be a designated portion of the memory of vehicle computer 205. The buffer may be used to collect a sufficiently large set of first vibration data for implementing a machine learning model. As long as the startup condition is met, vehicle computer 205 may continue to store additional first vibration data in the buffer, and may stop storing additional first vibration data in the buffer in response to the startup condition no longer being met.

[0060] The vehicle computer 205 can determine whether the analysis conditions are met. The analysis conditions can be selected to occur periodically and at a frequency that allows for the collection of a sufficiently large set of first vibration data. For example, the analysis condition could be that the first vehicle 100a is off (i.e., the journey of the first vehicle 100a is complete). For the purposes of this disclosure, an "on state" is defined as a state in which all electrical power is supplied to the electrical components of the first vehicle 100a and the first vehicle 100a is ready to be driven (e.g., the engine is running); a "off state" is defined as a state in which a small amount of electrical power is supplied to selected electrical components of the first vehicle 100a (typically used when the first vehicle 100a is stored); and an "accessory energized state" is defined as a state in which all electrical power is supplied to more electrical components compared to the off state and the first vehicle 100a is not ready to be driven. Typically, the operator puts the first vehicle 100a into the open state when the operator is about to operate it, puts it into the closed state when the operator is about to leave the first vehicle 100a, and puts the first vehicle 100a into the accessory energized state when the operator is about to sit in the first vehicle 100a but not operate it. When the first vehicle 100a changes from the open state to the closed state, the analysis conditions are met.

[0061] In response to the fulfillment of the analysis conditions, computers 205 and 210 receive vibration data from a buffer on vehicle 100. For example, vehicle computer 205 can read first vibration data from the buffer to implement a machine learning model. In another example, vehicle computer 205 can actuate transceiver 225 to transmit the first vibration data from the buffer to remote computer 210, and remote computer 210 receives the first vibration data.

[0062] Computers 205 and 210 implement a machine learning model to determine the amount of vibrational energy transmitted to the second component 125, as described above. The first vibrational data received from the buffer is the input to the machine learning model. The amount can be converted into a relevant quantity to be more useful for the desired evaluation of the second component 125. For example, the amount can be converted into the cumulative energy transmitted to the second component 125 over its lifetime, which is useful for evaluating the lifetime of the second component 125. As another example, the amount can be converted into the peak amount of energy transmitted to the second component 125 during a single driving maneuver, which is useful for evaluating driving maneuvers.

[0063] Vehicle computer 205 actuates a third component of the first vehicle 100a based on the amount of vibration energy. For example, vehicle computer 205 may actuate the third component in response to the amount of vibration energy exceeding a threshold. The threshold may be a preset value stored in the memory of vehicle computer 205. The amount of vibration energy may be the accumulated energy delivered to the second component 125, and the threshold may be selected to indicate that a certain percentage (e.g., 85%) of the lifespan of the second component 125 has been consumed. The amount of vibration energy may be the peak amount of energy delivered to the second component 125 during a single driving maneuver, and the threshold may be selected to indicate that the maneuver is not recommended.

[0064] For example, the third component may be a user interface 220, and the vehicle computer 205 may actuate the user interface 220 to output a message. The message may indicate that the amount of vibration energy exceeds a threshold. Alternatively, the third component may control the movement of the first vehicle 100a (e.g., a propulsion system), and the vehicle computer 205 may actuate the third component to mitigate the vibration energy applied to the second component 125 (e.g., by limiting the magnitude of the value that defines the manipulation of the first vehicle 100a).

[0065] Figure 3This is a flowchart illustrating an example process 300 for training a machine learning model to track the vibrational energy transfer of a second component 125 to a vehicle 100. The memory of a remote computer 210 stores executable instructions for performing the steps of process 300, and / or can be programmed using structures such as those mentioned above. The remote computer 210 running the training may differ from the remote computer 210 on which the trained machine learning model can be installed. As a general overview of process 300, the remote computer 210 receives first vibration data and second vibration data, performs preprocessing steps, performs coherence analysis on the component, filters the second vibration data based on the coherence analysis, trains the machine learning model, and validates the machine learning model.

[0066] Process 300 begins in block 305, where remote computer 210 receives first and second vibration data generated during trial operation, as described above.

[0067] Next, in box 310, remote computer 210 performs a preprocessing step as described above.

[0068] Next, in box 315, remote computer 210 performs a coherence analysis as described above.

[0069] Next, in box 320, remote computer 210 applies a bandpass filter to the second vibration data based on coherence analysis, as described above.

[0070] Next, in box 325, remote computer 210 trains a machine learning model as described above.

[0071] Next, in box 330, remote computer 210 validates the machine learning model, as described above. After box 330, process 300 ends.

[0072] Figure 4 This is a flowchart illustrating an example process 400 for tracking the vibration energy transfer to a second component 125 of a first vehicle 100a. The memories of computers 205 and 210 store executable instructions for performing the steps of process 400, and / or can be programmed in a structure as mentioned above. As a general overview of process 400, computers 205 and 210 receive vehicle data. Computers 205 and 210 store first vibration data in a buffer as soon as a start-up condition is met. In response to the satisfaction of analysis conditions, computers 205 and 210 implement a machine learning model. In response to the amount of vibration energy exceeding a threshold, computers 205 and 210 actuate a third component.

[0073] Process 400 begins in box 405, where computers 205 and 210 receive vehicle data. The vehicle data includes first vibration data generated by the first vibration sensor 105. The vehicle data may also include data used to assess startup conditions and / or analyze conditions. For example, the vehicle data may include the speed of the first vehicle 100a, ambient temperature, tire pressure, steering wheel angle, train weight, ignition status, etc.

[0074] Next, in decision box 410, computers 205 and 210 determine whether the startup conditions are met, as described above. If the startup conditions are met, process 400 proceeds to box 415. If the startup conditions are not met, process 400 proceeds to decision box 420.

[0075] In box 415, computers 205 and 210 store the first vibration data in a buffer, as described above. After box 415, process 400 returns to box 405 to continue collecting vehicle data.

[0076] In decision box 420, computers 205 and 210 determine whether the analysis conditions are met, as described above. If the analysis conditions are met, process 400 proceeds to box 425. If the analysis conditions are not met, process 400 returns to box 405 to continue collecting vehicle data until the analysis conditions are met.

[0077] In box 425, computers 205 and 210 receive first vibration data from the buffer and implement a machine learning model to determine the amount of vibration energy transmitted to the second component 125, as described above.

[0078] Next, in decision box 430, computers 205 and 210 determine whether any of the vibration energy quantities exceeds a threshold, as described above. In response to at least one vibration energy quantity exceeding the threshold, process 400 proceeds to box 435. In response to no vibration energy quantity exceeding the threshold, process 400 terminates.

[0079] In box 435, computers 205 and 210 actuate the third component as described above. After box 435, process 400 ends.

[0080] Generally, the described computing system and / or device may employ any of a number of computer operating systems, including, but not limited to, the following versions and / or types: Ford Sync® applications; AppLink / Smart Device Link middleware; Microsoft Automotive® operating system; Microsoft Windows® operating system; Unix operating system (e.g., Solaris® operating system released by Oracle Corporation, Redwood Coast, California); AIX UNIX operating system released by International Business Machines, Inc., Armonk, New York; Linux operating system; Mac OSX and iOS operating systems released by Apple Inc., Inc., Cupertino, California; BlackBerry OS released by BlackBerry Ltd., Inc., Waterloo, Canada; and Android operating system developed by Google and the Open Handset Alliance; or the QNX® CAR infotainment platform provided by QNX Software Systems, Inc. Examples of computing devices include, but are not limited to, in-vehicle computers, computer workstations, servers, desktop computers, laptops, notebook computers, or handheld computers, or any other computing system and / or device.

[0081] Computing devices typically include computer-executable instructions, which can be executed by one or more computing devices (such as those listed above). Computer-executable instructions can be compiled or interpreted from computer programs created using a variety of programming languages ​​and / or technologies, which, individually or in combination, include, but are not limited to, Java™, C, C++, Matlab, Simulink, Stateflow, Visual Basic, JavaScript, Python, Perl, HTML, etc. Some of these applications can be compiled and executed on virtual machines such as the Java Virtual Machine, the Dalvik Virtual Machine, etc. Generally, a processor (e.g., a microprocessor) receives instructions (e.g., from memory, computer-readable media, etc.) and executes those instructions to perform one or more processes, including one or more processes described herein. Such instructions and other data can be stored and transferred using a variety of computer-readable media. Files in a computing device are typically collections of data stored on computer-readable media such as storage media, random access memory, etc.

[0082] Computer-readable media (also known as processor-readable media) include any non-transitory (e.g., tangible) medium that contributes to providing data (e.g., instructions) that can be read by a computer (e.g., by the computer's processor). Such media can take many forms, including but not limited to non-volatile and volatile media. Instructions can be transmitted via one or more transmission media, including optical fibers, wires, wireless communications, and internals that constitute a system bus coupled to the computer's processor. Common forms of computer-readable media include, for example, RAM, PROM, EPROM, flash EEPROM, any other memory chip or magnetic tape, or any other medium from which a computer can read.

[0083] The databases, data repositories, or other data stores described herein can include various mechanisms for storing, accessing / retrieving various types of data, including hierarchical databases, file sets in file systems, application databases in proprietary formats, relational database management systems (RDBMS), NoSQL databases, graph databases (GDB), and so on. Each such data store is typically contained within a computing device employing a computer operating system such as those mentioned above, and is accessed via a network in any one or more of various ways. File systems can be accessed from the computer operating system and can include files stored in various formats. In addition to languages ​​used to create, store, edit, and execute the stored programs (such as PL / SQL as described above), RDBMS typically employs Structured Query Language (SQL).

[0084] In some examples, system elements may be implemented as computer-readable instructions (e.g., software) stored on one or more computing devices (e.g., servers, personal computers, etc.) and on computer-readable media (e.g., disks, storage, etc.) associated therewith. Computer program products may include such instructions stored on computer-readable media for performing the functions described herein.

[0085] In the accompanying drawings, the same reference numerals indicate the same elements. Furthermore, some or all of these elements may be changed. Regarding the media, processes, systems, methods, inspirations, etc., described herein, it should be understood that although the steps of such processes, etc., are described as occurring in a certain ordered order, such processes can be practiced by performing the steps in a different order than that described herein. It should also be understood that some steps may be performed simultaneously, other steps may be added, or some steps described herein may be omitted. The operations, systems, and methods described herein should always be implemented and / or performed in accordance with the applicable owner / user manual and / or safety guidelines.

[0086] This disclosure has been described in an illustrative manner, and it should be understood that the terminology used is intended to be descriptive in nature, not restrictive. The use of “in response to,” “after determining,” and “after transforming” indicates a causal relationship, not just a temporal one. The adjectives “first,” “second,” and “third” are used throughout this document as identifiers and are not intended to indicate importance, order, or quantity. In view of the foregoing teachings, many modifications and variations of this disclosure are possible, and this disclosure may be practiced in ways other than those specifically described.

[0087] According to the present invention, a computer is provided having a processor and a memory, the memory storing instructions executable by the processor to: receive first vibration data detected by a first vibration sensor, the first vibration sensor being of a first type, the first vibration sensor being mounted on a first component of a first vehicle, the first vehicle including a coupled dynamic system comprising the first component and a second component, the second component of the first vehicle having no vibration sensor mounted thereon; implement a machine learning model to determine the amount of vibration energy transmitted to the second component, the first vibration data being input to the machine learning model; and actuate a third component of the first vehicle based on the amount of vibration energy; wherein the machine learning model is trained to match ground-based vibration data generated by a second vibration sensor, the second vibration sensor being mounted on the second component of a second vehicle, the second vibration sensor being of a second type different from the first type.

[0088] According to an embodiment, the first type is a road noise cancellation (RNC) accelerometer.

[0089] According to an embodiment, the second type is an internal circuit piezoelectric (ICP) sensor.

[0090] According to an embodiment: the connected dynamic system includes a plurality of second components, each of which has no vibration sensor mounted thereon; and the machine learning model is trained to match ground vibration data generated by the plurality of second vibration sensors respectively mounted on the second components.

[0091] According to an embodiment, the machine learning model is a one-to-many regression model.

[0092] According to an embodiment, each second vibration sensor is of the second type.

[0093] According to an embodiment, the connected dynamic system is the suspension system of the wheels of the first vehicle.

[0094] According to an embodiment, the first component is a control arm.

[0095] According to an embodiment, ground vibration data is processed by applying a bandpass filter before training the machine learning model.

[0096] According to an embodiment, the bandpass filter's frequency band is based on coherence analysis of ground vibration data.

[0097] According to an embodiment, the instructions also include instructions to actuate a third component in response to the amount of vibration energy exceeding a threshold.

[0098] According to an embodiment, the instruction for receiving the first vibration data includes an instruction for receiving the first vibration data from a buffer on the first vehicle.

[0099] According to an embodiment, the instructions also include storing the first vibration data in a buffer in response to the fulfillment of a start-up condition.

[0100] According to an embodiment, the activation condition is based on the speed of the first vehicle.

[0101] According to the present invention, a method includes: training a machine learning model to match ground vibration data generated by a second vibration sensor, the second vibration sensor being mounted on a second component of a second vehicle, the second vibration sensor being of a second type; receiving first vibration data detected by a first vibration sensor, the first vibration sensor being of a first type different from the second type, the first vibration sensor being mounted on a first component of a first vehicle, the first vehicle including a coupled dynamic system including the first component and the second component, the second component of the first vehicle having no vibration sensor mounted thereon; implementing the machine learning model to determine the amount of vibration energy transmitted to the second component, the first vibration data being input to the machine learning model; and actuating a third component of the first vehicle based on the amount of vibration energy.

[0102] According to an embodiment, the first type is a road noise cancellation (RNC) accelerometer, and the second type is an internal circuit piezoelectric (ICP) sensor.

[0103] According to an embodiment, the connected dynamic system includes a plurality of second components, each of which has no vibration sensor mounted thereon; the machine learning model is trained to match ground vibration data generated by a plurality of second vibration sensors respectively mounted on the second components; and each second vibration sensor is of a second type.

[0104] According to an embodiment, the machine learning model is a one-to-many regression model.

[0105] According to an embodiment, the connected dynamic system is the suspension system of the wheels of the first vehicle, and the first component is a control arm.

[0106] According to an embodiment, the invention is further characterized in that, before training the machine learning model, the ground vibration data is processed by applying a bandpass filter, wherein the frequency band of the bandpass filter is based on the coherence analysis of the ground vibration data.

Claims

1. A method comprising: Receive first vibration data detected by a first vibration sensor, the first vibration sensor being of a first type, the first vibration sensor being mounted on a first component of a first vehicle, the first vehicle including a coupled dynamic system, the coupled dynamic system including the first component and a second component, the second component of the first vehicle having no vibration sensor mounted thereon; A machine learning model is implemented to determine the amount of vibrational energy transmitted to the second component, wherein the first vibrational data is the input to the machine learning model; as well as The third component of the first vehicle is actuated based on the amount of vibration energy. The machine learning model is trained to match ground vibration data generated by a second vibration sensor, which is mounted on the second component of the second vehicle and is a second type different from the first type.

2. The method of claim 1, wherein the first type is a road noise cancellation (RNC) accelerometer.

3. The method of claim 1, wherein the second type is an internal circuit piezoelectric (ICP) sensor.

4. The method of claim 1, wherein: The connected dynamic system includes a plurality of second components, each of which has no vibration sensor mounted thereon; and The machine learning model is trained to match the ground vibration data generated by multiple second vibration sensors respectively mounted on the second component.

5. The method of claim 4, wherein the machine learning model is a one-to-many regression model.

6. The method of claim 4, wherein each second vibration sensor is of the second type.

7. The method of claim 1, wherein the dynamic system of the connection is the suspension system of the wheels of the first vehicle.

8. The method of claim 7, wherein the first component is a control arm.

9. The method of claim 1, wherein the ground vibration data is processed by applying a bandpass filter before training the machine learning model.

10. The method of claim 9, wherein the bandpass filter frequency band is based on coherence analysis of the ground vibration data.

11. The method of claim 1, further comprising actuating the third component in response to the amount of vibration energy exceeding a threshold.

12. The method of claim 1, wherein receiving the first vibration data includes receiving the first vibration data from a buffer on the first vehicle.

13. The method of claim 12, further comprising storing the first vibration data in the buffer in response to a start-up condition being met.

14. The method of claim 13, wherein the initiation condition is based on the speed of the first vehicle.

15. A computer comprising a processor and a memory, the memory storing instructions executable by the processor to perform the method as claimed in any one of claims 1 to 14.

Images