US20220194394A1 - Machine learning device and machine learning system - Google Patents
Machine learning device and machine learning system Download PDFInfo
- Publication number
- US20220194394A1 US20220194394A1 US17/690,725 US202217690725A US2022194394A1 US 20220194394 A1 US20220194394 A1 US 20220194394A1 US 202217690725 A US202217690725 A US 202217690725A US 2022194394 A1 US2022194394 A1 US 2022194394A1
- Authority
- US
- United States
- Prior art keywords
- vehicle
- machine learning
- replacement
- neural network
- learning model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
- 238000010801 machine learning Methods 0.000 title claims abstract description 54
- 238000012549 training Methods 0.000 claims abstract description 14
- 238000003062 neural network model Methods 0.000 claims description 63
- 230000004044 response Effects 0.000 claims 1
- 238000004891 communication Methods 0.000 description 31
- 230000000875 corresponding effect Effects 0.000 description 16
- 238000012545 processing Methods 0.000 description 12
- 230000006870 function Effects 0.000 description 8
- 238000002485 combustion reaction Methods 0.000 description 5
- 230000004913 activation Effects 0.000 description 4
- 239000004065 semiconductor Substances 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000446 fuel Substances 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000000034 method Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 239000003054 catalyst Substances 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000000746 purification Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W50/0097—Predicting future conditions
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/12—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to parameters of the vehicle itself, e.g. tyre models
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W50/04—Monitoring the functioning of the control system
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W50/06—Improving the dynamic response of the control system, e.g. improving the speed of regulation or avoiding hunting or overshoot
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
- G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W2050/0001—Details of the control system
- B60W2050/0002—Automatic control, details of type of controller or control system architecture
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W2050/0062—Adapting control system settings
- B60W2050/0075—Automatic parameter input, automatic initialising or calibrating means
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W2556/00—Input parameters relating to data
- B60W2556/45—External transmission of data to or from the vehicle
Definitions
- the present disclosure relates to a machine learning device and a machine learning system.
- a neural network is used to predict the temperature of an exhaust purification catalyst provided in an internal combustion engine.
- the neural network model is trained in a server, and the trained neural network model is sent from the server to a vehicle.
- a machine learning model used in a vehicle is usually tailored to the properties of the vehicle part before replacement. Therefore, if replacement of a vehicle part causes the properties of the vehicle part to change, the precision of prediction of the machine learning model is liable to fall.
- an object of the present disclosure is to keep the precision of prediction of a machine learning model used in a vehicle from falling if a vehicle part is replaced.
- a machine learning device provided in a vehicle, comprising: a predicting part configured to use a machine learning model to predict predetermined information; an updating part configured to update the machine learning model; and a part information acquiring part configured to detect replacement of a vehicle part and acquire identification information of the vehicle part after replacement, wherein the updating part is configured to receive a new machine learning model trained using training data sets corresponding to the vehicle part after replacement from a server and apply the new machine learning model to the vehicle, if a vehicle part relating to input data of the machine learning model is replaced with a vehicle part of a different configuration.
- a machine learning system comprising a server and a vehicle, wherein the server comprises: a first communication device able to communicate with the vehicle; and a control device, the vehicle comprises: a second communication device able to communicate with the server; a predicting part configured to use a machine learning model to predict predetermined information; an updating part configured to update the machine learning model; and a part information acquiring part configured to detect replacement of a vehicle part and acquire identification information of the vehicle part after replacement, the updating part is configured to transmit identification information of the vehicle part after replacement to a server, if a vehicle part relating to input data of the machine learning model is replaced with a vehicle part of a different configuration, the control device is configured to transmit a new machine learning model trained using training data sets corresponding to the vehicle part after replacement to the vehicle, and the updating part is configured to apply the new machine learning model to the vehicle.
- FIG. 1 is a schematic view of the configuration of a machine learning system according to an embodiment of the present disclosure.
- FIG. 2 is a view schematically showing a part of the configuration of a vehicle of FIG. 1 .
- FIG. 3 is a functional block diagram of an ECU of FIG. 2 .
- FIG. 4 shows one example of a neural network model having a simple configuration.
- FIG. 5 is a flow chart showing a control routine of processing for updating a model.
- FIG. 6 is a flow chart showing a control routine of processing for transmitting a model.
- FIG. 1 is a schematic view of a configuration of a machine learning system 1 according to the embodiment of the present disclosure.
- the machine learning system 1 is provided with a server 2 and a vehicle 3 .
- the server 2 is provided outside of the vehicle 3 and is provided with a communication interface 21 , a storage device 22 , a memory 23 , and a processor 24 .
- the server 2 may be further provided with an input device such as a keyboard and mouse and an output device such as a display etc. Further, the server 2 may be configured by a plurality of computers.
- the communication interface 21 can communicate with the vehicle 3 and enables the server 2 to communicate with the vehicle 3 .
- the communication interface 21 has an interface circuit for connecting the server 2 to the communication network 5 .
- the server 2 communicates with the vehicle 3 through the communication interface 21 , communication network 5 , and wireless base station 6 .
- the communication interface 21 is one example of a communication device.
- the storage device 22 for example, has a hard disk drive (HDD), solid state drive (SSD), optical storage medium, etc.
- the storage device 22 stores various types of data, for example, stores computer programs for the processor 24 to perform various processing, etc.
- the memory 23 for example, has a semiconductor memory such as a random access memory (RAM).
- the memory 23 for example, stores various data etc., used when various processing is performed by the processor 24 .
- the communication interface 21 , the storage device 22 , and the memory 23 are connected through signal wires to the processor 24 .
- the processor 24 has one or more CPUs and peripheral circuits and performs various processing. Note that, the processor 24 may further have processing circuits such as arithmetic logic units or numerical calculation units.
- the processor 24 is an example of a control device.
- FIG. 2 is a view schematically showing a part of the configuration of the vehicle 3 of FIG. 1 .
- the vehicle 3 is provided with actuators 31 , sensors 32 , a communication module 33 , and an electronic control unit (ECU) 40 . That is, the actuators 31 , the sensors 32 , the communication module 33 , and the ECU 40 are mounted in the vehicle 3 .
- the actuators 31 , the sensors 32 , and the communication module 33 are connected to be able to communicate with the ECU 40 through an internal vehicle network based on the CAN (Controller Area Network) or other standard.
- CAN Controller Area Network
- the actuators 31 are operating parts required for the vehicle 3 to run. For example, they include parts required for operation of the internal combustion engine (spark plugs, fuel injectors, a throttle valve-drive actuator, an EGR control valve, etc.), a motor, a clutch, etc.
- the ECU 40 controls the actuators 31 .
- the sensors 32 detect the states of the vehicle 3 , the internal combustion engine, the battery, etc., and include a vehicle speed sensor, a GPS receiver, an accelerator opening degree sensor, an air flowmeter, an air-fuel ratio sensor, a crank angle sensor, a torque sensor, a voltage sensor, a temperature sensor, etc.
- the outputs of the sensors 32 are sent to the ECU 40 .
- the actuators 31 and the sensors 32 constitute vehicle parts mounted in the vehicle 3 .
- the vehicle parts include the actuators 31 and the sensors 32 .
- the communication module 33 is a device enabling communication between the vehicle 3 and the outside of the vehicle 3 .
- the communication module 33 for example, is a data communication module (DCM) able to communicate with a communication network 5 through a wireless base station 6 .
- DCM data communication module
- a mobile terminal for example, a smart phone, a tablet terminal, a WiFi router, etc., may be used.
- the communication module 33 is one example of a communication device.
- the ECU 40 includes a communication interface 41 , a memory 42 , and a processor 43 and performs various control operations of the vehicle 3 . Note that, in the present embodiment, a single ECU 40 is provided, but a plurality of ECUs may be provided for the different functions.
- the communication interface 41 is an interface circuit for connecting the ECU 40 to an internal vehicle network based on the CAN or other standard.
- the ECU 40 communicates with other vehicle-mounted devices through the communication interface 41 .
- the memory 42 for example, has a volatile semiconductor memory (for example, a RAM) and nonvolatile semiconductor memory (for example, a ROM).
- the memory 42 stores programs run in the processor 43 , various data used when the various processings are performed by the processor 43 , etc.
- the processor 43 has one or more CPUs (central processing units) and their peripheral circuits and performs various processing. Note that, the processor 43 may further have processing circuits such as arithmetic logic units or numerical calculation units.
- the communication interface 41 , the memory 42 and the processor 43 are connected to each other through signal wires.
- FIG. 3 is a functional block diagram of the ECU 40 of FIG. 2 .
- the ECU 40 has a predicting part 51 , an updating part 52 , and a part information acquiring part 53 .
- the predicting part 51 , the updating part 52 , and the part information acquiring part 53 are functional blocks realized by a program stored in the memory 42 of the ECU 40 run by the processor 43 of the ECU 40 .
- the predicting part 51 predicts predetermined information using a machine learning model.
- the updating part 52 updates the machine learning model used in the vehicle 3 .
- the machine learning model a neural network model outputting at least one output data (which is also called “a target variable”) from a plurality of input data (which are also called “explanatory variables”) is used.
- FIG. 4 shows one example of a neural network model having a simple configuration.
- FIG. 4 The circle marks in FIG. 4 show artificial neurons.
- An artificial neuron is usually called a “node” or “unit” (in this Description, called a “node”).
- the activation function is for example a Sigmoid function 6 .
- the corresponding weights “w” and biases “b” are used to calculate the total input value “u” ( ⁇ z ⁇ w+b) or only the corresponding weights “w” are used to calculate the total input value “u” ( ⁇ z+w).
- an identity function is used as the activation function.
- the total input value “u” calculated at the node of the output layer is output as it is as the output value “y” from the node of the output layer.
- the neural network model used in the vehicle 3 is stored in the memory 42 of the ECU 40 or another storage device provided at the vehicle 3 .
- the predicting part 51 inputs values of a plurality of input data (input values) to the neural network model to thereby make the neural network model output the value of at least one output data (output value).
- the values of the input data for example, output values of the sensors 32 , command values from the ECU 40 to the actuators 31 , values obtained from outside the vehicle 3 through the communication module 33 , values calculated at the ECU 40 , etc., are used.
- the predicting part 51 predicts predetermined information corresponding to the value of the output data which is output from the neural network model.
- the predetermined information predicted by the predicting part 51 is used for control of the vehicle 3 .
- the neural network model may be either of a regression model and a classification model. If the neural network model is a regression model, the output data, for example, is the output torque of the internal combustion engine, the concentration of predetermined constituent in the exhaust gas (HC, NOx, etc.), the state of charge (SOC) of the battery, etc.
- the output data is, for example, a result of judgment of abnormality of a predetermined sensor 32 , a result of judgment of startup of the internal combustion engine in a hybrid vehicle, etc.
- a state quantity correlated with the output data is suitably selected.
- training data sets comprised of combinations of measured values of a plurality of input data and measured values of at least one output data corresponding to these measured values of the plurality of input data (truth data or teacher data) are used.
- the measured values of the input data and output data are for example acquired using an engine bench, etc., and the training data sets are created by combining the corresponding measured values.
- the neural network model In training the neural network model, a large number of training data sets are used and the known back propagation method is used to repeatedly update the parameters of the neural network model (weights “w” and biases “b”) so that the difference between the values of the output data output by the neural network model and the measured values of the output data becomes smaller.
- the neural network model is trained and a trained neural network model is created.
- the created trained neural network model is loaded in the ECU 40 of the vehicle 3 before shipment of the vehicle 3 . That is, information of the trained neural network model (structure, weights “w”, biases “b”, etc., of the model) is stored in the memory 42 of the ECU 40 or another storage device provided in the vehicle 3 .
- the trained neural network model may be transmitted from the server 2 to the vehicle 3 after the vehicle 3 is shipped.
- a vehicle part such as the actuator 31 or the sensor 32 deteriorates or breaks down along with use of the vehicle 3 , the vehicle part is replaced. Further, sometimes a vehicle part is replaced in order to mount a higher performance vehicle part in the vehicle 3 .
- the neural network model loaded in the ECU 40 is tailored to the properties of the vehicle part before replacement. For this reason, if the properties of a vehicle part change due to replacement of the vehicle part, the precision of prediction of the neural network model is liable to fall.
- the part information acquiring part 53 detects replacement of a vehicle part and acquires identification information of the vehicle part after replacement. Further, if a vehicle part relating to input data of the neural network model is replaced with a vehicle part of a different configuration, the updating part 52 receives a new neural network model trained using training data sets corresponding to the vehicle part after replacement from the server 2 and applies the new neural network model to the vehicle 3 .
- the updating part 52 transmits identification information of the vehicle part after replacement to the server 2 . If receiving identification information of the vehicle part after replacement from the vehicle 3 , the processor 24 of the server 2 transmits the new neural network model trained using training data sets corresponding to the vehicle part after replacement to the vehicle 3 . If receiving the new neural network model from the server 2 , the updating part 52 applies the new neural network model to the vehicle 3 .
- FIG. 5 is a flow chart showing a control routine of processing for updating a model.
- the present control routine is repeatedly performed at predetermined run intervals by the ECU 40 of the vehicle 3 .
- the part information acquiring part 53 judges whether a replacement flag F is “ 1 ”.
- the replacement flag F is a flag which is set to “ 1 ” when a vehicle part is replaced.
- the initial value of the replacement flag F is zero. If at step S 101 it is judged that the replacement flag F is zero, the present control routine proceeds to step S 102 .
- the part information acquiring part 53 judges whether a vehicle part relating to input data of the neural network model has been replaced with a vehicle part of a different configuration. For example, if an output value of the sensor 32 is used as input data of the neural network model, the sensor 32 corresponds to a vehicle part relating to input data of the neural network model. Further, if a command value from the ECU 40 to the actuator 31 is used as input data of the neural network model, the actuator 31 corresponds to a vehicle part relating to input data of the neural network model. Further, a “vehicle part of a different configuration” includes a vehicle part of a different manufacturer, part number, model number, etc., than the vehicle part before replacement.
- the part information acquiring part 53 reads the identification information of the vehicle part (manufacturer's name, model number, serial number, manufacturing number, etc.,) to detect replacement of a vehicle part. Further, the part information acquiring part 53 may detect an output waveform, an operating voltage, etc., of a vehicle part to detect replacement of a vehicle part. Further, a worker may input replacement of a vehicle part to an input device (HMI etc.) of the vehicle 3 , and the part information acquiring part 53 may detect replacement of a vehicle part based on input to the input device.
- the identification information of the vehicle part manufacturer's name, model number, serial number, manufacturing number, etc.
- the part information acquiring part 53 may detect an output waveform, an operating voltage, etc., of a vehicle part to detect replacement of a vehicle part.
- a worker may input replacement of a vehicle part to an input device (HMI etc.) of the vehicle 3 , and the part information acquiring part 53 may detect replacement of a vehicle part based on input to the input device
- step S 102 If at step S 102 it is judged that a vehicle part has not been replaced, the present control routine ends. On the other hand, if at step S 102 it is judged that a vehicle part has been replaced, the present control routine proceeds to step S 103 .
- step S 103 the part information acquiring part 53 acquires the identification information of the vehicle part.
- step S 104 the part information acquiring part 53 sets the replacement flag F to 4 , r.
- the updating part 52 transmits the identification information of the vehicle part after replacement to the server 2 through the communication module 33 , the wireless base station 6 , and the communication network 5 .
- step S 105 the present control routine proceeds to step S 106 .
- step S 101 it is judged that the replacement flag F is “ 1 ”
- step S 102 the present control routine skips step S 102 to step S 105 and proceeds to step S 106 .
- step S 106 the updating part 52 judges whether it has received a new neural network model from the server 2 . If it is judged that it has not received a new neural network model, the present control routine ends. On the other hand, if it is judged that it has received a new neural network model, the present control routine proceeds to step S 107 .
- the updating part 52 applies the new neural network model to the vehicle 3 .
- the updating part 52 updates the neural network model used in the vehicle 3 to the new neural network model. That is, the information on the neural network model stored in the memory 42 of the ECU 40 or another storage device is rewritten.
- step S 108 the updating part 52 sets the replacement flag F to zero.
- step S 108 the present control routine ends.
- FIG. 6 is a flow chart showing a control routine of processing for transmitting a model.
- the present control routine is repeatedly performed at predetermined run intervals by the processor 24 of the server 2 .
- step S 201 the processor 24 judges whether it has received identification information on a vehicle part from the vehicle 3 . If it is judged that it has not received identification information on a vehicle part, the present control routine ends. On the other hand, if it is judged that it has received identification information on a vehicle part, the present control routine proceeds to step S 202 .
- the processor 24 identifies the vehicle part after replacement based on the identification information of the vehicle part and identifies the neural network model corresponding to the vehicle part after replacement. Specifically, the processor 24 extracts a neural network model trained using training data sets corresponding to the vehicle part after replacement from the plurality of neural network models stored in the storage device 22 of the server 2 . In the training data set corresponding to the vehicle part after replacement, the output value of the vehicle part after replacement, the command value to the vehicle part after replacement, etc., is used as one of the measured values of the input data.
- step S 203 the processor 24 transmits the new neural network model identified at step S 202 to the vehicle 3 .
- step S 203 the present control routine ends.
- the updating part 52 may transmit the identification information of the vehicle part after replacement and the identification information of the vehicle 3 (vehicle name, model, etc.) to the server 2 and, at step S 203 of FIG. 6 , the processor 24 of the server 2 may transmit, as a new neural network model, a neural network model corresponding to the identification information of the vehicle part after replacement and the identification information of the vehicle 3 to the vehicle 3 .
- the processor 24 of the server 2 may transmit, as a new neural network model, a neural network model corresponding to the identification information of the vehicle part after replacement and the identification information of the vehicle 3 to the vehicle 3 .
- the identification information of the vehicle 3 is stored in the memory 42 of the ECU 40 etc., in advance.
- a neural network model created for each type of vehicle for example, a neural network model trained using training data sets created from measured values of input data and output data acquired using an engine bench etc., prepared for each of the types of vehicles, is stored in advance in the storage device 22 of the server 2 .
- the machine learning model used in the vehicle 3 may be a machine learning model other than a neural network such as a random forest, a k-nearest neighbor method, a support vector machine, etc.
- ECU electronice control unit
Landscapes
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Computer Interaction (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Neurology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Feedback Control In General (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Vehicle Cleaning, Maintenance, Repair, Refitting, And Outriggers (AREA)
Abstract
The machine learning device includes a predicting part configured to use a machine learning model to predict predetermined information, an updating part configured to update the machine learning model, and a part information acquiring part configured to detect replacement of a vehicle part and acquire identification information of the vehicle part after replacement. The updating part is configured to receive a new machine learning model trained using training data sets corresponding to the vehicle part after replacement from a server and apply the new machine learning model to the vehicle, if a vehicle part relating to input data of the machine learning model is replaced with a vehicle part of a different configuration.
Description
- This application claims priority to U.S. patent application Ser. No. 17/405,390 filed on Aug. 18, 2021, the entire contents of which are hereby incorporated
- The present disclosure relates to a machine learning device and a machine learning system.
- In recent years, along with the development of AI (artificial intelligence) technology, in vehicles, control using machine learning models such as neural network models has been studied. For example, in the machine learning system described in
PTL 1, a neural network is used to predict the temperature of an exhaust purification catalyst provided in an internal combustion engine. - In order to improve the precision of prediction of such a neural network model, it is necessary to train the neural network model in advance. For example, as described in
PTL 1, the neural network model is trained in a server, and the trained neural network model is sent from the server to a vehicle. - [PTL 1] Japanese Unexamined Patent Publication No. 2019-183698
- In this regard, if, along with use of a vehicle, the vehicle parts become degraded or break down, the vehicle parts are replaced. Further, vehicle parts are sometimes replaced for mounting higher performance vehicle parts in a vehicle.
- However, a machine learning model used in a vehicle is usually tailored to the properties of the vehicle part before replacement. Therefore, if replacement of a vehicle part causes the properties of the vehicle part to change, the precision of prediction of the machine learning model is liable to fall.
- Therefore, in consideration of the above problem, an object of the present disclosure is to keep the precision of prediction of a machine learning model used in a vehicle from falling if a vehicle part is replaced.
- The summary of the present disclosure is as follows.
- (1) A machine learning device provided in a vehicle, comprising: a predicting part configured to use a machine learning model to predict predetermined information; an updating part configured to update the machine learning model; and a part information acquiring part configured to detect replacement of a vehicle part and acquire identification information of the vehicle part after replacement, wherein the updating part is configured to receive a new machine learning model trained using training data sets corresponding to the vehicle part after replacement from a server and apply the new machine learning model to the vehicle, if a vehicle part relating to input data of the machine learning model is replaced with a vehicle part of a different configuration.
- (2) The machine learning device described in above (1), wherein the updating part is configured to transmit identification information of the vehicle part after replacement and identification information of the vehicle to the server, and the new machine learning model is a machine learning model corresponding to the identification information of the vehicle.
- (3) The machine learning device described in above (1) or (2), wherein the machine learning model is a neural network model.
- (4) A machine learning system comprising a server and a vehicle, wherein the server comprises: a first communication device able to communicate with the vehicle; and a control device, the vehicle comprises: a second communication device able to communicate with the server; a predicting part configured to use a machine learning model to predict predetermined information; an updating part configured to update the machine learning model; and a part information acquiring part configured to detect replacement of a vehicle part and acquire identification information of the vehicle part after replacement, the updating part is configured to transmit identification information of the vehicle part after replacement to a server, if a vehicle part relating to input data of the machine learning model is replaced with a vehicle part of a different configuration, the control device is configured to transmit a new machine learning model trained using training data sets corresponding to the vehicle part after replacement to the vehicle, and the updating part is configured to apply the new machine learning model to the vehicle.
- According to the present disclosure, it is possible to keep the precision of prediction of a machine learning model used in a vehicle from falling if a vehicle part is replaced.
-
FIG. 1 is a schematic view of the configuration of a machine learning system according to an embodiment of the present disclosure. -
FIG. 2 is a view schematically showing a part of the configuration of a vehicle ofFIG. 1 . -
FIG. 3 is a functional block diagram of an ECU ofFIG. 2 . -
FIG. 4 shows one example of a neural network model having a simple configuration. -
FIG. 5 is a flow chart showing a control routine of processing for updating a model. -
FIG. 6 is a flow chart showing a control routine of processing for transmitting a model. - Below, referring to the drawings, embodiments of the present disclosure will be explained in detail.
FIG. 1 is a schematic view of a configuration of amachine learning system 1 according to the embodiment of the present disclosure. Themachine learning system 1 is provided with aserver 2 and avehicle 3. - As shown in
FIG. 1 , theserver 2 is provided outside of thevehicle 3 and is provided with acommunication interface 21, astorage device 22, amemory 23, and aprocessor 24. Note that, theserver 2 may be further provided with an input device such as a keyboard and mouse and an output device such as a display etc. Further, theserver 2 may be configured by a plurality of computers. - The
communication interface 21 can communicate with thevehicle 3 and enables theserver 2 to communicate with thevehicle 3. Specifically, thecommunication interface 21 has an interface circuit for connecting theserver 2 to the communication network 5. Theserver 2 communicates with thevehicle 3 through thecommunication interface 21, communication network 5, and wireless base station 6. Thecommunication interface 21 is one example of a communication device. - The
storage device 22, for example, has a hard disk drive (HDD), solid state drive (SSD), optical storage medium, etc. Thestorage device 22 stores various types of data, for example, stores computer programs for theprocessor 24 to perform various processing, etc. - The
memory 23, for example, has a semiconductor memory such as a random access memory (RAM). Thememory 23, for example, stores various data etc., used when various processing is performed by theprocessor 24. - The
communication interface 21, thestorage device 22, and thememory 23 are connected through signal wires to theprocessor 24. Theprocessor 24 has one or more CPUs and peripheral circuits and performs various processing. Note that, theprocessor 24 may further have processing circuits such as arithmetic logic units or numerical calculation units. Theprocessor 24 is an example of a control device. -
FIG. 2 is a view schematically showing a part of the configuration of thevehicle 3 ofFIG. 1 . As shown inFIG. 2 , thevehicle 3 is provided withactuators 31,sensors 32, acommunication module 33, and an electronic control unit (ECU) 40. That is, theactuators 31, thesensors 32, thecommunication module 33, and theECU 40 are mounted in thevehicle 3. Theactuators 31, thesensors 32, and thecommunication module 33 are connected to be able to communicate with theECU 40 through an internal vehicle network based on the CAN (Controller Area Network) or other standard. - The
actuators 31 are operating parts required for thevehicle 3 to run. For example, they include parts required for operation of the internal combustion engine (spark plugs, fuel injectors, a throttle valve-drive actuator, an EGR control valve, etc.), a motor, a clutch, etc. The ECU 40 controls theactuators 31. - The
sensors 32 detect the states of thevehicle 3, the internal combustion engine, the battery, etc., and include a vehicle speed sensor, a GPS receiver, an accelerator opening degree sensor, an air flowmeter, an air-fuel ratio sensor, a crank angle sensor, a torque sensor, a voltage sensor, a temperature sensor, etc. The outputs of thesensors 32 are sent to theECU 40. - The
actuators 31 and thesensors 32 constitute vehicle parts mounted in thevehicle 3. In other words, the vehicle parts include theactuators 31 and thesensors 32. - The
communication module 33 is a device enabling communication between thevehicle 3 and the outside of thevehicle 3. Thecommunication module 33, for example, is a data communication module (DCM) able to communicate with a communication network 5 through a wireless base station 6. Note that, as thecommunication module 33, a mobile terminal (for example, a smart phone, a tablet terminal, a WiFi router, etc.,) may be used. Thecommunication module 33 is one example of a communication device. - The
ECU 40 includes acommunication interface 41, amemory 42, and aprocessor 43 and performs various control operations of thevehicle 3. Note that, in the present embodiment, asingle ECU 40 is provided, but a plurality of ECUs may be provided for the different functions. - The
communication interface 41 is an interface circuit for connecting theECU 40 to an internal vehicle network based on the CAN or other standard. TheECU 40 communicates with other vehicle-mounted devices through thecommunication interface 41. - The
memory 42, for example, has a volatile semiconductor memory (for example, a RAM) and nonvolatile semiconductor memory (for example, a ROM). Thememory 42 stores programs run in theprocessor 43, various data used when the various processings are performed by theprocessor 43, etc. - The
processor 43 has one or more CPUs (central processing units) and their peripheral circuits and performs various processing. Note that, theprocessor 43 may further have processing circuits such as arithmetic logic units or numerical calculation units. Thecommunication interface 41, thememory 42 and theprocessor 43 are connected to each other through signal wires. - In the present embodiment, the
ECU 40 functions as a machine learning device.FIG. 3 is a functional block diagram of theECU 40 ofFIG. 2 . TheECU 40 has a predictingpart 51, an updatingpart 52, and a partinformation acquiring part 53. The predictingpart 51, the updatingpart 52, and the partinformation acquiring part 53 are functional blocks realized by a program stored in thememory 42 of theECU 40 run by theprocessor 43 of theECU 40. - The predicting
part 51 predicts predetermined information using a machine learning model. The updatingpart 52 updates the machine learning model used in thevehicle 3. In the present embodiment, as the machine learning model, a neural network model outputting at least one output data (which is also called “a target variable”) from a plurality of input data (which are also called “explanatory variables”) is used. - First, referring to
FIG. 4 , a neural network model will be explained in outline.FIG. 4 shows one example of a neural network model having a simple configuration. - The circle marks in
FIG. 4 show artificial neurons. An artificial neuron is usually called a “node” or “unit” (in this Description, called a “node”). InFIG. 4 , L=1 indicates an input layer, L=2 and L=3 indicates hidden layers, and L=4 indicates an output layer. Note that, the hidden layers are also called “intermediate layers”. - In
FIG. 4 , x1 and x2 indicate nodes of the input layer (L=1) and output values from the nodes, while “y” indicates a node of the output layer (L=4) and its output values. Similarly, the z1 (L=2), z2 (L=2), and z3 (L=2) indicates nodes of the hidden layers (L=2) and the output values from the nodes, while zk (L=3) and z2 (L=3) indicate nodes of the hidden layers (L=3) and the output values from the nodes. - At the nodes of the input layer, inputs are output as they are. On the other hand, at the nodes of the hidden layer (L=2), the output values x1 and x2 of the nodes of the input layer are input. At the nodes of the hidden layers (L=2), corresponding weights “w” and biases “b” are used to calculate total input values “u”. For example, in
FIG. 4 , the total input values uk (L=2) calculated at the nodes shown by zk (L=2) (k=1, 2, 3) of the hidden layer (L=2) become as in the following equation (M is the number of nodes of the input layer). -
- Next, this total input values uk (L=2) are converted by the activation function “f” and are output as the output values zk (L=2) (=f(uk (L=2))) from the nodes shown by zk (L=2) of the hidden layers (L=2). On the other hand, the nodes of the hidden layer (L=3) receive as input the output values z1 (L=2), z2 (L=2), and z3 (L=2) of the nodes of the hidden layer (L=2). At the nodes of the hidden layer (L=3), the corresponding weights “w” and biases “b” are used to calculate the total input values “u” (=Σz·w+b). The total input values “u” are similarly converted by an activation function and are output from the nodes of the hidden layers (L=3) as the output values z1 (L=3) and z2 (L=3). The activation function is for example a Sigmoid function 6.
- Further, the node of the output layer (L=4) receives as input the output values z1 (L=3) and z2 (L=3) of the nodes of the hidden layer (L=3). At the node of the output layer, the corresponding weights “w” and biases “b” are used to calculate the total input value “u” (Σz·w+b) or only the corresponding weights “w” are used to calculate the total input value “u” (Σz+w). For example, at the node of the output layer, an identity function is used as the activation function. In this case, the total input value “u” calculated at the node of the output layer is output as it is as the output value “y” from the node of the output layer.
- The neural network model used in the
vehicle 3 is stored in thememory 42 of theECU 40 or another storage device provided at thevehicle 3. The predictingpart 51 inputs values of a plurality of input data (input values) to the neural network model to thereby make the neural network model output the value of at least one output data (output value). At this time, as the values of the input data, for example, output values of thesensors 32, command values from theECU 40 to theactuators 31, values obtained from outside thevehicle 3 through thecommunication module 33, values calculated at theECU 40, etc., are used. - The predicting
part 51 predicts predetermined information corresponding to the value of the output data which is output from the neural network model. The predetermined information predicted by the predictingpart 51 is used for control of thevehicle 3. Note that, the neural network model may be either of a regression model and a classification model. If the neural network model is a regression model, the output data, for example, is the output torque of the internal combustion engine, the concentration of predetermined constituent in the exhaust gas (HC, NOx, etc.), the state of charge (SOC) of the battery, etc. On the other hand, if the neural network model is a classification model, the output data is, for example, a result of judgment of abnormality of apredetermined sensor 32, a result of judgment of startup of the internal combustion engine in a hybrid vehicle, etc. As input data, a state quantity correlated with the output data is suitably selected. - In order to improve the precision of this neural network model, it is necessary to train the neural network model in advance. For this reason, for example, the neural network model is trained in the vehicle production plant etc.
- In training the neural network model, training data sets comprised of combinations of measured values of a plurality of input data and measured values of at least one output data corresponding to these measured values of the plurality of input data (truth data or teacher data) are used. The measured values of the input data and output data are for example acquired using an engine bench, etc., and the training data sets are created by combining the corresponding measured values.
- In training the neural network model, a large number of training data sets are used and the known back propagation method is used to repeatedly update the parameters of the neural network model (weights “w” and biases “b”) so that the difference between the values of the output data output by the neural network model and the measured values of the output data becomes smaller. As a result, the neural network model is trained and a trained neural network model is created. The created trained neural network model is loaded in the
ECU 40 of thevehicle 3 before shipment of thevehicle 3. That is, information of the trained neural network model (structure, weights “w”, biases “b”, etc., of the model) is stored in thememory 42 of theECU 40 or another storage device provided in thevehicle 3. Note that, the trained neural network model may be transmitted from theserver 2 to thevehicle 3 after thevehicle 3 is shipped. - In this regard, if a vehicle part such as the
actuator 31 or thesensor 32 deteriorates or breaks down along with use of thevehicle 3, the vehicle part is replaced. Further, sometimes a vehicle part is replaced in order to mount a higher performance vehicle part in thevehicle 3. - However, the neural network model loaded in the
ECU 40 is tailored to the properties of the vehicle part before replacement. For this reason, if the properties of a vehicle part change due to replacement of the vehicle part, the precision of prediction of the neural network model is liable to fall. - Therefore, in the present embodiment, the part
information acquiring part 53 detects replacement of a vehicle part and acquires identification information of the vehicle part after replacement. Further, if a vehicle part relating to input data of the neural network model is replaced with a vehicle part of a different configuration, the updatingpart 52 receives a new neural network model trained using training data sets corresponding to the vehicle part after replacement from theserver 2 and applies the new neural network model to thevehicle 3. - Specifically, if a vehicle part relating to input data of the neural network model is replaced with a vehicle part of a different configuration, the updating
part 52 transmits identification information of the vehicle part after replacement to theserver 2. If receiving identification information of the vehicle part after replacement from thevehicle 3, theprocessor 24 of theserver 2 transmits the new neural network model trained using training data sets corresponding to the vehicle part after replacement to thevehicle 3. If receiving the new neural network model from theserver 2, the updatingpart 52 applies the new neural network model to thevehicle 3. - As explained above, by updating the neural network model in accordance with replacement of a vehicle part, it is possible to keep the precision of prediction of the neural network model used in the
vehicle 3 from falling if a vehicle part is replaced. - Below, flow charts of
FIG. 5 andFIG. 6 will be used to explain in detail the above control.FIG. 5 is a flow chart showing a control routine of processing for updating a model. The present control routine is repeatedly performed at predetermined run intervals by theECU 40 of thevehicle 3. - First, at step S101, the part
information acquiring part 53 judges whether a replacement flag F is “1”. The replacement flag F is a flag which is set to “1” when a vehicle part is replaced. The initial value of the replacement flag F is zero. If at step S101 it is judged that the replacement flag F is zero, the present control routine proceeds to step S102. - At step S102, the part
information acquiring part 53 judges whether a vehicle part relating to input data of the neural network model has been replaced with a vehicle part of a different configuration. For example, if an output value of thesensor 32 is used as input data of the neural network model, thesensor 32 corresponds to a vehicle part relating to input data of the neural network model. Further, if a command value from theECU 40 to theactuator 31 is used as input data of the neural network model, theactuator 31 corresponds to a vehicle part relating to input data of the neural network model. Further, a “vehicle part of a different configuration” includes a vehicle part of a different manufacturer, part number, model number, etc., than the vehicle part before replacement. - The part
information acquiring part 53, for example, reads the identification information of the vehicle part (manufacturer's name, model number, serial number, manufacturing number, etc.,) to detect replacement of a vehicle part. Further, the partinformation acquiring part 53 may detect an output waveform, an operating voltage, etc., of a vehicle part to detect replacement of a vehicle part. Further, a worker may input replacement of a vehicle part to an input device (HMI etc.) of thevehicle 3, and the partinformation acquiring part 53 may detect replacement of a vehicle part based on input to the input device. - If at step S102 it is judged that a vehicle part has not been replaced, the present control routine ends. On the other hand, if at step S102 it is judged that a vehicle part has been replaced, the present control routine proceeds to step S103. At step S103, the part
information acquiring part 53 acquires the identification information of the vehicle part. - Next, at step S104, the part
information acquiring part 53 sets the replacement flag F to 4,r. - Next, at step S105, the updating
part 52 transmits the identification information of the vehicle part after replacement to theserver 2 through thecommunication module 33, the wireless base station 6, and the communication network 5. - After step S105, the present control routine proceeds to step S106. On the other hand, if at step S101 it is judged that the replacement flag F is “1”, the present control routine skips step S102 to step S105 and proceeds to step S106.
- At step S106, the updating
part 52 judges whether it has received a new neural network model from theserver 2. If it is judged that it has not received a new neural network model, the present control routine ends. On the other hand, if it is judged that it has received a new neural network model, the present control routine proceeds to step S107. - At step S107, the updating
part 52 applies the new neural network model to thevehicle 3. In other words, the updatingpart 52 updates the neural network model used in thevehicle 3 to the new neural network model. That is, the information on the neural network model stored in thememory 42 of theECU 40 or another storage device is rewritten. - Next, at step S108, the updating
part 52 sets the replacement flag F to zero. After step S108, the present control routine ends. -
FIG. 6 is a flow chart showing a control routine of processing for transmitting a model. The present control routine is repeatedly performed at predetermined run intervals by theprocessor 24 of theserver 2. - First, at step S201, the
processor 24 judges whether it has received identification information on a vehicle part from thevehicle 3. If it is judged that it has not received identification information on a vehicle part, the present control routine ends. On the other hand, if it is judged that it has received identification information on a vehicle part, the present control routine proceeds to step S202. - At step S202, the
processor 24 identifies the vehicle part after replacement based on the identification information of the vehicle part and identifies the neural network model corresponding to the vehicle part after replacement. Specifically, theprocessor 24 extracts a neural network model trained using training data sets corresponding to the vehicle part after replacement from the plurality of neural network models stored in thestorage device 22 of theserver 2. In the training data set corresponding to the vehicle part after replacement, the output value of the vehicle part after replacement, the command value to the vehicle part after replacement, etc., is used as one of the measured values of the input data. - Next, at step S203, the
processor 24 transmits the new neural network model identified at step S202 to thevehicle 3. After step S203, the present control routine ends. - Note that, at step S105 of
FIG. 5 , the updatingpart 52 may transmit the identification information of the vehicle part after replacement and the identification information of the vehicle 3 (vehicle name, model, etc.) to theserver 2 and, at step S203 ofFIG. 6 , theprocessor 24 of theserver 2 may transmit, as a new neural network model, a neural network model corresponding to the identification information of the vehicle part after replacement and the identification information of thevehicle 3 to thevehicle 3. By doing this, it is possible to apply a neural network model better suited to the properties of thevehicle 3 to thevehicle 3 and in turn possible to better keep the precision of prediction of the neural network model from falling. - In this case, the identification information of the
vehicle 3 is stored in thememory 42 of theECU 40 etc., in advance. Further, at theserver 2, a neural network model created for each type of vehicle, for example, a neural network model trained using training data sets created from measured values of input data and output data acquired using an engine bench etc., prepared for each of the types of vehicles, is stored in advance in thestorage device 22 of theserver 2. - Above, embodiments according to the present disclosure were explained, but the present disclosure is not limited to these embodiments and can be corrected and changed in various ways within the language of the claims. For example, the machine learning model used in the
vehicle 3 may be a machine learning model other than a neural network such as a random forest, a k-nearest neighbor method, a support vector machine, etc. - 1 machine learning system
- 2 server
- 21 communication interface
- 24 processor
- 3 vehicle
- 33 communication module
- 40 electronic control unit (ECU)
- 51 predicting part
- 52 updating part
- 53 part information acquiring part
Claims (3)
1. A machine learning device provided in an apparatus, comprising a processor configured to:
predict predetermined information with a machine learning model;
update the machine learning model;
detect an electronic signal indicative of a replacement of a first apparatus part;
acquire identification information of a second apparatus part which is installed instead of the first apparatus part after the replacement;
transmit the identification information of the second apparatus part to a server to identify a new machine learning model trained using training data sets corresponding to the second apparatus part after the replacement; and
apply the new machine learning model to the apparatus, in response to the first apparatus part relating to input data of the machine learning model having a different configuration than the second apparatus part.
2. The machine learning device according to claim 1 , wherein the processor is configured to transmit the identification information of the second apparatus part after replacement and identification information of the apparatus to the server, and the new machine learning model is a machine learning model corresponding to the identification information of the apparatus.
3. The machine learning device according to claim 1 , wherein the machine learning model is a neural network model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US17/690,725 US20220194394A1 (en) | 2020-08-20 | 2022-03-09 | Machine learning device and machine learning system |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2020-139396 | 2020-08-20 | ||
JP2020139396A JP6935837B1 (en) | 2020-08-20 | 2020-08-20 | Machine learning device and machine learning system |
US17/405,390 US11472420B2 (en) | 2020-08-20 | 2021-08-18 | Machine learning device and machine learning system |
US17/690,725 US20220194394A1 (en) | 2020-08-20 | 2022-03-09 | Machine learning device and machine learning system |
Related Parent Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US17/405,390 Continuation US11472420B2 (en) | 2020-08-20 | 2021-08-18 | Machine learning device and machine learning system |
Publications (1)
Publication Number | Publication Date |
---|---|
US20220194394A1 true US20220194394A1 (en) | 2022-06-23 |
Family
ID=77657903
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US17/405,390 Active US11472420B2 (en) | 2020-08-20 | 2021-08-18 | Machine learning device and machine learning system |
US17/690,725 Abandoned US20220194394A1 (en) | 2020-08-20 | 2022-03-09 | Machine learning device and machine learning system |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US17/405,390 Active US11472420B2 (en) | 2020-08-20 | 2021-08-18 | Machine learning device and machine learning system |
Country Status (4)
Country | Link |
---|---|
US (2) | US11472420B2 (en) |
JP (1) | JP6935837B1 (en) |
CN (1) | CN114077193A (en) |
DE (1) | DE102021118572A1 (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11720668B2 (en) * | 2021-10-11 | 2023-08-08 | Sift Science, Inc. | Systems and methods for accelerated detection and replacement of anomalous machine learning-based digital threat scoring ensembles and intelligent generation of anomalous artifacts for anomalous ensembles |
CN114483271B (en) * | 2021-12-30 | 2022-09-30 | 特斯联科技集团有限公司 | Vehicle exhaust waste heat recovery system based on artificial intelligence |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6546321B1 (en) * | 1999-08-25 | 2003-04-08 | Unisia Jecs Corporation | Method and apparatus for rewriting data of control unit for vehicle |
US20150015389A1 (en) * | 2013-07-15 | 2015-01-15 | Continental Automotive Systems, Inc. | Methods, systems and devices for recording and transmitting identification information of tire pressure monitoring sensors to a vehicle |
US9296266B1 (en) * | 2014-12-16 | 2016-03-29 | E-Lead Electronic Co., Ltd. | Method for correcting ID codes after installation of tire pressure sensors on a vehicle |
US20170028796A1 (en) * | 2015-07-28 | 2017-02-02 | Ford Global Technologies, Llc | System and method for managing tire pressure for a trailer |
US9898923B1 (en) * | 2016-08-31 | 2018-02-20 | Ecolink Intelligent Technology, Inc. | In-field sensor programming |
US20180216557A1 (en) * | 2015-07-22 | 2018-08-02 | Robert Bosch Gmbh | Method for ascertaining an accuracy of a torque transmitted by a belt-driven starter generator of an internal combustion engine to the internal combustion engine |
US20190205765A1 (en) * | 2017-12-29 | 2019-07-04 | Micron Technology, Inc. | Self-Learning in Distributed Architecture for Enhancing Artificial Neural Network |
US20190215339A1 (en) * | 2018-01-05 | 2019-07-11 | Byton Limited | System and method for enforcing security with a vehicle gateway |
US20200125930A1 (en) * | 2018-10-23 | 2020-04-23 | Hrl Laboratories, Llc | Artificial neural network and method of training an artificial neural network with epigenetic neurogenesis |
Family Cites Families (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH1185719A (en) * | 1997-09-03 | 1999-03-30 | Matsushita Electric Ind Co Ltd | Parameter estimation device |
JPH1182137A (en) * | 1998-02-09 | 1999-03-26 | Matsushita Electric Ind Co Ltd | Parameter estimating device |
JP2008286251A (en) | 2007-05-15 | 2008-11-27 | Toyota Motor Corp | Learning controller |
WO2018119423A1 (en) | 2016-12-22 | 2018-06-28 | Surround. Io Corporation | Method and system for providing artificial intelligence analytic (aia) services for performance prediction |
US20210390624A1 (en) | 2017-09-27 | 2021-12-16 | State Farm Mutual Automobile Insurance Company | Real Property Monitoring Systems and Methods for Risk Determination |
DE102017217444B4 (en) * | 2017-09-29 | 2024-03-07 | Volkswagen Ag | Method and system for updating a control model for automatic control of at least one mobile unit |
US10726638B2 (en) | 2017-12-21 | 2020-07-28 | Micron Technology, Inc. | Providing autonomous vehicle maintenance |
JP6956028B2 (en) * | 2018-02-22 | 2021-10-27 | ファナック株式会社 | Failure diagnosis device and machine learning device |
JP6477951B1 (en) | 2018-04-05 | 2019-03-06 | トヨタ自動車株式会社 | In-vehicle electronic control unit |
JP6741057B2 (en) * | 2018-11-01 | 2020-08-19 | トヨタ自動車株式会社 | Internal combustion engine control system, electronic control unit, server, and internal combustion engine control method |
CN109829479B (en) * | 2019-01-02 | 2022-06-21 | 大连理工大学 | Automatic classifier model information updating system for sensor and sensor replacing method |
JP7200694B2 (en) | 2019-01-23 | 2023-01-10 | 日本電信電話株式会社 | SENSOR NODE, SERVER DEVICE, IDENTIFICATION SYSTEM, METHOD AND PROGRAM |
CN113811473A (en) | 2019-03-29 | 2021-12-17 | 英特尔公司 | Autonomous vehicle system |
JP7514063B2 (en) | 2019-06-14 | 2024-07-10 | 株式会社シマノ | DETECTION APPARATUS, DETECTION METHOD, COMPUTER PROGRAM, AND STORAGE MEDIUM |
JP7432321B2 (en) * | 2019-08-23 | 2024-02-16 | 株式会社デンソーテン | Control program, control method and control device |
CN110996286A (en) * | 2019-11-28 | 2020-04-10 | 的卢技术有限公司 | Vehicle control method and system based on wireless sensor |
CN111539474B (en) * | 2020-04-23 | 2022-05-10 | 大连理工大学 | Classifier model transfer learning method |
JP7010343B1 (en) | 2020-08-20 | 2022-01-26 | トヨタ自動車株式会社 | Machine learning device |
-
2020
- 2020-08-20 JP JP2020139396A patent/JP6935837B1/en active Active
-
2021
- 2021-07-19 DE DE102021118572.6A patent/DE102021118572A1/en active Pending
- 2021-08-18 CN CN202110947766.0A patent/CN114077193A/en active Pending
- 2021-08-18 US US17/405,390 patent/US11472420B2/en active Active
-
2022
- 2022-03-09 US US17/690,725 patent/US20220194394A1/en not_active Abandoned
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6546321B1 (en) * | 1999-08-25 | 2003-04-08 | Unisia Jecs Corporation | Method and apparatus for rewriting data of control unit for vehicle |
US20150015389A1 (en) * | 2013-07-15 | 2015-01-15 | Continental Automotive Systems, Inc. | Methods, systems and devices for recording and transmitting identification information of tire pressure monitoring sensors to a vehicle |
US9296266B1 (en) * | 2014-12-16 | 2016-03-29 | E-Lead Electronic Co., Ltd. | Method for correcting ID codes after installation of tire pressure sensors on a vehicle |
US20180216557A1 (en) * | 2015-07-22 | 2018-08-02 | Robert Bosch Gmbh | Method for ascertaining an accuracy of a torque transmitted by a belt-driven starter generator of an internal combustion engine to the internal combustion engine |
US20170028796A1 (en) * | 2015-07-28 | 2017-02-02 | Ford Global Technologies, Llc | System and method for managing tire pressure for a trailer |
US9898923B1 (en) * | 2016-08-31 | 2018-02-20 | Ecolink Intelligent Technology, Inc. | In-field sensor programming |
US20190205765A1 (en) * | 2017-12-29 | 2019-07-04 | Micron Technology, Inc. | Self-Learning in Distributed Architecture for Enhancing Artificial Neural Network |
US20190215339A1 (en) * | 2018-01-05 | 2019-07-11 | Byton Limited | System and method for enforcing security with a vehicle gateway |
US20200125930A1 (en) * | 2018-10-23 | 2020-04-23 | Hrl Laboratories, Llc | Artificial neural network and method of training an artificial neural network with epigenetic neurogenesis |
Also Published As
Publication number | Publication date |
---|---|
JP6935837B1 (en) | 2021-09-15 |
JP2022035227A (en) | 2022-03-04 |
DE102021118572A1 (en) | 2022-02-24 |
CN114077193A (en) | 2022-02-22 |
US11472420B2 (en) | 2022-10-18 |
US20220055635A1 (en) | 2022-02-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20220194394A1 (en) | Machine learning device and machine learning system | |
US11675999B2 (en) | Machine learning device | |
US20120245791A1 (en) | Apparatus and method for predicting mixed problems with vehicle | |
US20120323343A1 (en) | Virtual sensor system and method | |
KR102118088B1 (en) | Method for real driving emission prediction using artificial intelligence technology | |
US20210383215A1 (en) | Vehicle, model training system and server | |
US11623652B2 (en) | Machine learning method and machine learning system | |
US11040721B2 (en) | Causal analytics for powertrain management | |
US11820398B2 (en) | Learning apparatus and model learning system | |
US20220058522A1 (en) | Model learning system, model learning method, and server | |
US20210192872A1 (en) | Model diagnosis device and model diagnosis system | |
US11377110B2 (en) | Machine learning device | |
KR102552699B1 (en) | Method for training artificial neural network for predicting trouble of vehicle, method for predicting trouble of vehicle using artificial neural network, and computing system performing the same | |
JP2022035222A (en) | Machine learning apparatus | |
Luo et al. | Agent-based real-time fault diagnosis | |
US20220292350A1 (en) | Model updating apparatus, model updating method, and model updating program | |
JP2022079938A (en) | Machine learning system | |
JP2022076257A (en) | Machine learning system, vehicle, and server | |
JP7276298B2 (en) | MODEL LEARNING SYSTEM, VEHICLE CONTROL DEVICE, AND MODEL LEARNING METHOD | |
Jurgen | Electronic engine control technologies | |
Mahmud et al. | A Service Model of Predictive Maintenance for Autonomous and Connected Vehicles Using 5G | |
JP2022007079A (en) | vehicle |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |
|
AS | Assignment |
Owner name: TOYOTA JIDOSHA KABUSHIKI KAISHA, JAPAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:YOKOYAMA, DAIKI;HAREYAMA, YOHEI;REEL/FRAME:059301/0944 Effective date: 20210623 |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NON FINAL ACTION MAILED |
|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION |