TCU-based random assignment water pump control method and system
By adopting a random assignment water pump control method based on TCU in the vehicle temperature control system, using a multi-layer structure control model and a random assignment mechanism to dynamically adjust the water pump and fan, the problem of insufficient flexibility in the traditional system is solved and the adaptation of temperature control is improved Sex and efficiency.
Patent Information
- Application Number
- CN202411554306.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-04
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional vehicle temperature control systems lack flexibility and adaptability and cannot effectively deal with complex and changing working conditions, resulting in limited control performance and user experience.
The random assignment water pump control method based on TCU is adopted, and two-dimensional structure data is generated by collecting on-board information, inputting a multi-layer structure control model, outputting a control strategy, and random assignment method assigns control ID or feedback ID to the actuator, and dynamically adjusting the water pump or fan , realize accurate temperature control.
It improves the adaptability, reliability and efficiency of the on-board temperature control system, enhances the flexibility and stability of the system, quickly responds to different working conditions, and achieves more accurate and efficient temperature control.
Smart Images

Figure CN119065356A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of deep learning technology, and more specifically, to a TCU-based random assignment water pump control method and system. Background Art
[0002] With the continuous development of automobile technology, vehicle control systems are becoming increasingly complex. Among them, temperature control, as one of the key factors affecting vehicle performance and passenger comfort, has received widespread attention.
[0003] Traditional vehicle temperature control systems use fixed control strategies, lack flexibility and adaptability, and cannot effectively cope with complex and changing working conditions. Although the application of TCU improves the real-time and targeted nature of temperature control, it still relies on preset control logic and rules, making it difficult to achieve intelligent vehicle temperature control, which limits the improvement of control performance and user experience. Summary of the invention
[0004] The present invention provides a TCU-based random assignment water pump control method and system, which generates No. 1 two-dimensional structure data by collecting vehicle-mounted information including temperature information, vehicle status information, actuator information and pipeline information, and the structure data includes a data matrix and a relationship matrix, and each unit of the data matrix represents a description text of an independent object (such as temperature, vehicle status, actuator, pipeline); then the No. 1 two-dimensional structure data is time-sorted to obtain No. 1 one-dimensional structure data, which is input into a control model including a multi-layer structure, and finally outputs a control strategy after processing and data transmission in the middle layer; the temperature control device assigns a control ID or a feedback ID to the actuator through a random assignment method according to the control strategy, and the actuator adjusts the water pump or the fan according to the corresponding control amount, so as to realize accurate control and optimization of the temperature in the vehicle, thereby improving the adaptability, reliability and efficiency of the system.
[0005] The present invention provides a TCU-based random assignment water pump control method, comprising the following steps:
[0006] Step 100, collecting vehicle information, the vehicle information includes temperature information, vehicle status information, actuator information and pipeline information;
[0007] Step 200, based on the collected vehicle information, obtain No. 1 two-dimensional structure data, the No. 1 two-dimensional structure data includes a data matrix and a relationship matrix, a unit of the data matrix represents data of an independent object, the independent objects include temperature, vehicle status, actuator, pipeline, a unit of the data matrix representing temperature only contains description text of temperature information, a unit of the data matrix representing vehicle status only contains description text of vehicle status information, a unit of the data matrix representing actuator only contains description text of actuator information, and a unit of the data matrix representing pipeline only contains description text of pipeline information;
[0008] Step 300, obtaining No. 1 one-dimensional structure data based on the sorting of No. 1 two-dimensional structure data, wherein the No. 1 one-dimensional structure data includes n data items sorted by time, and the t-th data item represents the No. 1 two-dimensional structure data generated by the vehicle-mounted information collected at the t-th moment;
[0009] Step 400, input the one-dimensional structure data No. 1 into the control model, the control model includes a first intermediate layer, a second intermediate layer, a first output layer, a third intermediate layer, and a second output layer, wherein the first intermediate layer inputs the two-dimensional structure data No. 1, and outputs the first intermediate representation data to the second intermediate layer; the second intermediate layer outputs the second intermediate representation data to the first output layer, and outputs the control strategy; the third intermediate layer inputs the one-dimensional structure data No. 1, and outputs the third intermediate representation data to the second output layer; and outputs the failure probability of the actuator.
[0010] In a preferred embodiment, the calculation formula of the first intermediate layer is as follows:
[0011] ;
[0012] in The first unit of the vth unit of the data matrix represents the data in the middle, represents the state weight parameter, and The vth and pth cells of the data matrix represent the vehicle information of independent objects, respectively. represents the set of cells of the data matrix that are associated with the vth cell of the data matrix, express The total number of cells in ;
[0013] ;
[0014] Where tanh represents the hyperbolic tangent function, represents adjustable parameters, The default value of is 1 / M. Represents a transpose operation;
[0015] ;
[0016] in represents the first intermediate representation data of the output, The data matrix represents the data in the middle of the v-th cell of the data matrix. It means that the first unit intermediate representation data of the data matrix of all the data matrix units in the data matrix are concatenated, and N represents the set of all the data matrix units. is the weight parameter, is the bias parameter, Represents the sigmoid function.
[0017] In a preferred embodiment, the calculation formula of the second intermediate layer is as follows:
[0018] ;
[0019] ;
[0020] ;
[0021] ;
[0022] in, , , represents the first, second, and third weight parameters, , , represents the first, second and third bias parameters, represents the dot product, where , and Respectively represent the first, second, and third intermediate states, It represents the first intermediate representation data output when the tth data item is input into the first intermediate layer. and Respectively represent the t-th and t-1-th second intermediate representation data, n≥t≥1, n represents the total number of data items of the one-dimensional structure data, t=1 , tanh is the hyperbolic tangent function, Represents the sigmoid function.
[0023] In a preferred embodiment, the calculation formula of the first output layer is as follows:
[0024] ;
[0025] in represents the first output vector, and its cth component value represents the probability value of the cth strategy. The strategy with the largest probability value is selected as the output. The strategy group contains all executable strategies. A strategy includes a group of control quantities of actuators, and the control quantity represents the flow rate of the water pump or the speed of the fan. represents the nth second intermediate representation data, represents the nth third intermediate representation data, n represents the total number of data items of the one-dimensional structure data, is the first output weight parameter, is the first output bias parameter, Represents a sigmoid function.
[0026] In a preferred embodiment, the calculation formula of the third intermediate layer is as follows:
[0027] ;
[0028] in represents the Lth third intermediate representation data, represents the L-1th third intermediate representation data, n≥L≥1, n represents the total number of data items of the one-dimensional structure data, represents the Lth data item of the one-dimensional structure data, and is the weight parameter, is the bias parameter and tanh is the hyperbolic tangent function.
[0029] In a preferred embodiment, the calculation formula of the second output layer is as follows:
[0030] ;
[0031] in represents the second output vector. The fth component of the vector represents the probability value of the fth actuator failing. If the probability value is greater than 0.5, it indicates a failure. Otherwise, it indicates no failure. represents the nth third intermediate representation data, is the weight parameter, is the bias parameter, Represents the sigmoid function.
[0032] In a preferred embodiment, only the first output layer is used during the control model training, and the training steps include:
[0033] Step 101, initializing the parameters of the control model;
[0034] Step 102: Observe the vehicle information at time e , the strategy executed at time e 、In-vehicle information at e+1 time , execution strategy Rewards ;
[0035] Step 103, then calculate the policy error:
[0036] ;
[0037] in represents the policy error at time e, represents the discount factor, , Represents the control model input The maximum probability value in the first output vector output when Represents the control model input The first output vector corresponding to the strategy The probability value of
[0038] Step 104, update the control model, the updated formula is as follows:
[0039] ;
[0040] , represents the step size of deep learning, Indicates delivery update;
[0041] Step 105, iterate steps 102-104 until the control model converges or the number of iterations reaches a set value, the default value of which is 30.
[0042] In a preferred embodiment, the reward function It is expressed as follows:
[0043] ;
[0044] in, Indicates the difference between the maximum temperature and the actual temperature. Indicates the difference between the maximum temperature and the target temperature, Indicates the actual power consumption of the actuator, Indicates the rated power consumption of the actuator.
[0045] In a preferred embodiment, the TCU random assignment method includes the following steps:
[0046] TCU sends UDS unified diagnostic service instructions to control the pump to report random numbers, the reporting range includes 0x01-0x0A;
[0047] The actuator self-checks whether the Flag is 1. When Flag=0, the actuator reports a random number 0x01-0x0A to the TCU and sets the reported value to its own control or feedback ID. When Flag=1, the actuator does not report a number.
[0048] After TCU sends the UDS report command, it waits for 5000ms to receive the report. If the report is not received within 5000ms, the ID assignment process ends directly.
[0049] If the TCU receives a report, it determines whether the received random number is a unique value. If the random number is a unique value, the TCU records the control message ID corresponding to the random number and sends a command through the ID to make the actuator Flag = 1. If the random number is not unique, it is skipped;
[0050] When all random number judgments are completed, return to step 1 and proceed to the next cycle. The range of random numbers reported in the second round includes 0x0B-0x14.
[0051] In a preferred embodiment, a TCU-based random assignment water pump control system includes the following modules:
[0052] A data acquisition module, used to start collecting two-dimensional structure data when the car is driving;
[0053] A data encoding module, used for encoding the collected two-dimensional structure data into one-dimensional structure data of time series;
[0054] The data processing module inputs the one-dimensional structure data into the control model and outputs the control quantity of the actuator;
[0055] The on-board control unit sends the actuator control quantity to the actuator. The actuator obtains the control ID or feedback ID of the water pump and fan it controls through a random assignment method, and controls the corresponding water pump or fan according to the control quantity corresponding to the control ID.
[0056] The beneficial effects of the present invention are:
[0057] 1. Improve the adaptability and flexibility of the vehicle temperature control system. By introducing a random assignment mechanism and dynamically generating control strategies for water pumps and fans, the control parameters and strategies can be adaptively adjusted according to real-time collected vehicle information and environmental changes, avoiding the limitations of traditional fixed control methods, enabling the system to quickly respond to different working conditions and achieve more accurate and efficient temperature control.
[0058] 2. Enhanced system reliability and robustness. A multi-layer control model is adopted to optimize and adjust the control strategy through data processing and transmission in the middle layer. At the same time, a randomly assigned feedback mechanism is introduced. The actuator dynamically adjusts the control amount according to the feedback ID to form a closed-loop control, which improves the stability and anti-interference ability of the system. In addition, the centralized control architecture based on TCU simplifies the system design, reduces the communication and computing burden, and improves the real-time performance and reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] Figure 1 It is a flow chart of a TCU-based random assignment water pump control method of the present invention. DETAILED DESCRIPTION
[0060] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the present specification. Various examples may omit, replace, or add various processes or components as needed. In addition, the features described in some examples may also be combined in other examples.
[0061] At least one embodiment of the present invention discloses a TCU-based random assignment water pump control method, such as Figure 1 As shown, the following steps are included:
[0062] Step 100, collecting vehicle information, the vehicle information includes temperature information, vehicle status information, actuator information and pipeline information;
[0063] Temperature information includes battery temperature, motor temperature, battery circuit radiator air inlet temperature, motor circuit radiator air inlet temperature, battery circuit coolant temperature, and motor circuit coolant temperature;
[0064] The vehicle status information includes the vehicle's driving or braking status and the vehicle's speed;
[0065] The actuator information includes the operating parameters of the battery circuit water pump, the motor circuit water pump, the battery circuit fan and the motor circuit fan. The operating parameters include speed and power.
[0066] The pipeline information includes the pressure values of the inlet and outlet of the battery circuit and the motor circuit pipeline and the flow rate of the coolant;
[0067] The vehicle information is collected at fixed time intervals while the vehicle is driving.
[0068] Step 200, based on the collected vehicle information, obtain No. 1 two-dimensional structure data, the No. 1 two-dimensional structure data includes a data matrix and a relationship matrix, a unit of the data matrix represents data of an independent object, the independent objects include temperature, vehicle status, actuator, pipeline, a unit of the data matrix representing temperature only contains description text of temperature information, a unit of the data matrix representing vehicle status only contains description text of vehicle status information, a unit of the data matrix representing actuator only contains description text of actuator information, and a unit of the data matrix representing pipeline only contains description text of pipeline information;
[0069] The element in the i-th row and j-th column of the relationship matrix represents the association between the i-th unit of the data matrix and the independent object represented by the j-th unit. If there is an association, the value of this element of the relationship matrix is 1, otherwise it is 0;
[0070] The relationship between vehicle status and temperature refers to: the vehicle status can directly affect the temperature of the motor and battery;
[0071] The relationship between temperature and actuator means that the actuator executes the water pump and fan in the corresponding circuit to cool down according to the temperature of the battery and motor;
[0072] The association between pipelines and actuators means that: each circuit of the actuator contains pipelines;
[0073] Step 300, obtaining No. 1 one-dimensional structure data based on the sorting of No. 1 two-dimensional structure data, wherein the No. 1 one-dimensional structure data includes n data items sorted by time, and the t-th data item represents the No. 1 two-dimensional structure data generated by the vehicle-mounted information collected at the t-th moment;
[0074] Step 400, input the one-dimensional structure data No. 1 into the control model, the control model includes a first intermediate layer, a second intermediate layer, a first output layer, a third intermediate layer, and a second output layer, wherein the first intermediate layer inputs the two-dimensional structure data No. 1, and outputs the first intermediate representation data to the second intermediate layer; the second intermediate layer outputs the second intermediate representation data to the first output layer, and outputs the control strategy; the third intermediate layer inputs the one-dimensional structure data No. 1, and outputs the third intermediate representation data to the second output layer; and outputs the failure probability of the actuator.
[0075] In one embodiment of the present invention, the execution process of the actuator is as follows:
[0076] The actuator controls the corresponding water pump or fan according to the control quantity corresponding to the control ID;
[0077] If the control quantity indicates that the water pump or fan needs to be turned on, the actuator will send an opening instruction to the water pump or fan with the corresponding ID, and set the appropriate speed and power parameters according to the control quantity;
[0078] If the control quantity indicates that the water pump or fan needs to be shut down, the actuator will send a shutdown command to the water pump or fan with the corresponding ID;
[0079] If the control quantity indicates that the pump or fan needs to adjust its operating parameters, the actuator will send an adjustment instruction to the pump or fan with the corresponding ID and update its speed and power settings;
[0080] The actuator monitors the status of the controlled water pump and fan according to the feedback ID and obtains their actual working parameters, such as speed, power, and whether they are working normally;
[0081] The actuator feeds back the monitored status information of the water pump and fan to the control model so that the control model can dynamically adjust the control strategy according to the feedback;
[0082] The actuator repeats the above steps and controls the water pump and fan in real time according to the instructions of the control model to achieve the purpose of regulating the temperature of the battery and motor.
[0083] In one embodiment of the present invention, the TCU random assignment method includes the following steps:
[0084] TCU sends UDS unified diagnostic service instructions to control the pump to report random numbers, the reporting range includes 0x01-0x0A;
[0085] The actuator self-checks whether the Flag is 1. When Flag=0, the actuator reports a random number 0x01-0x0A to the TCU and sets the reported value to its own control or feedback ID. When Flag=1, the actuator does not report a number.
[0086] After TCU sends the UDS report command, it waits for 5000ms to receive the report. If the report is not received within 5000ms, the ID assignment process ends directly.
[0087] If the TCU receives the report, it determines whether the received random number is a unique value. If the random number is a unique value, the TCU records the control message ID corresponding to the random number and sends a command through the ID to make the actuator Flag = 1; if the random number is not unique, it is skipped;
[0088] When all random number judgments are completed, return to step 1 and proceed to the next cycle. The range of the second round of random number reporting includes 0x0B-0x14;
[0089] In one embodiment of the present invention, when the ID allocation is completed, a certain actuator fails, and after replacing a new actuator, the steps of re-allocating the ID are as follows:
[0090] TCU sends the recorded control message ID. If there is no feedback message ID, it is the control or feedback message ID of the replaced actuator.
[0091] TCU reads the flags of all actuators in the system through UDS;
[0092] When a certain actuator Flag=0 is read, the actuator is an updated actuator, and the TCU uses the UDS command to report the actuator number and change the actuator ID; when all actuator Flags=1, the TCU directly sets all actuator Flags in the system to 0, clears the ID, and re-enters the random report number assignment ID process;
[0093] The actuator self-checks whether the Flag is 1. When Flag=0, the actuator reports to the TCU that the actuator ID has been changed and set to its own ID. The TCU then sends a command through the ID to make the pump Flag=1. When Flag=1, the actuator does not report.
[0094] In one embodiment of the present invention, the random assignment method includes:
[0095] Save software development costs;
[0096] Reduce the difficulty of production software management and differentiation, improve production rhythm, and reduce production line costs;
[0097] Reduce the difficulty of managing and distinguishing end users, reduce the difficulty of production scheduling, and reduce costs;
[0098] If any actuator fails, you can directly replace any actuator and then configure the ID, which reduces terminal inventory and reduces inventory pressure;
[0099] Save hardware costs. This method does not require external mechanical structure support and can be run by connecting to TCU.
[0100] In one embodiment of the present invention, the calculation formula of the first intermediate layer is as follows:
[0101] ;
[0102] in The first unit of the vth unit of the data matrix represents the data in the middle, represents the state weight parameter, and The vth and pth cells of the data matrix represent the vehicle information of independent objects, respectively. represents the set of cells of the data matrix that are associated with the vth cell of the data matrix, express The total number of cells in ;
[0103] ;
[0104] Where tanh represents the hyperbolic tangent function, represents adjustable parameters, The default value of is 1 / M. Represents a transpose operation;
[0105] ;
[0106] in represents the first intermediate representation data of the output, The data matrix represents the data in the middle of the v-th cell of the data matrix. It means that the first unit intermediate representation data of the data matrix of all the data matrix units in the data matrix are concatenated, and N represents the set of all the data matrix units. is the weight parameter, is the bias parameter, Represents the sigmoid function.
[0107] In one embodiment of the present invention, the calculation formula of the second intermediate layer is as follows:
[0108] ;
[0109] ;
[0110] ;
[0111] ;
[0112] in, , , represents the first, second, and third weight parameters, , , represents the first, second and third bias parameters, represents the dot product, where , and Respectively represent the first, second, and third intermediate states, It represents the first intermediate representation data output when the tth data item is input into the first intermediate layer. and Respectively represent the t-th and t-1-th second intermediate representation data, n≥t≥1, n represents the total number of data items of the one-dimensional structure data, t=1 , tanh is the hyperbolic tangent function, Represents the sigmoid function.
[0113] In one embodiment of the present invention, the calculation formula of the third intermediate layer is as follows:
[0114] ;
[0115] in represents the Lth third intermediate representation data, represents the L-1th third intermediate representation data, n≥L≥1, n represents the total number of data items of the one-dimensional structure data, represents the Lth data item of the one-dimensional structure data, and is the weight parameter, is the bias parameter and tanh is the hyperbolic tangent function.
[0116] In one embodiment of the present invention, the calculation formula of the first output layer is as follows:
[0117] ;
[0118] in represents the first output vector, and its cth component value represents the probability value of the cth strategy. The strategy with the largest probability value is selected as the output. The strategy group contains all executable strategies. A strategy includes a group of control quantities of actuators, and the control quantity represents the flow rate of the water pump or the speed of the fan. represents the nth second intermediate representation data, represents the nth third intermediate representation data, n represents the total number of data items of the one-dimensional structure data, is the first output weight parameter, is the first output bias parameter, Represents a sigmoid function.
[0119] In one embodiment of the present invention, the control amount of the controller is divided into 5 gears, and the value between 0 and 0.2 represents the first gear flow rate or speed, between 0.2 and 0.4 represents the second gear flow rate or speed, between 0.4 and 0.6 represents the third gear flow rate or speed, between 0.6 and 0.8 represents the fourth gear flow rate or speed, and between 0.8 and 1 represents the fifth gear flow rate or speed;
[0120] In one embodiment of the present invention, the calculation formula of the second output layer is as follows:
[0121] ;
[0122] in represents the second output vector. The fth component of the vector represents the probability value of the fth actuator failing. If the probability value is greater than 0.5, it indicates a failure. Otherwise, it indicates no failure. represents the nth third intermediate representation data, is the weight parameter, is the bias parameter, Represents the sigmoid function
[0123] In one embodiment of the present invention, only the first output layer is used during the control model training, and the training steps include:
[0124] Step 101, initializing the parameters of the control model;
[0125] Step 102: Observe the vehicle information at time e , the strategy executed at time e 、In-vehicle information at e+1 time , execution strategy Rewards ;
[0126] Step 103, then calculate the policy error:
[0127] ;
[0128] in represents the policy error at time e, represents the discount factor, , Represents the control model input The maximum probability value in the first output vector output when Represents the control model input The first output vector corresponding to the strategy The probability value of
[0129] Step 104, update the control model, the updated formula is as follows:
[0130] ;
[0131] , represents the step size of deep learning, Indicates delivery update;
[0132] Step 105, iterating steps 102-104 until the control model converges or the number of iterations reaches a set value, the default value of which is 30;
[0133] Reward Function It is expressed as follows:
[0134] ;
[0135] in, Indicates the difference between the maximum temperature and the actual temperature. Indicates the difference between the maximum temperature and the target temperature, Indicates the actual power consumption of the actuator, Indicates the rated power consumption of the actuator.
[0136] In one embodiment of the present invention, the temperature description text, the vehicle status description text, the actuator description text, and the pipeline description text are subjected to feature engineering and then input into the model;
[0137] The temperature description text, vehicle status description text, actuator description text, and pipeline description text are used as a feature engineering method through the WordEmbedding algorithm;
[0138] In one embodiment of the present invention, a TCU-based random assignment water pump control system includes the following modules:
[0139] A data acquisition module, used to start collecting two-dimensional structure data when the car is driving;
[0140] A data encoding module, used for encoding the collected two-dimensional structure data into one-dimensional structure data of time series;
[0141] The data processing module inputs the one-dimensional structure data into the control model and outputs the control quantity of the actuator;
[0142] The on-board control unit sends the actuator control quantity to the actuator. The actuator obtains the control ID or feedback ID of the water pump and fan it controls through a random assignment method, and controls the corresponding water pump or fan according to the control quantity corresponding to the control ID.
[0143] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation mode. The above-mentioned specific implementation mode is merely illustrative and not restrictive. Under the guidance of this embodiment, ordinary technicians in this field can also make more forms of equivalent embodiments, all of which are within the protection of this embodiment.
Claims
1. A TCU-based random assignment water pump control method, characterized in that: The following steps are involved: Step 100, collecting vehicle information, the vehicle information includes temperature information, vehicle status information, actuator information and pipeline information; Step 200, based on the collected vehicle information, obtain No. 1 two-dimensional structure data, the No. 1 two-dimensional structure data includes a data matrix and a relationship matrix, a unit of the data matrix represents data of an independent object, the independent objects include temperature, vehicle status, actuator, pipeline, a unit of the data matrix representing temperature only contains description text of temperature information, a unit of the data matrix representing vehicle status only contains description text of vehicle status information, a unit of the data matrix representing actuator only contains description text of actuator information, and a unit of the data matrix representing pipeline only contains description text of pipeline information; Step 300, obtaining No. 1 one-dimensional structure data based on the sorting of No. 1 two-dimensional structure data, wherein the No. 1 one-dimensional structure data includes n data items sorted by time, and the t-th data item represents the No. 1 two-dimensional structure data generated by the vehicle-mounted information collected at the t-th moment; Step 400, inputting the first one-dimensional structure data into the control model, the control model comprising a first intermediate layer, a second intermediate layer, a first output layer, a third intermediate layer, and a second output layer, wherein the first intermediate layer inputs the first two-dimensional structure data and outputs the first intermediate representation data to the second intermediate layer; The second intermediate layer outputs the second intermediate representation data to the first output layer, and outputs the control strategy; The third intermediate layer inputs the one-dimensional structure data, outputs the third intermediate representation data to the second output layer, and outputs the failure probability of the actuator.
2. A TCU-based random assignment water pump control method according to claim 1, characterized in that: The calculation formula for the first intermediate layer is as follows: ; in The first unit of the vth unit of the data matrix represents the data in the middle, represents the state weight parameter, and The vth and pth cells of the data matrix represent the vehicle information of independent objects, respectively. represents the set of cells of the data matrix that are associated with the vth cell of the data matrix, express The total number of cells in ; ; Where tanh represents the hyperbolic tangent function, represents adjustable parameters, The default value of is 1 / M. Represents a transpose operation; ; in represents the first intermediate representation data of the output, The data matrix represents the data in the middle of the v-th cell of the data matrix. It means that the first unit intermediate representation data of the data matrix of all the data matrix units in the data matrix are concatenated, and N represents the set of all the data matrix units. is the weight parameter, is the bias parameter, Represents the sigmoid function.
3. A TCU-based random assignment water pump control method according to claim 2, characterized in that: The calculation formula for the second intermediate layer is as follows: ; ; ; ; in, , , represents the first, second, and third weight parameters, , , represents the first, second and third bias parameters, represents the dot product, where , and Respectively represent the first, second, and third intermediate states, It represents the first intermediate representation data output when the tth data item is input into the first intermediate layer. and They represent the t-th and t-1-th second intermediate representation data, respectively. n≥t≥1, where n represents the total number of data items of the one-dimensional structure data. When t=1 , tanh is the hyperbolic tangent function, Represents the sigmoid function.
4. The TCU-based random assignment water pump control method according to claim 3 is characterized in that: The calculation formula of the first output layer is as follows: ; in represents the first output vector, and its cth component value represents the probability value of the cth strategy. The strategy with the largest probability value is selected as the output. The strategy group contains all executable strategies. A strategy includes a group of control quantities of actuators, and the control quantity represents the flow rate of the water pump or the speed of the fan. represents the nth second intermediate representation data, represents the nth third intermediate representation data, n represents the total number of data items of the one-dimensional structure data, is the first output weight parameter, is the first output bias parameter, Represents a sigmoid function.
5. The TCU-based random assignment water pump control method according to claim 1 is characterized in that: The calculation formula for the third intermediate layer is as follows: ; in represents the Lth third intermediate representation data, represents the L-1th third intermediate representation data, n≥L≥1, n represents the total number of data items of the one-dimensional structure data, represents the Lth data item of the one-dimensional structure data, and is the weight parameter, is the bias parameter and tanh is the hyperbolic tangent function.
6. A TCU-based random assignment water pump control method according to claim 5, characterized in that: The calculation formula of the second output layer is as follows: ; in represents the second output vector. The fth component of the vector represents the probability value of the fth actuator failing. If the probability value is greater than 0.5, it indicates a failure. Otherwise, it indicates no failure. represents the nth third intermediate representation data, is the weight parameter, is the bias parameter, Represents the sigmoid function.
7. The TCU-based random assignment water pump control method according to claim 1 is characterized in that: Only the first output layer is used when controlling the model training. The training steps include: Step 101, initializing the parameters of the control model; Step 102: Observe the vehicle information at time e , the strategy executed at time e 、In-vehicle information at e+1 time , execution strategy Rewards ; Step 103, then calculate the policy error: ; in represents the policy error at time e, represents the discount factor, , Represents the control model input The maximum probability value in the first output vector output when Represents the control model input The first output vector corresponding to the strategy The probability value of Step 104, update the control model, the updated formula is as follows: ; , represents the step size of deep learning, Indicates delivery update; Step 105, iterate steps 102-104 until the control model converges or the number of iterations reaches a set value, the default value of which is 30.
8. The TCU-based random assignment water pump control method according to claim 7 is characterized in that: Reward Function It is expressed as follows: ; in, Indicates the difference between the maximum temperature and the actual temperature. Indicates the difference between the maximum temperature and the target temperature, Indicates the actual power consumption of the actuator, Indicates the rated power consumption of the actuator.
9. The TCU-based random assignment water pump control method according to claim 1, characterized in that: The TCU random assignment method includes the following steps: TCU sends UDS unified diagnostic service instructions to control the pump to report random numbers, the reporting range includes 0x01-0x0A; The actuator self-checks whether the Flag is 1. When Flag=0, the actuator reports a random number 0x01-0x0A to the TCU and sets the reported value to its own control or feedback ID. When Flag=1, the actuator does not report a number. After TCU sends the UDS report command, it waits for 5000ms to receive the report. If the report is not received within 5000ms, the ID assignment process ends directly. If the TCU receives a report, it determines whether the received random number is a unique value. If the random number is a unique value, the TCU records the control message ID corresponding to the random number and sends a command through the ID to make the actuator Flag = 1. If the random number is not unique, it is skipped; When all random number judgments are completed, return to step 1 and proceed to the next cycle. The range of random numbers reported in the second round includes 0x0B-0x14.
10. A TCU-based random assignment water pump control system, characterized in that: A TCU-based random assignment water pump control method for executing any one of claims 1-9, a TCU-based random assignment water pump control system comprising the following modules: A data acquisition module, used to start collecting two-dimensional structure data when the car is driving; A data encoding module, used for encoding the collected two-dimensional structure data into one-dimensional structure data of time series; The data processing module inputs the one-dimensional structure data into the control model and outputs the control quantity of the actuator; The on-board control unit sends the actuator control quantity to the actuator. The actuator obtains the control ID or feedback ID of the water pump and fan it controls through a random assignment method, and controls the corresponding water pump or fan according to the control quantity corresponding to the control ID.
Citation Information
Patent Citations
Electric vehicle charging control strategy optimization method and system based on self-learning
CN111532150A
Neural network-based networked vehicle decision control model establishment and evaluation method
CN111967087A
Vehicle control method based on reinforcement learning control strategy in hybrid fleet
CN112162555A
Intelligent dispatching control method and system for high-speed train
CN113537603A
Hybrid vehicle thermal management strategy generation method based on deep reinforcement learning
CN115840987A
Cited By
Temporary wireless bridge adjusting method and system for capital construction site
CN119136228A
A temporary wireless bridge adjustment method and system for infrastructure construction site
CN119136228B
