Intelligent control method and system for throttle based on internet of things

Abstract

The application relates to the technical fields of Internet of Things and automobile engine control, and discloses a method and system for intelligent control of a throttle valve based on the Internet of Things, which comprises the following steps: obtaining sensor data and processing the same to obtain a real-time driving state recognition result; determining a calculation resource allocation requirement according to the real-time driving state recognition result, and adjusting the power of an edge processor according to the requirement to obtain an adjusted power level; fusing real-time data according to the adjusted power level, calculating a throttle valve opening degree driving signal, and generating a throttle valve actuator instruction; monitoring the execution response time of the instruction, and adjusting the power again according to the response time to obtain a final balanced throttle valve control signal; and determining an updated driving state recognition result according to the final balanced throttle valve control signal. The method solves the problem that the control precision and energy efficiency of the prior art are difficult to be considered simultaneously under complex working conditions by constructing a double-layer dynamic power adjustment mechanism combining feedforward and feedback.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things technology and automobile engine control technology, and particularly relates to a throttle valve intelligent control method and system based on Internet of Things. BACKGROUND

[0002] At present, the application of Internet of Things technology in the field of automobile engine control is increasingly widespread. Through real-time data acquisition and processing, it can significantly improve vehicle performance and energy efficiency, and is an important part of advanced combustion optimization technology.

[0003] In one prior art, engine throttle valve control, as a key link affecting combustion efficiency, often relies on edge processors for calculation and signal generation, but its power regulation and computing resource allocation strategy is relatively fixed. This mode is difficult to achieve accurate balance between power consumption and computing performance when facing different driving scenarios. For example, when accelerating quickly, if the processor does not timely increase power, the throttle opening response may lag due to calculation delay, affecting vehicle power output; conversely, if the power is too high at idle speed, unnecessary energy consumption will be caused. Therefore, the existing method cannot dynamically adjust resources according to the working condition, so that the system cannot achieve both control accuracy and efficiency when quickly switching between driving states.

[0004] In summary, the prior art has the technical problem that the accuracy and efficiency of throttle valve control cannot be considered. SUMMARY

[0005] The present application provides a throttle valve intelligent control method and system based on Internet of Things to solve the technical problem that the accuracy and efficiency of throttle valve control cannot be considered in the prior art.

[0006] In a first aspect, to solve the above technical problem, the present application provides a throttle valve intelligent control method based on Internet of Things, comprising:

[0007] Obtaining sensor data and processing to obtain real-time driving state recognition results;

[0008] According to the real-time driving state recognition result, a preset threshold rule is used to determine the computing resource allocation requirement;

[0009] According to the computing resource allocation requirement, the power of the edge processor is adjusted to obtain an adjusted power level;

[0010] According to the adjusted power level, and combining real-time monitoring data, a throttle opening driving signal is calculated, and a throttle actuator instruction is generated;

[0011] monitoring an execution response time of the throttle actuator instruction, and adjusting the edge processor power again according to the execution response time to obtain a final balanced throttle control signal;

[0012] According to the final balanced throttle control signal, the latest sensor data is preprocessed and classified to determine an updated driving state recognition result.

[0013] Preferably, the acquisition of sensor data and processing to obtain a real-time driving state recognition result comprises:

[0014] Acquiring an original data stream of engine speed and throttle pedal position;

[0015] Timestamp alignment processing is performed on the original data stream to obtain a synchronous data stream;

[0016] Filtering and denoising processing is performed on the synchronous data stream to obtain a smooth data sequence;

[0017] The data in the smooth data sequence is normalized and weighted to obtain a weighted feature value, and the weighted feature value is taken as the real-time driving state recognition result.

[0018] Preferably, the calculation resource allocation requirement is determined according to the real-time driving state recognition result by using a preset threshold rule, comprising:

[0019] If the real-time driving state recognition result exceeds a preset dynamic scene threshold, an indication of needing to improve the calculation resource allocation is generated as the calculation resource allocation requirement.

[0020] Preferably, the edge processor power is adjusted according to the calculation resource allocation requirement to obtain an adjusted power level, comprising:

[0021] The voltage and frequency parameters of the edge processor are adjusted by using a dynamic voltage and frequency scaling method to obtain adjusted voltage and frequency parameters;

[0022] The power output of the edge processor is reconfigured according to the adjusted voltage and frequency parameters to obtain the adjusted power level.

[0023] Preferably, the throttle opening degree driving signal is calculated according to the adjusted power level and fused with real-time monitoring data, and a throttle actuator instruction is generated, comprising:

[0024] The current energy consumption index is obtained from the adjusted power level;

[0025] If the current energy consumption index is lower than a preset energy consumption threshold, the driving intensity index is obtained and fused with engine speed data and throttle pedal position data.

[0026] inputting the driving intensity index into a pre-trained logistic regression model to obtain a probability value of resource promotion demand;

[0027] comparing the probability value with a preset scheduling threshold to select one from a plurality of preset scheduling schemes as an optimized resource scheduling scheme;

[0028] According to the optimized resource scheduling scheme, and using a pre-established multi-dimensional mapping table defining the relationship between sensor data and throttle opening, the throttle opening driving signal is calculated;

[0029] The throttle opening driving signal is modulated to obtain the throttle actuator command.

[0030] Preferably, the execution response time of the throttle actuator command is monitored, and the edge processor power is adjusted again according to the execution response time to obtain a final balanced throttle control signal, comprising:

[0031] The execution response time of the throttle actuator command is continuously monitored to obtain a response time data set;

[0032] If the response time in the response time data set exceeds the preset response time threshold, the power output is recalculated to obtain a new power target value as the first power output data set;

[0033] According to the first power output data set, the performance level of the edge processor is adjusted;

[0034] Under the adjusted performance level, the initial throttle control signal is recalculated based on the latest sensor real-time data;

[0035] The initial throttle control signal is fused with the actual throttle opening feedback value at the previous time to generate the final balanced throttle control signal.

[0036] Preferably, the final balanced throttle control signal is used to pre-process and classify the latest sensor data to determine an updated driving state recognition result, comprising:

[0037] According to the final balanced throttle control signal, the sensor data is obtained;

[0038] The sensor data is subjected to signal noise verification to obtain a verification result;

[0039] According to the verification result, the sensor data is subjected to conditional filtering processing to obtain purified sensor data;

[0040] The purified sensor data is normalized to obtain to-be-classified data;

[0041] The to-be-classified data is input into a pre-established driving state model for classification processing to obtain the updated driving state recognition result.

[0042] In a second aspect, the present application provides a throttle valve intelligent control system based on Internet of Things, comprising:

[0043] A state recognition module is configured to acquire sensor data and process the sensor data to obtain a real-time driving state recognition result;

[0044] A demand determination module is configured to determine a computing resource allocation demand according to the real-time driving state recognition result and by using a preset threshold rule;

[0045] A power adjustment module is configured to adjust the power of an edge processor according to the computing resource allocation demand to obtain an adjusted power level;

[0046] An instruction generation module is configured to calculate a throttle valve opening degree driving signal according to the adjusted power level and by fusing real-time monitoring data, and generate a throttle valve actuator instruction;

[0047] A feedback control module is configured to monitor the execution response time of the throttle valve actuator instruction, and adjust the power of the edge processor again according to the execution response time to obtain a final balanced throttle valve control signal;

[0048] A state update module is configured to pre-process and classify the latest sensor data according to the final balanced throttle valve control signal, and determine an updated driving state recognition result.

[0049] In a third aspect, the present application further provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the throttle valve intelligent control method based on Internet of Things according to any one of the above.

[0050] In a fourth aspect, the present application further provides a computer readable storage medium, comprising a stored computer program, wherein the computer program controls the device where the computer readable storage medium is located to execute the throttle valve intelligent control method based on Internet of Things according to any one of the above when the computer program is running.

[0051] Compared with the prior art, the present application has the following beneficial effects:

[0052] (1) The present application can accurately identify the real-time driving state of the vehicle under different working conditions such as idle, acceleration or cruising by fusing multi-dimensional sensor data such as engine speed and throttle pedal position and using signal filtering to remove noise. This multi-dimensional accurate perception process solves the problem that the prior art only relies on single feedback and cannot fully grasp the dynamic working condition, providing scientific and reliable data basis for subsequent intelligent control.

[0053] (2) The present application can adjust the edge processor performance in advance by using dynamic voltage frequency scaling method under high dynamic scene by constructing the feedforward control path of "state recognition-demand judgment-power regulation". This predictive resource allocation mechanism deduces and solves the response lag problem caused by fixed control parameters in the prior art, improving the active adjustment ability and response speed of the control system.

[0054] (3) The present application generates balanced control signals by establishing a closed-loop feedback adjustment mechanism based on the response time of the actuator instruction after the initial adjustment and making a secondary dynamic adjustment to the processor power. This method changes the control strategy from passive open-loop execution to active closed-loop optimization, ensuring the control accuracy and energy balance of the system under complex working conditions and improving the operation efficiency of the entire throttle control system. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 is a flowchart of a throttle intelligent control method based on Internet of Things provided by the first embodiment of the present application;

[0056] Figure 2 is a structure schematic diagram of a throttle intelligent control system based on Internet of Things provided by the second embodiment of the present application. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0058] Referring to Figure 1 , the first embodiment of the present application provides a throttle intelligent control method based on Internet of Things, comprising the following steps:

[0059] S11, acquiring sensor data and processing to obtain real-time driving state recognition result;

[0060] S12, determining the calculation resource allocation demand according to the real-time driving state recognition result by using the preset threshold rule;

[0061] S13, adjusting the edge processor power according to the computing resource allocation requirement, to obtain an adjusted power level;

[0062] S14, calculating a throttle opening degree driving signal according to the adjusted power level and fusing real-time monitoring data, and generating a throttle actuator instruction;

[0063] S15, monitoring an execution response time of the throttle actuator instruction, and adjusting the edge processor power again according to the execution response time, to obtain a final balanced throttle control signal;

[0064] S16, pre-processing and classifying the latest sensor data according to the final balanced throttle control signal, to determine an updated driving state recognition result.

[0065] In step S11, sensor data is acquired and processed to obtain a real-time driving state recognition result, including:

[0066] Raw data streams of engine speed and throttle pedal position are acquired;

[0067] Timestamp alignment processing is performed on the raw data streams to obtain synchronized data streams;

[0068] Filtering and denoising processing is performed on the synchronized data streams to obtain smooth data sequences;

[0069] Data in the smooth data sequences is normalized and weighted to obtain weighted feature values, and the weighted feature values are taken as the real-time driving state recognition result.

[0070] In an implementation manner, a speed sensor deployed on an engine and a position sensor installed on a throttle pedal are used to respectively collect raw data of engine speed and raw data of throttle pedal position, to constitute the raw data streams.

[0071] It should be noted that due to differences in different sensor hardware, there may be millisecond-level deviations in timestamps of data points in the raw data streams. To ensure that subsequent analysis is based on a consistent time reference, the system performs timestamp alignment processing on the raw data streams, and combines speed data and pedal data according to a unified time axis to obtain the synchronized data streams.

[0072] It should be noted that to eliminate high-frequency noise in the synchronized data streams caused by engine vibration or signal interference, the system uses a Kalman filtering algorithm to perform filtering processing on the synchronized data streams.

[0073] Exemplarily, an original engine speed data can have a fluctuation of ±50 RPM due to vibration, and after Kalman filtering, the fluctuation can be reduced to ±10 RPM, obtaining more stable data. In another implementation, a low-pass filtering method can also be combined to further eliminate high-frequency interference and obtain the smoothed data sequence.

[0074] In an implementation, the system first normalizes the engine speed and throttle pedal position data in the smoothed data sequence before data fusion, mapping values in different physical units to a unified dimensionless interval; then, the weighted feature value is obtained by weighted calculation on the normalized engine speed and throttle pedal position.

[0075] It should be noted that the normalization process can be realized by using the min-max normalization method, i.e., linearly mapping the original data to the [0, 1] interval according to the preset maximum and minimum values. The maximum and minimum values are determined in advance according to the design specifications and physical operation upper limit of the engine or sensor.

[0076] It should be noted that the weight value used in the weighted calculation is determined based on offline optimization of a historical data set containing multiple known driving states (e.g., idle, smooth driving, and rapid acceleration). Exemplarily, the optimization process includes: setting multiple groups of candidate weight combinations, applying these weight combinations to the historical data set for feature value calculation and state classification respectively, and calculating the classification accuracy of each group of weights; finally, selecting a group of weights that can achieve the highest classification accuracy (e.g., engine speed weight 0.6 and throttle pedal position weight 0.4) as the final configuration.

[0077] It should be noted that the calculation process of the weighted feature value is: multiplying the normalized engine speed value by its corresponding speed weight to obtain a first product; multiplying the normalized throttle pedal position value by its corresponding pedal position weight to obtain a second product; and adding the first product and the second product to obtain the weighted feature value.

[0078] Exemplarily, assuming that the preset maximum value of the engine speed is 8000 RPM and the minimum value is 0 RPM, and the preset maximum value of the throttle pedal opening is 100% and the minimum value is 0%. If the current speed is 6000 RPM (normalized value 0.75) and the pedal opening is 30% (normalized value 0.3), the calculated dimensionless weighted feature value is 0.75x0.6+0.3x0.4=0.57. The calculated weighted feature value is taken as the real-time driving state recognition result.

[0079] In step S12, according to the real-time driving state recognition result, a preset threshold rule is used to determine the computing resource allocation requirement, including:

[0080] If the real-time driving state recognition result exceeds a preset dynamic scene threshold, an indication of needing to improve computing resource allocation is generated as the computing resource allocation requirement.

[0081] In an implementation manner, the real-time driving state recognition result is compared with a preset dynamic scene threshold. If the value exceeds the dynamic scene threshold, it indicates that the current driving scene is relatively strong in dynamics, and the system will generate an indication of needing to improve computing resource allocation, which is the computing resource allocation requirement.

[0082] It should be noted that the determination of the dynamic scene threshold is based on statistical analysis of the real-time driving state recognition result data under a large number of historical high-dynamic driving scenes (such as sudden acceleration and emergency avoidance). The analysis process includes calculating the probability density distribution of the recognition result under all historical high-dynamic scenes, and selecting a quantile that can cover 95% of the confidence interval as the preset dynamic scene threshold.

[0083] In step S13, according to the computing resource allocation requirement, the power of the edge processor is adjusted to obtain an adjusted power level, including:

[0084] The voltage and frequency parameters of the edge processor are adjusted by using a dynamic voltage and frequency scaling method to obtain adjusted voltage and frequency parameters;

[0085] According to the adjusted voltage and frequency parameters, the power output of the edge processor is reconfigured to obtain the adjusted power level.

[0086] In an implementation manner, when the system receives the computing resource allocation requirement generated in S12, a dynamic voltage and frequency scaling (DVFS) method is used to adjust the core voltage and working frequency parameters of the edge processor. It should be noted that dynamic voltage and frequency scaling is a control technology for dynamically balancing the power consumption and computing performance of a processor. The implementation of this method relies on a preset voltage and frequency mapping table, which defines multiple stable and reliable performance levels, each level corresponding to a specific voltage and frequency combination.

[0087] It should be noted that the voltage and frequency mapping table is usually provided by the manufacturer of the edge processor, or determined through experiments such as stress testing and power consumption analysis of the processor, to ensure the stability of system operation at each performance level.

[0088] Exemplarily, the voltage frequency mapping table can define a "standard performance level" (corresponding to a 1.2V voltage and a 1GHz frequency) and a "high performance level" (corresponding to a 1.4V voltage and a 1.5GHz frequency). When receiving the indication of the need to improve the allocation of computing resources, the system selects the "high performance level" from the mapping table, and controls the processor to switch to running at a 1.4V voltage and a 1.5GHz frequency, to obtain the adjusted voltage frequency parameters.

[0089] In an implementation manner, after obtaining the adjusted voltage frequency parameters, the system will reconfigure the power output of the edge processor through the power management module according to the parameters, to obtain the adjusted power level, which can support faster sensor data processing and decision response.

[0090] In step S14, the throttle opening degree driving signal is calculated according to the adjusted power level and the fusion of real-time monitoring data, and the throttle actuator instruction is generated, including:

[0091] The current energy consumption index is obtained from the adjusted power level;

[0092] If the current energy consumption index is lower than the preset energy consumption threshold, the engine speed data and the accelerator pedal position data are obtained and fused to obtain the driving intensity index;

[0093] The driving intensity index is input into a pre-trained logistic regression model to obtain a probability value of resource improvement demand;

[0094] According to the comparison between the probability value and the preset scheduling threshold, one of a plurality of preset scheduling schemes is selected as an optimized resource scheduling scheme;

[0095] According to the optimized resource scheduling scheme and using a multidimensional mapping table pre-established and defining the relationship between sensor data and throttle opening degree, the throttle opening degree driving signal is calculated;

[0096] The throttle opening degree driving signal is modulated to obtain the throttle actuator instruction.

[0097] In an implementation manner, the system uses an energy consumption sensor deployed on the power supply line of the edge processor to collect the actual power consumption data of the edge processor from the adjusted power level obtained in S13 through the real-time monitoring module, and the data is the current energy consumption index.

[0098] It should be noted that the system compares the current energy consumption indicator with a preset energy consumption threshold. If the current energy consumption indicator is lower than the threshold, it indicates that the processor is currently in an energy-saving state, and the system will trigger the subsequent resource scheduling optimization process. The preset energy consumption threshold is obtained based on clustering analysis of historical power consumption data of the processor. The determination process includes: collecting continuous power consumption data points of the processor under various historical load conditions; using the K-means clustering algorithm to divide these data points into at least two clusters, respectively corresponding to "low power consumption state cluster" and "high power consumption state cluster"; determining the boundary value between the two clusters as the preset energy consumption threshold. This threshold represents the critical point of the processor from the energy-saving state to the higher performance state.

[0099] It should be noted that after triggering the optimization process, the system will again obtain engine speed data and accelerator pedal position data, and use the same normalization method and weight value as in step S11 to perform weighted summation processing on the data to calculate a comprehensive driving intensity indicator. Subsequently, the system uses a pre-trained logistic regression model to analyze the driving intensity indicator to determine whether resources need to be improved.

[0100] It should be noted that the structure of the logistic regression model includes an input layer (receiving the driving intensity indicator), a linear layer with weight and bias parameters, and a Sigmoid activation function output layer. The training process includes: obtaining a training data set containing a large number of historical "driving intensity indicators" and corresponding "resource demand labels" (0 represents standard demand, 1 represents high performance demand); using supervised learning, taking binary cross-entropy as the loss function, and using gradient descent and other optimization algorithms to iteratively train the weight and bias parameters of the model; when the loss value of the model on the independent validation set decreases by less than a preset convergence threshold in consecutive multiple periods, the model is determined to be converged, and the training is completed. It should be noted that the determination of the convergence threshold is obtained through a grid search experiment. The experiment process includes: defining a set of candidate convergence thresholds in advance, for example {1e-3, 1e-4, 1e-5, 1e-6}; for each candidate threshold, use the threshold as the convergence standard to perform multiple cross-validation training on the training data set; record the average training time required for the model to converge under each candidate threshold, as well as the average classification accuracy of the model on the validation set after convergence; from all candidate thresholds, select the threshold (e.g. 1e-5) that can achieve the highest classification accuracy of the model within an acceptable training time as the final preset convergence threshold.

[0101] In one implementation, the system inputs the real-time calculated driving intensity indicator into the logistic regression model, and the model outputs a probability value representing the resource upgrade demand. The system compares the probability value with a preset dispatching threshold. The dispatching threshold is determined by ROC curve analysis on the model validation set, and the selection principle is to select the probability value that can make the false positive rate the lowest under the premise of not less than 98% true positive rate as the dispatching threshold.

[0102] It is worth noting that if the probability value exceeds the dispatching threshold, the system selects a preset "high-performance" resource scheduling scheme; otherwise, a "standard" resource scheduling scheme is selected. The parameters of these preset schemes are determined by a systematic performance calibration experiment.

[0103] It is worth noting that the calibration experiment process includes: defining key performance indicators (KPIs) and their target thresholds under different driving modes (such as standard and high performance), for example, the end-to-end delay of throttle control in high-performance mode must be less than 5 milliseconds; running the simulation load under different resource parameter combinations (such as different CPU priorities and sensor sampling frequencies) through the benchmark test program, and recording the actual performance indicators (such as average delay) and power consumption corresponding to each combination; select the resource parameter combination that can meet the performance indicator requirements of each mode and achieve the lowest power consumption, and solidify it as the corresponding preset scheme. These schemes are stored in the form of configuration files or parameter tables, which clearly define the allocation strategies of each resource under different schemes. For example, the standard scheme may configure the CPU priority to be 5 and the sensor sampling frequency to be 100 Hz; while the high-performance scheme configures the CPU priority to be 10 and the sensor sampling frequency to be 500 Hz. The selected scheme is the optimized resource scheduling scheme.

[0104] In an implementation, the system calculates a target throttle opening degree according to the optimized resource scheduling scheme. The calculation is based on a multi-dimensional mapping table calibrated in advance through engine bench test. The calibration process includes: mounting the engine on a dynamometer, traversing all stable operating points in a test matrix containing multiple engine speed and accelerator pedal position combinations; traversing a preset throttle opening degree search range at a preset small step (e.g. 0.1 degree) at each operating point, and recording the engine output torque at each opening degree in real time; determining the opening degree point at which the output torque reaches the maximum value as the optimal target opening degree value at the operating point; recording the correspondence between all operating points (input: engine speed, accelerator pedal position) and optimal target opening degrees (output) to form a mapping table. At the same time, independent mapping table areas are calibrated for different modes such as standard and high performance. It should be noted that the preset throttle opening degree search range is set according to the physical stroke limit of the throttle actuator and the effective response interval determined in combination with the engine bench test. The preset small step (e.g. 0.1 degree) is selected according to the minimum control resolution of the throttle actuator.

[0105] It should be noted that in order to adapt to the control requirements of the actuator, the system can use pulse width modulation (PWM) method to modulate the drive signal. The modulation controls the signal strength by adjusting the pulse width of the signal.

[0106] Exemplarily, if the finally calculated target opening degree is 40.2%, the signal generator can generate a corresponding voltage signal of 2.5V, and then the pulse width modulation module modulates the signal at a duty cycle of 50% to generate the final drive signal. The drive signal is then processed by the actuator control module, which includes encapsulating parameters such as the duty cycle (50%) of the drive signal into a digital message conforming to the Controller Area Network (CAN) bus protocol, and the digital message is the throttle actuator instruction.

[0107] In step S15, the execution response time of the throttle actuator instruction is monitored, and the edge processor power is adjusted again according to the execution response time to obtain a final balanced throttle control signal, including:

[0108] The execution response time of the throttle actuator instruction is continuously monitored to obtain a response time data set;

[0109] If the response time in the response time data set exceeds the preset response time threshold, the power output is recalculated to obtain a new power target value as the first power output data set;

[0110] According to the first power output data set, the performance level of the edge processor is adjusted;

[0111] At the adjusted performance level, the initial throttle control signal is recalculated based on the latest sensor real-time data;

[0112] The initial throttle control signal is fused with the actual throttle opening degree feedback value at the previous time to generate the final balanced throttle control signal.

[0113] In an implementation, the system continuously monitors the throttle actuator command issued by S14 through a high-precision timer. This monitoring process records the time difference from the command issuance time to the completion time of the actuator physical action, and stores this time difference as the response time in a response time dataset.

[0114] It should be noted that the system compares the response time recorded in the response time dataset with a preset response time threshold. If the response time exceeds the threshold, it indicates that the current power distribution may not meet the real-time load demand (for example, the engine load suddenly increases when the vehicle is climbing uphill), and the system determines a power output adjustment requirement.

[0115] It is worth noting that the determination of the preset response time threshold is based on statistical analysis of historical response time data of the system under different loads. The process includes: collecting the response time distribution of the system under multiple known load levels; analyzing the average response time and standard deviation under each load level; selecting a critical value (for example, the average value plus three times the standard deviation) that can distinguish between normal response and significant lag response as the preset response time threshold.

[0116] In an implementation, after determining the power output adjustment requirement, the system uses a power distribution module to recalculate the power output through an incremental PI controller algorithm. It should be noted that this algorithm is a well-known technology in the field of industrial control for dynamic correction. The calculation process is as follows: the over-limit part of the response time is input as an error; the power adjustment increment is calculated according to the following exemplary discrete PI control formula:

[0117] ;

[0118] Where, represents the power adjustment increment calculated at the current time; represents the preset proportional coefficient; represents the preset integral coefficient; represents the error at the current time, which is the value of "actual response time" minus "preset response time threshold"; represents the error at the previous time error of the instant.

[0119] adding the current power output and the power adjustment increment to obtain a new power target value, which is the first power output data set. It should be noted that the proportional coefficient Kp and the integral coefficient Ki are determined in advance according to the dynamic characteristics of the controlled system by the experimental method known in the art such as Ziegler-Nichols tuning method.

[0120] For example, if the current power level of the edge processor is 5W, but the response time is out of limit, and the power adjustment increment calculated by the incremental PI controller according to the error is +0.5W, then the system determines the new power target value as 5.5W, which is the first power output data set.

[0121] It should be noted that the system first applies the first power output data set (i.e. the new power target value) to the edge processor, and adjusts the computing performance of the processor to the level matched with the new target value through power management technologies such as dynamic voltage and frequency scaling (DVFS).

[0122] Under this enhanced performance level, the system recalculates a new and ideal throttle opening command according to the real-time data fed back (such as the current engine speed and the driver's throttle pedal request) and using the multi-dimensional mapping table described in S14.

[0123] To ensure smooth transition of the control signal and avoid vehicle shaking caused by sudden change of the command, the system further integrates the newly calculated ideal throttle opening command and the actual throttle opening value fed back in the previous control cycle through a data fusion algorithm to generate the final balanced throttle control signal. The fusion process can use the weighted average method, and the calculation process is as follows: the newly calculated ideal command and the actual opening value fed back are multiplied by their corresponding weights and then summed. The determination of the weight value is based on the off-line sensitivity analysis experiment, which measures the influence of the weight ratio on the stability and response speed of the final control signal in the simulation environment; the features with higher influence degree can be given higher weight. For example, the weight of the newly calculated command can be set to 0.7, and the weight of the historical actual opening feedback can be set to 0.3, so as to ensure the smoothness of the control while ensuring the fast response.

[0124] In step S16, the latest sensor data is preprocessed and classified according to the final balanced throttle control signal to determine the updated driving state recognition result, including:

[0125] According to the final balanced throttle control signal, the sensor data is obtained;

[0126] The sensor data is subjected to signal noise verification to obtain a verification result;

[0127] According to the verification result, the sensor data is subjected to conditional filtering processing to obtain purified sensor data;

[0128] The purified sensor data is subjected to normalization processing to obtain data to be classified;

[0129] The data to be classified is input into a pre-established driving state model for classification processing to obtain the updated driving state recognition result.

[0130] In an implementation manner, after the final balanced throttle control signal is generated and issued at S15, the system again acquires the latest real-time sensor data through the sensor network, which reflects the system state after the control action is executed.

[0131] It should be noted that the system performs signal noise verification on the newly acquired sensor data. The verification process includes calculating the standard deviation of the data sequence within a period of time, and taking the standard deviation as a quantitative indicator of signal noise to obtain the verification result.

[0132] In an implementation manner, the system compares the verification result with a preset signal noise threshold. If the verification result exceeds the threshold, the sensor data is subjected to conditional filtering processing; if it does not exceed, the original sensor data is directly taken as data to be classified.

[0133] It should be noted that the determination of the signal noise threshold is based on statistical analysis of a large amount of historical sensor data under normal working conditions. The analysis process includes calculating the probability density distribution of the signal noise indicator under all historical normal working conditions, and selecting the upper limit value that can cover the 95% confidence interval as the preset signal noise threshold.

[0134] It should be noted that the conditional filtering processing can be realized by using a Kalman filter. Through the prediction and update steps, the Kalman filter combines historical data and current measurement values to smooth the data containing out-of-limit noise, and generates a more stable and lower-noise data set, which is the data to be classified.

[0135] It should be noted that the system performs normalization processing on the data to be classified before inputting it into the driving state model; then, the normalized data is input into a pre-established driving state model for classification processing by the model to obtain the updated driving state recognition result.

[0136] It should be noted that the normalization processing is realized by using the same minimum-maximum normalization method as in step S11, that is, the data after the conditional filtering processing is linearly mapped to the interval [0, 1] according to the preset maximum value and minimum value, so as to ensure that the data scale input to the model is completely consistent with the data scale used during the model training.

[0137] It should be noted that the driving state model is a classification model constructed based on a support vector machine (SVM) algorithm. The training process is as follows: a training data set containing a large amount of normalized historical sensor data (as input features) and corresponding real driving state labels (such as "idling", "cruising", etc., as output) labeled by experts is obtained; the SVM model is trained by finding a hyperplane that can maximize the interval between different categories of samples in a supervised learning manner; when the classification accuracy of the model on the independent verification set is improved by less than a preset convergence threshold (for example, 0.01%) in continuous multiple training periods, the training is terminated. It should be noted that the convergence threshold is obtained through a grid search experiment. The experiment process includes: a group of candidate convergence thresholds are defined in advance, for example, {0.1%, 0.01%, 0.001%}; for each candidate threshold, use the threshold as the stopping condition for training, and perform multiple cross-validation training on the training data set; record the average training time required for the model to reach convergence under each candidate threshold, and the average classification accuracy of the model on the verification set after convergence; from all candidate thresholds, select the threshold (for example, 0.01%) that can achieve the highest classification accuracy of the model within an acceptable training time as the final preset convergence threshold.

[0138] It should be noted that the key hyperparameters of the model, such as the selection of the kernel function (such as the radial basis function (RBF) kernel) and its corresponding parameters (such as C and gamma), are determined through a systematic grid search and cross-validation experiment. The experiment process includes: defining a grid of candidate values for each hyperparameter to be optimized (such as C and gamma); for each parameter combination in the grid, perform model training and evaluation on the training data set using k-fold cross-validation; record the average classification accuracy of each parameter combination on all validation folds; select the parameter combination that can achieve the highest average classification accuracy to train the final SVM model.

[0139] In actual use, the system inputs the latest to-be-classified data (after normalization processing) as input, and outputs a final driving state recognition result corrected in a closed loop by the trained SVM model.

[0140] In order to facilitate the understanding of the present application, some preferred embodiments of the present application will be further described below.

[0141] In an implementation, the method of the present application can be applied to a smart connected commercial vehicle equipped with an edge computing unit to achieve precise and energy-saving control of the engine throttle. A complete workflow example is as follows:

[0142] Step one, the vehicle is driving smoothly on urban roads, the system continuously obtains the original data stream of engine speed (for example, 2000 RPM) and throttle pedal position (for example, 20%). After time stamp alignment, Kalman filtering and minimum-maximum normalization processing, the normalized speed value is 0.25 and the pedal position value is 0.2. The system calculates the weighted feature value as 0.23 according to the preset weight, and takes this value as the real-time driving state recognition result.

[0143] Step two, the system compares the recognition result 0.23 with the preset state type threshold (for example, idle speed threshold 0.2, cruise threshold 0.5), and classifies it as "cruise state". At the same time, since the value does not exceed the preset dynamic scene threshold (for example, 0.6), the system determines that there is no need to upgrade the computing resources, and generates an indication to maintain the current resource allocation.

[0144] Step three, since there is no need to upgrade resources, the edge processor is maintained at a standard energy-saving power level (for example, 1.2V, 1GHz). The energy consumption index obtained from this power level is lower than the preset energy consumption threshold, indicating that there is surplus computing power, which triggers the optimization process immediately, selects the "standard" resource scheduling scheme, and calculates the corresponding throttle opening command based on the mapping table calibrated by the engine bench.

[0145] Step four, an obstacle appears in front of the vehicle, the driver steps on the accelerator pedal urgently to overtake, and the pedal position quickly changes to 90%. The system calculates a new feature value of 0.75 in step one. Step two immediately determines that it is a high dynamic scene and generates an indication to upgrade resources.

[0146] Step five, the power adjustment module receives the indication and immediately adjusts the processor to "high performance level" (for example, 1.4V, 1.5GHz) through DVFS. However, since the engine load increases sharply, the throttle command issued by the instruction generation module is monitored by the feedback control module, and the actual execution response time is 0.12 seconds, which exceeds the preset response threshold of 0.1 seconds. The feedback control module is triggered immediately, calculates the power adjustment increment through the incremental PI controller algorithm, adjusts the power again, and generates the final balanced throttle control signal by fusing real-time feedback data, ensuring that the throttle can respond at the fastest speed.

[0147] Step six, after the overtaking action is completed, the state updating module updates the driving state of the vehicle to "smooth driving" according to the latest sensor data after noise verification and normalization processing, and inputs the sensor data into the pre-trained SVM model.

[0148] In another implementation, the system deployment architecture of the application can also be configured differently. For example, in a large fleet management scenario, the sensor data can be uploaded to the cloud server in real time through the vehicle-mounted 5G module. All complex model training, state recognition and power regulation decisions are completed in the cloud, and the final throttle control signal calculated is issued to the local controller of the vehicle through the network for execution.

[0149] In summary, the application discloses a kind of based on Internet of Things's intelligent control method and system of throttle valve. The method comprises: obtaining sensor data and processing, to obtain real-time driving state recognition result;According to the real-time driving state recognition result, the threshold rule of pre-set is used to determine the computing resource allocation demand;According to the computing resource allocation demand, the power of edge processor is adjusted, to obtain the power level after adjustment;According to the power level after adjustment, and fusion real-time monitoring data, calculate the throttle opening degree drive signal, and generate throttle actuator instruction;Monitoring the execution response time of the instruction and adjusting the power again, to obtain the final balanced throttle control signal;Finally, according to the final balanced throttle control signal, determine the updated driving state recognition result.The application realizes the intelligent balance of computing resource and system response in throttle control by constructing the complete technical process from multi-dimensional data perception, feedforward and feedback combined double-layer dynamic power regulation to the final closed-loop state update, solves the problem that control precision and energy efficiency are difficult to consider in complex dynamic working condition in prior art.

[0150] Reference Figure 2 The second embodiment of the application provides a kind of based on Internet of Things's intelligent control system of throttle valve, comprising:

[0151] State recognition module, for obtaining sensor data and processing, to obtain real-time driving state recognition result;

[0152] Demand determination module, for according to the real-time driving state recognition result, the threshold rule of pre-set is used to determine the computing resource allocation demand;

[0153] Power regulation module, for according to the computing resource allocation demand, the power of edge processor is adjusted, to obtain the power level after adjustment;

[0154] Instruction generation module, for according to the power level after adjustment, and fusion real-time monitoring data, calculate the throttle opening degree drive signal, and generate throttle actuator instruction;

[0155] A feedback control module is configured to monitor an execution response time of the throttle actuator instruction, and adjust the edge processor power again according to the execution response time to obtain a final balanced throttle control signal.

[0156] A state updating module is configured to pre-process and classify the latest sensor data according to the final balanced throttle control signal, and determine an updated driving state recognition result.

[0157] It should be noted that the above-mentioned embodiment of the application provides a kind of based on Internet of Things's throttle intelligent control system for executing the all process steps of a kind of based on Internet of Things's throttle intelligent control method of above-mentioned embodiment, the working principle and beneficial effects of the two are one-to-one correspondence, thus no longer repeat.

[0158] The embodiment of the application further provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a throttle intelligent control program based on Internet of Things. The processor executes the computer program to implement the steps in each of the above-mentioned throttle intelligent control method embodiments based on Internet of Things, such as Figure 1 The step S11 shown. Alternatively, the processor executes the computer program to implement the functions of each module / unit in each of the above-mentioned device embodiments, such as a state recognition module.

[0159] For example, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the application. The one or more modules / units can be a series of computer program instruction segments that can complete a specific function, which are used to describe the execution process of the computer program in the electronic device.

[0160] The electronic device can be a desktop computer, a notebook computer, a palm computer, and a smart tablet computer, etc. The electronic device can include, but is not limited to, a processor, a memory. Those skilled in the art can understand that the above-mentioned components are only examples of the electronic device, and do not constitute a limitation on the electronic device, and can include more or fewer components than the above-mentioned, or combine certain components, or different components, for example, the electronic device can also include an input / output device, a network access device, a bus, etc.

[0161] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The processor is a control center of the electronic device, and connects various parts of the electronic device through various interfaces and lines.

[0162] The memory can be used to store the computer program and / or modules, and the processor realizes various functions of the electronic device by running or executing the computer program and / or modules stored in the memory, and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), etc.; and the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a nonvolatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory device.

[0163] The modules / units integrated in the electronic device, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. that can carry the computer program code. It should be noted that the contents included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0164] It should be noted that the above-described device embodiments are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. In addition, the connection relationship between the modules in the device embodiment provided by the present application indicates that there is a communication connection between them, which can be realized as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0165] The above-described specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above-described specific embodiments are only for the specific embodiments of the present application and do not limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An Internet of Things-based intelligent control method for a throttle valve, characterized by, The method comprises the following steps: acquiring sensor data and processing the same to obtain a real-time driving state recognition result; determining a computing resource allocation requirement according to the real-time driving state recognition result by using a preset threshold rule; adjusting the power of an edge processor according to the computing resource allocation requirement to obtain an adjusted power level; calculating a throttle opening degree driving signal and generating a throttle actuator instruction according to the adjusted power level and by fusing real-time monitoring data; monitoring the execution response time of the throttle actuator instruction and adjusting the power of the edge processor again according to the execution response time to obtain a final balanced throttle control signal; preprocessing and classifying the latest sensor data according to the final balanced throttle control signal to determine an updated driving state recognition result. 2.The IoT-based intelligent control method of a throttle valve according to claim 1, wherein, The method of acquiring sensor data and processing the same to obtain a real-time driving state recognition result comprises the following steps: acquiring an original data stream of engine speed and accelerator pedal position; performing timestamp alignment processing on the original data stream to obtain a synchronous data stream; performing filtering and denoising processing on the synchronous data stream to obtain a smooth data sequence; performing normalization processing on the data in the smooth data sequence and performing weighted calculation to obtain weighted feature values, and taking the weighted feature values as the real-time driving state recognition result. 3.The IoT-based intelligent control method of a throttle according to claim 1, wherein, The method of determining a computing resource allocation requirement according to the real-time driving state recognition result by using a preset threshold rule comprises the following steps: if the real-time driving state recognition result exceeds a preset dynamic scene threshold, generating an indication of needing to improve computing resource allocation as the computing resource allocation requirement. 4.The IoT-based intelligent control method of a throttle valve according to claim 1, wherein, The method of adjusting the power of an edge processor according to the computing resource allocation requirement to obtain an adjusted power level comprises the following steps: adjusting the voltage and frequency parameters of the edge processor by using a dynamic voltage and frequency scaling method to obtain adjusted voltage and frequency parameters; reconfiguring the power output of the edge processor according to the adjusted voltage and frequency parameters to obtain the adjusted power level. 5.The IoT-based intelligent control method of a throttle valve according to claim 1, wherein, The method of calculating a throttle opening degree driving signal and generating a throttle actuator instruction according to the adjusted power level and by fusing real-time monitoring data comprises the following steps: obtaining a current energy consumption indicator from the adjusted power level; if the current energy consumption indicator is lower than a preset energy consumption threshold, obtaining and fusing engine speed data and accelerator pedal position data to obtain a driving intensity indicator; inputting the driving intensity indicator into a pre-trained logistic regression model to obtain a probability value of resource improvement requirement; comparing the probability value with a preset scheduling threshold to select one from a plurality of preset scheduling schemes as an optimized resource scheduling scheme; calculating a throttle opening degree driving signal according to the optimized resource scheduling scheme and by using a pre-established multidimensional mapping table defining the relationship between sensor data and throttle opening degree; performing modulation processing on the throttle opening degree driving signal to obtain the throttle actuator instruction. 6.The IoT-based intelligent control method of a throttle valve according to claim 1, wherein, The monitoring the execution response time of the throttle actuator instruction, and adjusting the edge processor power according to the execution response time to obtain a final balanced throttle control signal, comprises: The execution response time of the throttle actuator instruction is continuously monitored to obtain a response time data set; If there is a response time in the response time data set that exceeds a preset response time threshold, the power output is recalculated to obtain a new power target value as a first power output data set; According to the first power output data set, the performance level of the edge processor is adjusted; At the adjusted performance level, the initial throttle control signal is recalculated based on the latest sensor real-time data; The initial throttle control signal is fused with the actual throttle opening feedback value at the previous time to generate the final balanced throttle control signal. 7.The IoT-based intelligent control method of a throttle valve according to claim 1, wherein, According to the final balanced throttle control signal, the latest sensor data is preprocessed and classified to determine an updated driving state recognition result, comprising: According to the final balanced throttle control signal, the sensor data is obtained; The sensor data is subjected to signal noise verification to obtain a verification result; According to the verification result, the sensor data is subjected to conditional filtering processing to obtain purified sensor data; The purified sensor data is subjected to normalization processing to obtain classification data; The classification data is input into a pre-established driving state model for classification processing to obtain the updated driving state recognition result.

8. An intelligent control system for a throttle based on an Internet of Things, characterized in that, Comprise: The state recognition module is used for obtaining sensor data and processing to obtain real-time driving state recognition results; The demand determination module is used for determining the calculation resource allocation demand according to the real-time driving state recognition results by using a preset threshold rule; The power adjustment module is used for adjusting the edge processor power according to the calculation resource allocation demand to obtain an adjusted power level; The instruction generation module is used for calculating the throttle opening driving signal according to the adjusted power level and fusing real-time monitoring data, and generating the throttle actuator instruction; The feedback control module is used for monitoring the execution response time of the throttle actuator instruction, and adjusting the edge processor power according to the execution response time to obtain a final balanced throttle control signal; The state update module is used for preprocessing and classifying the latest sensor data according to the final balanced throttle control signal to determine an updated driving state recognition result.

Images