Commercial vehicle abnormal energy consumption identification system and method based on Internet of Vehicles multidimensional data

Abstract

The invention provides a commercial vehicle abnormal energy consumption identification system and method based on Internet of Vehicles multi-dimensional data. The commercial vehicle abnormal energy consumption identification system based on Internet of Vehicles multi-dimensional data comprises a data acquisition module, a data processing module, an abnormal energy consumption identification module, an abnormal energy consumption reason diagnosis module and a pushing module. The commercial vehicle abnormal energy consumption identification method based on the Internet of Vehicles multi-dimensional data comprises the following steps: S1, collecting vehicle data and carrying out preprocessing; s2, abnormal energy consumption identification is carried out based on the vehicle data, and high-energy-consumption vehicles are screened out; s3, diagnosing the cause of high energy consumption; s4, pushing a diagnosis result; according to the method, comprehensive energy consumption analysis is carried out on multi-dimensional data such as vehicle states, driving tracks and driving behaviors, multi-level similarity analysis is carried out on similar vehicles, routes and road characteristics, high-energy-consumption vehicles are recognized more accurately, and errors of a single standard are avoided; abnormal energy consumption reasons can be accurately and timely diagnosed, motorcade management is optimized, and operation cost is reduced.

Description

Technical Field

[0001] The present invention relates to the field of vehicle networking technology, and in particular to a system and method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the vehicle networking. Background Art

[0002] Commercial vehicles play a vital role in the transportation industry, particularly in logistics and long-distance transport. Fuel consumption is a core indicator of fleet management and operating costs. Fleet managers and drivers often notice abnormally higher-than-expected fuel consumption, especially when compared with similar vehicles or manufacturer-provided reference fuel consumption. This discrepancy can lead to complaints. Fuel consumption directly impacts operating costs and, given the fluctuating fuel prices, is a particular focus in the commercial vehicle sector.

[0003] Commercial vehicle energy consumption is affected by many factors, including but not limited to vehicle configuration, route, road conditions, gross vehicle weight, vehicle status, driving habits, vehicle operating conditions, and environmental factors. While the application of connected vehicle technology and big data analytics has simplified the process of fuel consumption monitoring and analysis to a certain extent, allowing fleet managers to monitor vehicle fuel consumption in real time and identify potential energy consumption anomalies through connected vehicle data, existing methods for commercial vehicle energy consumption analysis still have the following drawbacks: Low accuracy: The current mainstream method for analyzing commercial vehicle energy consumption is the baseline method. This method uses big data analysis to confirm the vehicle's normal energy consumption range. Any energy consumption outside this range is considered an anomaly. This method relies on a single energy consumption baseline to determine anomalies, ignoring multi-dimensional data analysis. Consequently, energy consumption assessment and anomaly diagnosis lack persuasiveness and accuracy. Lack of a unified evaluation benchmark: Existing methods fail to unify and align the evaluation benchmarks when conducting energy consumption diagnosis, resulting in inconsistent energy consumption assessment standards for different vehicles and different operating environments, affecting the energy consumption analysis results. Summary of the Invention

[0004] The purpose of the present invention is to address the shortcomings of the prior art and to propose a system and method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the Internet of Vehicles.

[0005] To achieve the above object, the present invention adopts the following technical solutions: A system and method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the Internet of Vehicles. The system includes a data acquisition module, a data processing module, an abnormal energy consumption identification module, an abnormal energy consumption cause diagnosis module, and a push module. The data acquisition module is used to collect vehicle data in real time; The data processing module is used to pre-process the collected vehicle data; The abnormal energy consumption identification module is used to identify abnormal energy consumption of vehicles and screen out high energy consumption vehicles; The abnormal energy consumption cause diagnosis module is used to diagnose the cause of high energy consumption and obtain a diagnosis result; The push module is used to push the diagnosis results to relevant personnel.

[0006] The method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the Internet of Vehicles includes the following steps: S1: Collect vehicle data and preprocess it; The following sub-steps are included: S11: Real-time collection of vehicle data; The data acquisition module collects vehicle data in real time through the vehicle terminal Tbox. The vehicle data includes vehicle identification code, vehicle status information, trajectory data, fuel consumption rate, etc. The vehicle identification code is unique; the vehicle status information includes engine speed, accelerator pedal position, etc.; the trajectory data is composed of trajectory points, each of which includes real-time latitude and longitude, vehicle speed, timestamp, road section elevation, horizontal distance, etc.; The data acquisition module collects driving behavior information in real time through sensors, including gear position, throttle, brake, etc. The data acquisition module transmits the collected vehicle data to the data processing layer; S12: pre-processing vehicle data; The data processing layer receives vehicle data and pre-processes the vehicle data; The preprocessing includes abnormal data removal, spatiotemporal alignment, feature fusion, etc. S2: Identify abnormal energy consumption based on vehicle data and screen out high-energy-consuming vehicles; The abnormal energy consumption identification module is pre-set with the cosine similarity threshold of similar vehicles, similarity coefficient threshold, cosine similarity threshold of the characteristic vector of the driving trajectory, average speed difference threshold, and specific percentage of fuel consumption rate exceeding the average value; The following sub-steps are included: S21: performing vehicle similarity recognition; The abnormal energy consumption identification module divides the key parameters affecting commercial vehicle fuel consumption into categorical parameters and numerical parameters. The categorical parameters include vehicle model, tire specifications, engine model, and whether there is a diversion device; the numerical parameters include the vehicle's total speed ratio. By directly comparing the category parameters corresponding to each vehicle, vehicles with consistent category parameters are selected; Convert all numerical parameters corresponding to each vehicle with the same category parameters into vectors, one vehicle corresponding to one vector, and each dimension in the vector corresponding to a numerical parameter; calculate the cosine similarity of the vectors of each two vehicles in turn. The cosine similarity formula is as follows: ; Among them, C is the cosine similarity, A and B are the vectors of the two cars, is the vector dot product, and is the magnitude of each vector; Filter out vehicles whose cosine similarity is within the cosine similarity threshold as similar vehicle group A; S22: performing similar route identification; Use Kalman filter algorithm to denoise the trajectory data; The abnormal energy consumption identification module uses a grid method to divide the driving area involved in the vehicles in the similar vehicle group A into multiple small grids to obtain the total number of grids; all the trajectory points of each vehicle are mapped to the grids, and the grids that each vehicle has traveled through are marked; The similarity coefficient of each two vehicle's driving trajectories is calculated in sequence using the spatial overlap method. The specific calculation method is as follows: The number of times the driving trajectories of each two vehicles pass through the same grid is obtained in sequence to obtain the number of identical grids; the similarity coefficient is calculated using a similarity coefficient formula, wherein the similarity coefficient formula is: similarity coefficient = number of identical grids / total number of grids; Filter out vehicles whose similarity coefficients are within the similarity coefficient threshold from the similar vehicle group A as the similar vehicle group B; S23: performing similar road feature recognition; Route similarity takes precedence over road feature similarity. For vehicles that are in similar vehicle group A but not in similar vehicle group B, the abnormal energy consumption identification module calculates the slope of each vehicle's trajectory based on the road segment elevation and horizontal distance in the trajectory data, as follows: Based on the elevation of the road section, the height difference between the starting point and the end point of the trajectory is obtained, which is recorded as Δelevation; the horizontal distance corresponding to the trajectory is recorded as Δhorizontal distance; ; The abnormal energy consumption identification module classifies slopes into different categories based on their numerical values. It then calculates the proportion of each slope category in each vehicle's driving trajectory and constructs a feature vector. Each dimension in the feature vector corresponds to the proportion of a slope category. Calculate the cosine similarity of the feature vectors of the corresponding driving trajectories of every two vehicles in turn, and select the vehicles whose feature vectors of the driving trajectories are within the set cosine similarity threshold as the similar vehicle group C; S24: perform similar transport feature identification; For vehicles in similar vehicle group B and similar vehicle group C, the abnormal energy consumption identification module calculates the average speed of each vehicle in a specific time period based on the timestamp and speed obtained from the trajectory data, and compares the average speed difference of the vehicles. If the average speed difference of the vehicles is less than the set average speed difference threshold, the vehicles have basically the same transportation time requirements, and vehicles with basically the same transportation time requirements are screened out; S25: perform transport weight identification; The abnormal energy consumption identification module identifies and corrects the total transport weight of the vehicle obtained in step S24 to obtain the transport weight of the vehicle in the corresponding time period; The method for predicting and correcting the gross transport weight is from the patent "Method for Gross Weight Identification and Confidence Assessment of Heavy Freight Vehicles Based on Cloud Computing" (Publication No. CN117951487A). The present invention only applies the gross weight identification method and does not innovate the gross weight identification method. The principle of the gross weight identification method will not be elaborated on in this invention. S26: Eliminate abnormal data and obtain vehicles with similar transportation characteristics; The abnormal energy consumption identification module extracts the actual energy consumption of each vehicle from the energy consumption data uploaded by the vehicle bus based on the vehicle identification code corresponding to the vehicle obtained in step S25; Based on historical data and statistical methods, the actual energy consumption that deviates from the average is detected and eliminated. The remaining actual energy consumption corresponds to vehicles with similar transportation characteristics. S27: Screen out high energy consumption vehicles; The abnormal energy consumption identification module sets the ratio A; Based on the vehicle identification codes corresponding to the vehicles with similar transportation characteristics obtained in step S26, the fuel consumption rate of each vehicle is obtained from the vehicle data collected in step S1; The average fuel consumption rate of the vehicles is calculated using the average value calculation formula, and the vehicles are ranked by fuel consumption rate; the abnormal energy consumption identification module selects the vehicles ranked in the top A and whose fuel consumption rate exceeds the average value by a specific percentage as high-energy-consuming vehicles; S3: Diagnose the cause of high energy consumption; The abnormal energy consumption cause diagnosis module pre-sets the fault energy consumption time / mileage ratio, high speed range / high rotation speed range, and difference threshold; The following sub-steps are included: S31: Perform energy consumption related fault diagnosis; The abnormal energy consumption cause diagnosis module uses intelligent diagnostic technology to identify whether a high-energy-consuming vehicle has a fault that significantly affects energy consumption during the analyzed time period. If so, the duration or mileage of the fault is compared with a pre-set fault energy consumption time / mileage ratio. If the duration or mileage ratio of the vehicle's fault is greater than the pre-set fault energy consumption time / mileage ratio, it is determined that one of the causes of the high energy consumption of the high-energy-consuming vehicle is an energy-related fault. For example, if the set fault energy consumption time / mileage ratio is 20%, if the duration or mileage of the vehicle fault ratio is greater than 20%, it is determined that one of the causes of high energy consumption of this high energy consumption vehicle is an energy consumption-related fault; S32: Diagnose bad driving behavior; The abnormal energy consumption diagnosis module evaluates driving behavior information. If there are frequent sudden acceleration or braking, long periods of idling, etc., it will determine that one of the causes of high energy consumption of this high-energy-consuming vehicle is bad driving behavior. S33: Perform operating condition difference diagnosis; The abnormal energy consumption cause diagnosis module obtains the proportion of high-energy-consuming vehicles in a set high-speed range and high-speed range based on the vehicle's speed and rotational speed, compares the proportion of high-energy-consuming vehicles in the set high-speed range / high-speed range with the proportion of non-high-energy-consuming vehicles in the set high-speed range / high-speed range, and if the difference exceeds a set difference threshold, determines that one of the causes of high energy consumption of the high-energy-consuming vehicle is a difference in operating conditions; Get diagnostic results for high energy consumption vehicles.

[0007] S4: Push the diagnosis results; The push module regularly pushes the diagnostic results to the business front desk through the diagnostic middle desk, and the business front desk notifies customers or relevant operation service personnel; after receiving the diagnostic results, corresponding measures are taken in real time.

[0008] Compared with the prior art, the present invention has the following beneficial effects: The present invention conducts a comprehensive energy consumption analysis by integrating multi-dimensional data such as vehicle status, driving trajectory, and driving behavior, and introduces multi-level similarity analysis such as similar vehicles, similar routes, and similar road characteristics. This can more accurately identify high-energy-consuming vehicles and avoid the errors of a single standard. For high-energy-consuming vehicles, the present invention conducts cause diagnosis from aspects such as fault diagnosis, poor driving behavior, and differences in operating conditions, so as to more accurately and promptly diagnose the causes of abnormal energy consumption, optimize fleet management, and reduce operating costs.

[0009] By classifying and standardizing vehicles, we ensure the uniformity of energy consumption assessment and avoid inconsistent standards in different vehicles or environments.

[0010] The system can push diagnostic results in real time, so fleet managers or drivers can take countermeasures immediately to avoid further waste of energy and improve the fleet's operational efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 This is a flowchart of the steps of the method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the Internet of Vehicles of the present invention. DETAILED DESCRIPTION

[0012] In order to provide a further understanding of the purpose, structure, features, and functions of the present invention, the present invention is described in detail below with reference to the embodiments.

[0013] like Figure 1 As shown, a commercial vehicle abnormal energy consumption identification system and method based on multi-dimensional data of the Internet of Vehicles includes a data acquisition module, a data processing module, an abnormal energy consumption identification module, an abnormal energy consumption cause diagnosis module, and a push module; The data acquisition module is used to collect vehicle data in real time; The data processing module is used to pre-process the collected vehicle data; The abnormal energy consumption identification module is used to identify abnormal energy consumption of vehicles and screen out high energy consumption vehicles; The abnormal energy consumption cause diagnosis module is used to diagnose the cause of high energy consumption and obtain a diagnosis result; The push module is used to push the diagnosis results to relevant personnel.

[0014] The method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the Internet of Vehicles includes the following steps: S1: Collect vehicle data and preprocess it; The following sub-steps are included: S11: Real-time collection of vehicle data; The data acquisition module collects vehicle data in real time through the vehicle terminal Tbox. The vehicle data includes vehicle identification code, vehicle status information, trajectory data, fuel consumption rate, etc. The vehicle identification code is unique; the vehicle status information includes engine speed, accelerator pedal position, etc.; the trajectory data is composed of trajectory points, each of which includes real-time latitude and longitude, vehicle speed, timestamp, road section elevation, horizontal distance, etc.; The data acquisition module collects driving behavior information in real time through sensors, including gear position, throttle, brake, etc. The data acquisition module transmits the collected vehicle data to the data processing layer; S12: pre-processing vehicle data; The data processing layer receives vehicle data and pre-processes the vehicle data; The preprocessing includes abnormal data removal, spatiotemporal alignment, feature fusion, etc. Specifically, machine learning algorithms such as isolation forests are used to detect and remove abnormal data in vehicle data. Interpolation is used to smoothly transition vehicle data with different time stamps to achieve time alignment. Geographic Information System (GIS) tools are used to adjust vehicle data in different spatial coordinate systems so that the vehicle data are aligned in a unified coordinate system to achieve spatial alignment. Vehicle data is standardized using methods such as Z-score standardization, and relevant features are selected from the vehicle data and feature fusion is performed using deep learning methods. S2: Identify abnormal energy consumption based on vehicle data and screen out high-energy-consuming vehicles; The abnormal energy consumption identification module is pre-set with the cosine similarity threshold of similar vehicles, similarity coefficient threshold, cosine similarity threshold of the characteristic vector of the driving trajectory, average speed difference threshold, and specific percentage of fuel consumption rate exceeding the average value; The following sub-steps are included: S21: performing vehicle similarity recognition; The abnormal energy consumption identification module divides the key parameters affecting commercial vehicle fuel consumption into categorical parameters and numerical parameters. The categorical parameters include vehicle model, tire specifications, engine model, and whether there is a diversion device; the numerical parameters include the vehicle's total speed ratio. By directly comparing the category parameters corresponding to each vehicle, vehicles with consistent category parameters are selected; Convert all numerical parameters corresponding to each vehicle with the same category parameters into vectors, one vehicle corresponding to one vector, and each dimension in the vector corresponding to a numerical parameter; calculate the cosine similarity of the vectors of each two vehicles in turn. The cosine similarity formula is as follows: ; Among them, C is the cosine similarity, A and B are the vectors of the two cars, is the vector dot product, and is the magnitude of each vector; Filter out vehicles whose cosine similarity is within the cosine similarity threshold as similar vehicle group A; By combining categorical and numerical parameters to calculate vehicle similarity, and using cosine similarity for vectorized comparison, the system not only considers basic vehicle attributes (such as vehicle model and engine model), but also uses trajectory data, road slope and other data to further accurately distinguish similar vehicles. This allows for more comprehensive identification of similar vehicles, avoiding errors caused by a single indicator or simple comparison, thereby improving recognition accuracy.

[0015] S22: performing similar route identification; Use Kalman filter algorithm to denoise the trajectory data; The abnormal energy consumption identification module uses a grid method to divide the driving area involved in the vehicles in the similar vehicle group A into multiple small grids to obtain the total number of grids; all the trajectory points of each vehicle are mapped to the grids, and the grids that each vehicle has traveled through are marked; The similarity coefficient of each two vehicle's driving trajectories is calculated in sequence using the spatial overlap method. The specific calculation method is as follows: The number of times the driving trajectories of each two vehicles pass through the same grid is obtained in sequence to obtain the number of identical grids; the similarity coefficient is calculated using a similarity coefficient formula, wherein the similarity coefficient formula is: similarity coefficient = number of identical grids / total number of grids; Filter out vehicles whose similarity coefficients are within the similarity coefficient threshold from the similar vehicle group A as the similar vehicle group B; S23: performing similar road feature recognition; Route similarity takes precedence over road feature similarity. For vehicles that are in similar vehicle group A but not in similar vehicle group B, the abnormal energy consumption identification module calculates the slope of each vehicle's trajectory based on the road segment elevation and horizontal distance in the trajectory data, as follows: Based on the elevation of the road section, the height difference between the starting point and the end point of the trajectory is obtained, which is recorded as Δelevation; the horizontal distance corresponding to the trajectory is recorded as Δhorizontal distance; ; The abnormal energy consumption identification module classifies slopes into different categories based on their numerical values. It then calculates the proportion of each slope category in each vehicle's driving trajectory and constructs a feature vector. Each dimension in the feature vector corresponds to the proportion of a slope category. For example, the slope is divided into five categories: flat road, gentle downhill, gentle uphill, steep downhill, and steep uphill. The slope of the flat road is between -1° and 1°, the slope of the gentle downhill is between -2.5° and -1°, the slope of the gentle uphill is between 1° and 2.5°, the slope of the steep downhill is less than -2.5°, and the slope of the steep uphill is greater than 2.5°. On a certain road section, 60% of the road may be flat, 20% is a gentle downhill slope, 10% is a gentle uphill slope, 5% is a steep downhill slope, and 5% is a steep uphill slope. These proportions will form a feature vector, feature vector = [0.60, 0.20, 0.10, 0.05, 0.05]; among them, 0.60 represents the proportion of flat roads, 0.20 represents the proportion of gentle downhill slopes, 0.10 represents the proportion of gentle uphill slopes, 0.05 represents the proportion of steep downhill slopes, and 0.05 represents the proportion of steep uphill slopes. Calculate the cosine similarity of the feature vectors of the corresponding driving trajectories of every two vehicles in turn, and select the vehicles whose feature vectors of the driving trajectories are within the set cosine similarity threshold as the similar vehicle group C; By combining the vehicle's driving area and geographical features (such as slope, etc.) for analysis, the ability to identify vehicle driving characteristics is improved, external interference can be effectively eliminated, the accuracy of the analysis can be ensured, and it can adapt to different road conditions and transportation environments.

[0016] S24: perform similar transport feature identification; For vehicles in similar vehicle group B and similar vehicle group C, the abnormal energy consumption identification module calculates the average speed of each vehicle in a specific time period based on the timestamp and speed obtained from the trajectory data, and compares the average speed difference of the vehicles. If the average speed difference of the vehicles is less than the set average speed difference threshold, the vehicles have basically the same transportation time requirements, and vehicles with basically the same transportation time requirements are screened out; S25: perform transport weight identification; The abnormal energy consumption identification module identifies and corrects the total transport weight of the vehicle obtained in step S24 to obtain the transport weight of the vehicle in the corresponding time period; The method for predicting and correcting the gross transport weight is from the patent "Method for Gross Weight Identification and Confidence Assessment of Heavy Freight Vehicles Based on Cloud Computing" (Publication No. CN117951487A). The present invention only applies the gross weight identification method and does not innovate the gross weight identification method. The principle of the gross weight identification method will not be elaborated on in this invention. By analyzing the differences in vehicle driving speeds and combining transport time and transport weight to identify the vehicle's transportation demand characteristics, and comprehensively considering factors such as transport efficiency and transport weight, we ensure that the screening of high-energy-consuming vehicles is more accurate, which helps to identify high-energy-consuming vehicles that truly do not meet transportation requirements and improve the accuracy of energy efficiency monitoring.

[0017] S26: Eliminate abnormal data and obtain vehicles with similar transportation characteristics; The abnormal energy consumption identification module extracts the actual energy consumption of each vehicle from the energy consumption data uploaded by the vehicle bus based on the vehicle identification code corresponding to the vehicle obtained in step S25; Based on historical data and statistical methods, the actual energy consumption that deviates from the average is detected and eliminated. The remaining actual energy consumption corresponds to vehicles with similar transportation characteristics. By using historical data and statistical methods to eliminate data whose energy consumption deviates from the average value, we ensure that only vehicles within the normal range can participate in the final high-energy consumption screening. This can reduce the impact of extreme data or abnormal fluctuations on the final results, improve the stability and accuracy of the screening process, and ensure that the high-energy consumption vehicles finally identified are more representative.

[0018] S27: Screen out high energy consumption vehicles; The abnormal energy consumption identification module sets the ratio A; Based on the vehicle identification codes corresponding to the vehicles with similar transportation characteristics obtained in step S26, the fuel consumption rate of each vehicle is obtained from the vehicle data collected in step S1; The average fuel consumption rate of the vehicles is calculated using the average value calculation formula, and the vehicles are ranked by fuel consumption rate; the abnormal energy consumption identification module selects the vehicles ranked in the top A and whose fuel consumption rate exceeds the average value by a specific percentage as high-energy-consuming vehicles; For example, if the ratio A is set to 20% and the specific percentage by which the fuel consumption rate exceeds the average is 5%, then the vehicles in the top 20% of the fuel consumption rate ranking and whose fuel consumption rate exceeds the average by more than 5% are high-energy-consuming vehicles.

[0019] S3: Diagnose the cause of high energy consumption; The abnormal energy consumption cause diagnosis module pre-sets the fault energy consumption time / mileage ratio, high speed range / high rotation speed range, and difference threshold; The following sub-steps are included: S31: Perform energy consumption related fault diagnosis; The abnormal energy consumption cause diagnosis module uses intelligent diagnostic technology to identify whether a high-energy-consuming vehicle has a fault that significantly affects energy consumption during the analyzed time period. If so, the duration or mileage of the fault is compared with a pre-set fault energy consumption time / mileage ratio. If the duration or mileage ratio of the vehicle's fault is greater than the pre-set fault energy consumption time / mileage ratio, it is determined that one of the causes of the high energy consumption of the high-energy-consuming vehicle is an energy-related fault. For example, if the set fault energy consumption time / mileage ratio is 20%, if the duration or mileage of the vehicle fault ratio is greater than 20%, it is determined that one of the causes of high energy consumption of this high energy consumption vehicle is an energy consumption-related fault; S32: Diagnose bad driving behavior; The abnormal energy consumption diagnosis module evaluates driving behavior information. If there are frequent sudden acceleration or braking, long periods of idling, etc., it will determine that one of the causes of high energy consumption of this high-energy-consuming vehicle is bad driving behavior. S33: Perform operating condition difference diagnosis; The abnormal energy consumption cause diagnosis module obtains the proportion of high-energy-consuming vehicles in a set high-speed range and high-speed range based on the vehicle's speed and rotational speed, compares the proportion of high-energy-consuming vehicles in the set high-speed range / high-speed range with the proportion of non-high-energy-consuming vehicles in the set high-speed range / high-speed range, and if the difference exceeds a set difference threshold, determines that one of the causes of high energy consumption of the high-energy-consuming vehicle is a difference in operating conditions; Get diagnostic results for high energy consumption vehicles.

[0020] S4: Push the diagnosis results; The push module regularly pushes the diagnosis results to the business front desk through the diagnosis middle desk, and the business front desk notifies customers or relevant operation service personnel.

[0021] The present invention has been described with reference to the above embodiments. However, the above embodiments are merely exemplary embodiments of the present invention. It should be noted that the disclosed embodiments do not limit the scope of the present invention. On the contrary, modifications and improvements that do not depart from the spirit and scope of the present invention are intended to be protected by the present invention.

Claims

1. A method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the Internet of Vehicles, characterized by: The following steps are involved: S1: Collect vehicle data and preprocess it; The following sub-steps are included: S11: Real-time collection of vehicle data; S12: pre-processing vehicle data; S2: Identify abnormal energy consumption based on vehicle data and screen out high-energy-consuming vehicles; The following sub-steps are included: S21: performing vehicle similarity recognition; S22: performing similar route identification; S23: performing similar road feature recognition; S24: perform similar transport feature identification; S25: perform transport weight identification; S26: Eliminate abnormal data and obtain vehicles with similar transportation characteristics; S27: Screen out high energy consumption vehicles; S3: Diagnose the cause of high energy consumption; The following sub-steps are included: S31: Perform energy consumption related fault diagnosis; S32: Diagnose bad driving behavior; S33: Perform operating condition difference diagnosis; S4: Push the diagnosis results; The push module regularly pushes the diagnosis results to the business front desk through the diagnosis middle desk, and the business front desk notifies customers or relevant operation service personnel.

2. The method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the Internet of Vehicles according to claim 1, characterized in that: The specific content of step S1 is as follows: The following sub-steps are included: S11: Real-time collection of vehicle data; The data acquisition module collects vehicle data in real time through the vehicle terminal Tbox. The vehicle data includes vehicle identification code, vehicle status information, trajectory data, and fuel consumption; The vehicle identification code is unique; the vehicle status information includes the rotation speed; the trajectory data is composed of trajectory points, each of which includes real-time latitude and longitude, vehicle speed, timestamp, road section elevation, and horizontal distance; The data acquisition module collects driving behavior information in real time through sensors; The data acquisition module transmits the collected vehicle data to the data processing layer; S12: pre-processing vehicle data; The data processing layer receives vehicle data and preprocesses the vehicle data.

3. The method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the Internet of Vehicles according to claim 1, characterized in that: The specific content of step S2 is as follows: The abnormal energy consumption identification module is pre-set with the cosine similarity threshold of similar vehicles, similarity coefficient threshold, cosine similarity threshold of the characteristic vector of the driving trajectory, average speed difference threshold, and specific percentage of fuel consumption rate exceeding the average value; The following sub-steps are included: S21: performing vehicle similarity recognition; The abnormal energy consumption identification module divides the key parameters affecting commercial vehicle fuel consumption into categorical parameters and numerical parameters. By directly comparing the categorical parameters corresponding to each vehicle, vehicles with consistent categorical parameters are selected. Convert all numerical parameters corresponding to each vehicle with the same category parameters into vectors, one vehicle corresponding to one vector, and each dimension in the vector corresponding to a numerical parameter; calculate the cosine similarity of the vectors of each two vehicles in turn. The cosine similarity formula is as follows: ； Among them, C is the cosine similarity, A and B are the vectors of the two cars, is the vector dot product, and is the magnitude of each vector; Filter out vehicles whose cosine similarity is within the cosine similarity threshold as similar vehicle group A; S22: performing similar route identification; Use Kalman filter algorithm to denoise the trajectory data; The abnormal energy consumption identification module uses a grid method to divide the driving area involved in the vehicles in the similar vehicle group A into multiple small grids to obtain the total number of grids; all the trajectory points of each vehicle are mapped to the grids, and the grids that each vehicle has traveled through are marked; The similarity coefficient of each two vehicle's driving trajectories is calculated in sequence using the spatial overlap method. The specific calculation method is as follows: The number of times the driving trajectories of each two vehicles pass through the same grid is obtained in sequence to obtain the number of identical grids; the similarity coefficient is calculated using a similarity coefficient formula, wherein the similarity coefficient formula is: similarity coefficient = number of identical grids / total number of grids; Filter out vehicles whose similarity coefficients are within the similarity coefficient threshold from the similar vehicle group A as the similar vehicle group B; S23: performing similar road feature recognition; Route similarity takes precedence over road feature similarity. For vehicles that are in similar vehicle group A but not in similar vehicle group B, the abnormal energy consumption identification module calculates the slope of each vehicle's trajectory based on the road segment elevation and horizontal distance in the trajectory data, as follows: Based on the elevation of the road section, the height difference between the starting point and the end point of the trajectory is obtained, which is recorded as Δelevation; the horizontal distance corresponding to the trajectory is recorded as Δhorizontal distance; ； The abnormal energy consumption identification module classifies slopes into different categories based on their numerical values. It then calculates the proportion of each slope category in each vehicle's driving trajectory and constructs a feature vector. Each dimension in the feature vector corresponds to the proportion of a slope category. Calculate the cosine similarity of the feature vectors of the corresponding driving trajectories of every two vehicles in turn, and select the vehicles whose feature vectors of the driving trajectories are within the set cosine similarity threshold as the similar vehicle group C; S24: perform similar transport feature identification; For vehicles in similar vehicle group B and similar vehicle group C, the abnormal energy consumption identification module calculates the average speed of each vehicle in a specific time period based on the timestamp and speed obtained from the trajectory data, and compares the average speed difference of the vehicles. If the average speed difference of the vehicles is less than the set average speed difference threshold, the vehicles have basically the same transportation time requirements, and vehicles with basically the same transportation time requirements are screened out; S25: perform transport weight identification; The abnormal energy consumption identification module identifies and corrects the total transport weight of the vehicle obtained in step S24 to obtain the transport weight of the vehicle in the corresponding time period; S26: Eliminate abnormal data and obtain vehicles with similar transportation characteristics; The abnormal energy consumption identification module extracts the actual energy consumption of each vehicle from the energy consumption data uploaded by the vehicle bus based on the vehicle identification code corresponding to the vehicle obtained in step S25; Based on statistical methods, the actual energy consumption that deviates from the average is detected and eliminated. The remaining actual energy consumption corresponds to vehicles with similar transportation characteristics. S27: Screen out high energy consumption vehicles; The abnormal energy consumption identification module sets the ratio A; Based on the vehicle identification codes corresponding to the vehicles with similar transportation characteristics obtained in step S26, the fuel consumption rate of each vehicle is obtained from the vehicle data collected in step S1; The average fuel consumption rate of the vehicles is calculated using the average value calculation formula, and the vehicles are ranked by fuel consumption rate; the abnormal energy consumption identification module selects vehicles ranked in the top A and whose fuel consumption rate exceeds a specific percentage of the average value as high-energy-consuming vehicles.

4. The method for identifying abnormal energy consumption of commercial vehicles based on multi-dimensional data of the Internet of Vehicles according to claim 1, characterized in that: The specific content of step S3 is as follows: The abnormal energy consumption cause diagnosis module pre-sets the fault energy consumption time / mileage ratio, high speed range / high rotation speed range, and difference threshold; The following sub-steps are included: S31: Perform energy consumption related fault diagnosis; The abnormal energy consumption cause diagnosis module uses intelligent diagnostic technology to identify whether a high-energy-consuming vehicle has a fault that significantly affects energy consumption during the analyzed time period. If so, the duration or mileage of the fault is compared with a pre-set fault energy consumption time / mileage ratio. If the duration or mileage ratio of the vehicle's fault is greater than the pre-set fault energy consumption time / mileage ratio, it is determined that one of the causes of the high energy consumption of the high-energy-consuming vehicle is an energy-related fault. S32: Diagnose bad driving behavior; The abnormal energy consumption cause diagnosis module evaluates the driving behavior information and, if bad driving behavior is present, determines that one of the causes of high energy consumption of the high energy consumption vehicle is bad driving behavior. S33: Perform operating condition difference diagnosis; The abnormal energy consumption cause diagnosis module obtains the proportion of high-energy-consuming vehicles in a set high-speed range and high-speed range based on the vehicle's speed and rotational speed, compares the proportion of high-energy-consuming vehicles in the set high-speed range / high-speed range with the proportion of non-high-energy-consuming vehicles in the set high-speed range / high-speed range, and if the difference exceeds a set difference threshold, determines that one of the causes of high energy consumption of the high-energy-consuming vehicle is a difference in operating conditions; Get diagnostic results for high energy consumption vehicles.

5. A commercial vehicle abnormal energy consumption identification system based on multi-dimensional data of the Internet of Vehicles for implementing the method described in claims 1-4, characterized in that: It includes data acquisition module, data processing module, abnormal energy consumption identification module, abnormal energy consumption cause diagnosis module and push module; The data acquisition module is used to collect vehicle data in real time; The data processing module is used to pre-process the collected vehicle data; The abnormal energy consumption identification module is used to identify abnormal energy consumption of vehicles and screen out high energy consumption vehicles; The abnormal energy consumption cause diagnosis module is used to diagnose the cause of high energy consumption and obtain a diagnosis result; The push module is used to push the diagnosis results to relevant personnel.

Images