Detailed Description
The method mainly solves the problems that the traditional traffic abnormality processing mode has higher processing cost and relatively lower processing efficiency, and monitoring results cannot be identified and early-warned in real time. By fusing the multi-source traffic data, the traffic organization can be helped to better know the traffic condition, so that the traffic flow is better managed and controlled, and the occurrence of congestion and traffic accidents is reduced.
For a better understanding of the foregoing technical solutions, the following detailed description will be given with reference to the accompanying drawings and specific embodiments of the present invention:
Example 1
A method for processing data for multi-object fusion as shown in fig. 1, the method comprising:
Real-time monitoring is carried out on the target vehicle to obtain a real-time monitoring result of the vehicle;
Specifically, the target vehicle is monitored in real time, and the real-time monitoring result of the vehicle can be obtained in various modes. For example by in-vehicle sensors: various sensors, such as a speed sensor, an acceleration sensor, a brake sensor, a steering sensor, etc., are mounted on the target vehicle, and state and operation information of the vehicle can be acquired through these sensors. Video monitoring: the camera is used for carrying out real-time video monitoring on the target vehicle, and information such as the running state, the running track, the driver behavior and the like of the vehicle can be obtained from the video. GPS monitoring: by installing the GPS equipment on the target vehicle, the information of the position, the speed, the running track and the like of the vehicle can be monitored in real time. And (3) communication monitoring: by monitoring the communication between the target vehicle and other vehicles or traffic infrastructure, the information of the running state, the running track, the traffic condition and the like of the vehicle can be obtained. The information of the running performance, the safety condition, the behavior habit of the driver and the like of the vehicle can be obtained, and decision support is provided for vehicle management.
Performing feature recognition based on the real-time monitoring result of the vehicle to obtain a plurality of real-time operation features;
Specifically, based on the vehicle real-time monitoring results, a plurality of real-time operating features may be extracted and identified. All include speed features: by monitoring the speed of the vehicle, the running speed status of the vehicle, including the highest speed, the average speed, the running time, etc., can be obtained. Track characteristics: by monitoring the travel track of the vehicle, characteristics such as the travel path, the travel direction, the turning radius, and the like of the vehicle can be obtained. Behavior characteristics: by monitoring the driver behavior of the vehicle, characteristics of the driving operation, driving habit, driving style, and the like of the driver can be obtained. Safety features: by monitoring the safety conditions of the vehicle, the characteristics of the vehicle such as braking performance, safety belt service condition, collision condition and the like can be obtained. Traffic environment characteristics: by monitoring the traffic environment in which the vehicle is located, characteristics such as traffic flow, traffic flow speed, road conditions and the like can be obtained.
Performing abnormality recognition based on the real-time operation characteristics to obtain operation abnormality indexes;
Specifically, based on the plurality of real-time operation characteristics, a plurality of abnormal operation conditions of the vehicle can be identified, and corresponding abnormal operation indexes can be calculated. Speed abnormality index: by comparing the actual speed with the normal speed range, a speed abnormality index can be obtained for judging whether the vehicle is overspeed or running at a low speed or not. Trace abnormality index: by analyzing the change rule of the vehicle running track and the deviation of the actual running track, the track abnormality index can be obtained and used for judging whether the vehicle deviates from the normal running track. Behavioral abnormality index: by analyzing the operation behaviors and driving habits of the driver, the behavior abnormality index can be obtained and used for judging whether unsafe behaviors such as fatigue driving, drunk driving and the like exist in the driver. Safety abnormality index: the safety abnormality index can be obtained by analyzing the safety condition of the vehicle, and is used for judging whether the vehicle has potential safety hazards such as braking failure, tire breakage, oil tank oil leakage and the like. Traffic environment abnormality index: by analyzing the traffic environment information of the vehicle, the traffic environment abnormality index can be obtained and used for judging whether traffic conditions have conditions such as congestion, accidents, poor road conditions and the like.
Performing multi-objective fusion based on the operation abnormality indexes to obtain a vehicle operation abnormality index;
Specifically, based on a plurality of operational anomaly indexes, the operational anomaly indexes of the vehicle can be obtained by processing through a multi-objective fusion method. Data normalization: and (3) carrying out standardization processing on each abnormal operation index so as to enable the abnormal operation indexes to fall into the same numerical range, thereby facilitating comparison and analysis. Weight determination: and determining the weight of each index according to the importance of each abnormal operation index and the influence degree of the abnormal operation index on the running state of the vehicle. Fusion calculation: multiplying each abnormal running index by the corresponding weight, and accumulating the products to obtain the abnormal running index of the vehicle. The specific calculation formula can be selected according to actual requirements and conditions. Abnormality determination: according to the magnitude of the abnormal running index of the vehicle and the set threshold value, whether the abnormal running condition exists in the vehicle can be judged.
Judging whether the abnormal running index of the vehicle is larger than a preset abnormal index;
Specifically, according to the magnitude of the abnormal running index of the vehicle, the abnormal running index of the vehicle can be compared with a preset abnormal running index to judge whether the abnormal running condition exists in the vehicle. In general, if the vehicle operation abnormality index is greater than the preset abnormality index, it can be considered that the vehicle has an operation abnormality. The determination of the preset abnormality index needs to comprehensively consider various factors such as the type of the vehicle, the use condition, the running environment and the like, and is set according to actual requirements and conditions. Meanwhile, if various abnormal conditions exist in the vehicle, the influence degree of various abnormal conditions on the running of the vehicle needs to be comprehensively considered so as to accurately judge the abnormal conditions. The method has the advantages that the abnormal running index of the vehicle is obtained through the multi-target fusion method, whether the index is larger than the preset abnormal index or not is judged, real-time monitoring and abnormal judgment can be effectively carried out on the running state of the vehicle, and decision support is provided for vehicle management.
If the vehicle running abnormality index is larger than the preset abnormality index, a vehicle abnormality early warning signal is obtained;
specifically, when the vehicle operation abnormality index is greater than the preset abnormality index, a vehicle abnormality warning signal may be acquired. The early warning signal can be transmitted and expressed in various modes: and (3) sound alarm: and a sound alarm device is arranged in or outside the vehicle, and when the running abnormality index of the vehicle exceeds a preset value, the alarm device gives out alarm sound so as to remind a driver or a pedestrian of paying attention to the abnormal condition of the vehicle. And (3) light alarm: and a lamplight alarm device is arranged in or outside the vehicle, and when the running abnormality index of the vehicle exceeds a preset value, the alarm device emits flashing lamplight or red alarm lamplight so as to remind a driver or a pedestrian of paying attention to the abnormal condition of the vehicle. Alarming by mobile phone short messages: and sending the abnormal running index of the vehicle to a mobile phone of a driver or a vehicle owner in a mode of mobile phone short messages so as to remind the driver or the vehicle owner to timely treat abnormal conditions of the vehicle. And displaying on an electronic display screen: and displaying the abnormal running index and the early warning information of the vehicle on an electronic display screen of the vehicle so as to remind a driver or a pedestrian of timely knowing and processing the abnormal condition of the vehicle. The transmission and expression modes of the early warning signals are selected according to actual demands and conditions, and meanwhile, factors such as the acceptance of drivers or pedestrians, the response time and the like are required to be considered.
And carrying out abnormality early warning on the target vehicle according to the vehicle abnormality early warning signal.
Specifically, according to the obtained vehicle abnormality early warning signal, different abnormality early warning measures can be adopted for remote early warning: and the early warning signals are received through remote equipment such as a mobile phone, a computer and the like, and the target vehicle is subjected to remote early warning, such as sending a short message, making a call and the like. Short-range early warning: and a short-range early-warning device such as sound, light and the like is arranged around the target vehicle, and when an early-warning signal appears, the short-range early-warning device immediately sends out the early-warning signal to remind a driver or a pedestrian of paying attention to the abnormal condition of the vehicle. Linkage early warning: the early warning signals are transmitted to third party institutions such as traffic management departments, insurance companies and the like, and corresponding early warning measures such as notifying traffic police, insurance claims and the like are adopted by the institutions. Self-adaptive early warning: according to the abnormal condition of the vehicle, the transmission mode and the frequency of the early warning signal are automatically adjusted so as to better adapt to the early warning requirements under different conditions. The selection and the use of the abnormal early warning measures are carried out according to actual conditions and requirements, and meanwhile, factors such as the acceptance of drivers or pedestrians, the response time and the like, as well as the accuracy and the reliability of early warning signals are required to be considered.
Further, as shown in fig. 2, the method of the present application performs feature recognition based on the real-time monitoring result of the vehicle to obtain a plurality of real-time operation features, including:
Obtaining a multi-dimensional preset data cleaning index, wherein the multi-dimensional preset data cleaning index comprises data deduplication, missing value processing, data standardization and data verification;
Performing data cleaning on the real-time monitoring result of the vehicle based on the multidimensional preset data cleaning index to obtain real-time monitoring data of the vehicle;
And carrying out feature recognition on the real-time monitoring data of the vehicle to generate the plurality of real-time operation features.
Specifically, a multidimensional preset data cleaning index is obtained: data cleaning index: data deduplication, missing value processing, data normalization, and data validation. Data deduplication is the removal of duplicate data, ensuring that each data item is unique. The missing value processing is to process missing data, such as filling in missing values, deleting rows or columns containing missing values, and the like. Data normalization is the conversion of the distribution of data into a standard normal distribution to facilitate subsequent data analysis and processing. Data verification is to check the validity of data, such as checking whether the data conforms to an expected range or format, etc. And carrying out data cleaning on the real-time monitoring result of the vehicle based on the multidimensional preset data cleaning index: namely, according to the four indexes, the real-time monitoring result of the vehicle is processed. For example, for duplicated data, deletion or merging may be required; for missing data, it may be necessary to fill in default values or use interpolation methods. For non-standardized data, scaling or panning may be required; for data that does not meet expectations, checks or corrections may be required. Performing feature recognition on the real-time monitoring data of the vehicle to generate the plurality of real-time operation features: after data cleaning, feature recognition can be performed based on the cleaned data. Feature recognition is the identification of the intrinsic rules and features of data through exploratory analysis and statistical modeling of the data. These characteristics may include speed, acceleration, steering angle, driver behavior, etc. In this process, techniques such as clustering, classification, regression, etc., may need to be used for some data mining and machine learning.
Further, the method of the present application performs abnormality recognition based on the plurality of real-time operation features to obtain a plurality of operation abnormality indexes, including:
Acquiring a first real-time operation feature based on the plurality of real-time operation features;
Constructing a first abnormal operation identification unit;
Inputting the first real-time operation characteristic into the first operation abnormality identification unit to obtain a first operation abnormality index;
the first operational anomaly index is added to the plurality of operational anomaly indexes.
Specifically, a plurality of real-time operating characteristics are identified from vehicle real-time monitoring data. Here you mention obtaining the first real-time operation feature, constructing the first operation abnormality recognition unit, and inputting the first real-time operation feature to obtain the first operation abnormality index. Then you add this first operational anomaly index to the multiple operational anomaly indexes. This process may be in evaluating and monitoring the operating conditions of the vehicle to find possible operating anomalies or faults. Acquiring the first real-time operating characteristic comprises selecting a specific operating characteristic, such as a specific vehicle speed characteristic, from previous data processing steps, or selecting a specific characteristic according to certain specific real-time operating conditions, such as according to a driver's behavior characteristic. Constructing the first operational anomaly identification unit includes designing and training a model or algorithm that identifies anomalies or faults based on the input operational characteristics. For example, this may be a machine learning based classifier that distinguishes between normal and abnormal operation. The first real-time operation characteristic is input into a first operation abnormality identification unit to obtain a first operation abnormality index, and the process comprises taking the real-time operation characteristic as input, and obtaining an index representing the operation abnormality degree through the inference of a model or an algorithm. The first operational anomaly index is then added to a plurality of operational anomaly indexes by combining the anomaly index with other anomaly indexes derived from different operational characteristics to obtain a composite anomaly index. The comprehensive abnormality index can provide more comprehensive assessment of the overall running state of the vehicle, and is helpful for more accurate fault early warning or decision support.
Further, as shown in fig. 3, the method of the present application constructs a first abnormal operation identification unit, including:
Historical data query is carried out based on the first real-time operation characteristics, and a first operation abnormality identification record is obtained;
Performing data division based on the first abnormal operation identification record to obtain a first training sequence, a first test sequence and a first verification sequence;
And training, testing and verifying based on the first training sequence, the first testing sequence and the first verifying sequence to generate the first abnormal operation identification unit.
Specifically, historical data query is performed based on the first real-time operation characteristic, and a first operation abnormality identification record is obtained: this step queries the historical database or record for data similar to the first real-time operational characteristics and obtains a record of relevant operational anomalies for the data. These anomaly records are marked data for training and validating the first operational anomaly identification unit. Based on the first abnormal operation identification record, data division is carried out, and a first training sequence, a first test sequence and a first verification sequence are obtained: in this step, a training set, a test set, and a validation set are partitioned from the first operational anomaly identification record. The training set is used to train the model, the test set is used to evaluate the performance of the model (during the training phase and the validation phase), and the validation set is used to adjust the parameters of the model to optimize the performance. Training, testing and verifying based on the first training sequence, the first testing sequence and the first verifying sequence, and generating the first abnormal operation identifying unit: this step is the actual machine learning process that involves training the model over a first training sequence, then testing the performance of the model over a first test sequence, and verifying the performance of the model over a first verification sequence. Through this process, the first abnormal operation recognition unit may be generated or trained. This process may include, but is not limited to, selecting an appropriate machine learning algorithm, adjusting parameters of the model, dealing with data imbalance, etc. The final objective is to obtain a model or algorithm that can accurately identify operational anomalies when the real-time operational characteristics are input.
Further, the method of the present application performs multi-objective fusion based on the plurality of operation abnormality indexes to obtain a vehicle operation abnormality index, and includes:
carrying out abnormal importance identification based on the plurality of real-time operation features to obtain a plurality of feature abnormal importance levels;
performing duty ratio calculation based on the feature anomaly importance levels to obtain multi-objective fusion constraint;
And carrying out weighted calculation on the plurality of running abnormality indexes according to the multi-objective fusion constraint to generate the vehicle running abnormality index.
Specifically, through the recognition of the abnormal importance of a plurality of real-time operation features, the calculation of multi-objective fusion constraints is carried out according to the abnormal importance of the features, and finally, the weighted calculation is carried out on each operation abnormality index according to the constraints, so that the vehicle operation abnormality index is obtained. Carrying out abnormal importance identification based on a plurality of real-time operation features to obtain a plurality of feature abnormal importance levels: this step performs individual anomaly detection for each of the real-time operating characteristics and evaluates its importance in the overall operating state evaluation. For example, certain real-time operating characteristics may better predict or represent operating anomalies of the vehicle, and thus anomaly detection on those characteristics may result in a higher anomaly importance. Performing duty ratio calculation based on a plurality of characteristic abnormal importance degrees to obtain a multi-objective fusion constraint: this step may determine the weight or duty cycle of the individual features in the final vehicle operation anomaly index calculation based on their anomaly importance. This process needs to be determined based on the actual vehicle operating environment and requirements. Weighting calculation is carried out on a plurality of running abnormality indexes according to the multi-objective fusion constraint, and a vehicle running abnormality index is generated: this is a procedure of applying the previous steps to actual operation abnormality index calculation. Including weighted averaging of the individual operational anomaly indices according to their corresponding characteristic anomaly importance levels, or processing using other multi-objective decision methods. The goal of this process is to obtain a comprehensive vehicle operation anomaly index by identifying and fusing the anomaly importance of multiple real-time operating characteristics, which better reflects the overall operating state and possible anomalies of the vehicle.
Further, the method of the application further comprises:
Activating a vehicle abnormal operation and maintenance library according to the vehicle abnormal early warning signal;
Inputting the plurality of abnormal operation indexes into the abnormal operation and maintenance library of the vehicle to obtain an abnormal operation and maintenance scheme of the vehicle;
and carrying out abnormality management on the target vehicle according to the abnormal operation and maintenance scheme of the vehicle.
Specifically, according to the vehicle abnormality early warning signal, activating a vehicle abnormality operation and maintenance library: the method comprises the steps of calling or activating a vehicle abnormal operation and maintenance library stored in the system after receiving a vehicle abnormal early warning signal. This library may contain various operational policies and schemes for different anomalies. Inputting a plurality of operation abnormality indexes into a vehicle abnormality operation and maintenance library to obtain a vehicle abnormality operation and maintenance scheme: the method comprises the steps of taking a plurality of calculated operation abnormality indexes as input, and transmitting the operation abnormality indexes to a vehicle abnormality operation and maintenance library so as to obtain an operation and maintenance scheme aiming at the abnormality indexes. These schemes may include specific operational steps, repair advice, emergency countermeasures, or the like. Performing anomaly management on the target vehicle according to the vehicle anomaly operation and maintenance scheme: the method comprises the step of implementing corresponding abnormal management measures for the target vehicle according to the obtained operation and maintenance scheme. This may include inspection and repair of the device, adjustment of system parameters, or optimization of the run-time flow, etc. The process aims to effectively manage the abnormal conditions of the target vehicle in time through double management of early warning and operation and maintenance, and ensure safe, stable and efficient operation of the vehicle.
Example two
Based on the same inventive concept as the data processing method of multi-object fusion of the previous embodiments, as shown in fig. 4, the present application provides a multi-object fusion data processing apparatus, the apparatus comprising:
The real-time monitoring result acquisition module 10 is used for carrying out real-time monitoring on a target vehicle to obtain a vehicle real-time monitoring result;
The real-time operation feature acquisition module 20 performs feature recognition based on the real-time monitoring result of the vehicle to acquire a plurality of real-time operation features;
The abnormality index obtaining module 30 is configured to perform abnormality identification based on the plurality of real-time operation features, so as to obtain a plurality of operation abnormality indexes;
a vehicle operation abnormality index obtaining module 40, wherein the vehicle operation abnormality index obtaining module 40 performs multi-objective fusion based on the plurality of operation abnormality indexes to obtain a vehicle operation abnormality index;
The abnormality index determination module 50, wherein the abnormality index determination module 50 is configured to determine whether the abnormality index of the vehicle operation is greater than a preset abnormality index;
the vehicle abnormality early warning signal acquisition module 60, wherein the vehicle abnormality early warning signal acquisition module 60 obtains a vehicle abnormality early warning signal if the vehicle operation abnormality index is greater than the preset abnormality index;
The abnormality early-warning module 70 is configured to perform abnormality early-warning on the target vehicle according to the vehicle abnormality early-warning signal.
Further, the apparatus further comprises:
The cleaning index acquisition module is used for acquiring a multi-dimensional preset data cleaning index, wherein the multi-dimensional preset data cleaning index comprises data deduplication, missing value processing, data standardization and data verification;
the real-time monitoring data acquisition module is used for carrying out data cleaning on the real-time monitoring result of the vehicle based on the multidimensional preset data cleaning index to acquire real-time monitoring data of the vehicle;
And the real-time operation feature generation module is used for carrying out feature recognition on the real-time monitoring data of the vehicle and generating the plurality of real-time operation features.
Further, the apparatus further comprises:
the first real-time operation feature acquisition module is used for acquiring the first real-time operation features based on the plurality of real-time operation features;
the identification unit construction module is used for constructing a first abnormal operation identification unit;
the first operation index acquisition module is used for inputting the first real-time operation characteristic into the first operation abnormality identification unit to obtain a first operation abnormality index;
and the index adding module is used for adding the first operation abnormality index to the operation abnormality indexes.
Further, the apparatus further comprises:
The first operation anomaly identification record acquisition module is used for inquiring historical data based on the first real-time operation characteristic to acquire a first operation anomaly identification record;
the training sequence acquisition module is used for carrying out data division based on the first abnormal operation identification record to acquire a first training sequence, a first test sequence and a first verification sequence;
And the identification unit generation module is used for training, testing and verifying based on the first training sequence, the first testing sequence and the first verification sequence to generate the first abnormal operation identification unit.
Further, the apparatus further comprises:
the abnormal importance obtaining module is used for carrying out abnormal importance identification based on the plurality of real-time operation features to obtain a plurality of feature abnormal importance;
the multi-target fusion constraint acquisition module is used for carrying out duty ratio calculation based on the feature anomaly importance levels to obtain multi-target fusion constraint;
And the vehicle operation abnormality index generation module is used for carrying out weighted calculation on the operation abnormality indexes according to the multi-objective fusion constraint to generate the vehicle operation abnormality index.
Further, the apparatus further comprises:
The vehicle abnormal operation and maintenance library activation module is used for activating the vehicle abnormal operation and maintenance library according to the vehicle abnormal early warning signal;
the vehicle abnormal operation and maintenance scheme acquisition module is used for inputting the plurality of operation abnormality indexes into the vehicle abnormal operation and maintenance library to acquire a vehicle abnormal operation and maintenance scheme;
And the abnormality management module is used for carrying out abnormality management on the target vehicle according to the vehicle abnormality operation and maintenance scheme.
The foregoing detailed description of the multi-object fusion data processing method will be clear to those skilled in the art, so that the description is relatively simple, and the relevant parts will be referred to in the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.