A remote operation and maintenance management platform for safe driving of intelligent fleets

Abstract

This invention relates to the field of fleet driving management technology, specifically to a remote operation and maintenance management platform for safe driving of intelligent fleets. The platform includes: a line graph construction unit for collecting driving data and GPS data for each vehicle, and obtaining driving line graphs and GPS line graphs respectively; a target vehicle acquisition unit for obtaining a first feature value sequence and a second feature value sequence based on the driving line graph and GPS line graph of a vehicle, and obtaining a difference feature value sequence to determine whether the vehicle is a target vehicle; a target time determination unit for obtaining the time to be corrected for the target vehicle, and determining whether the time to be corrected is the target time; and a correction unit for classifying the target time according to the distribution characteristics of the target time of a vehicle, and correcting the GPS data of the target time according to the classification results. This application can correct GPS data during the driving process of fleet vehicles.

Description

Technical Field

[0001] This invention relates to the field of fleet driving management technology, specifically to a remote operation and maintenance management platform for intelligent fleet safe driving. Background Technology

[0002] With the development of intelligent transportation and vehicle-to-everything (V2X) technologies, fleet management is rapidly upgrading towards intelligence and remote operation. Building an intelligent fleet safety driving remote operation and maintenance management platform that integrates positioning monitoring, driving behavior analysis, fault early warning, and task scheduling has become a key means to improve operational efficiency and ensure driving safety. In this platform, GPS positioning data, as the core basic information, is widely used in functions such as trajectory playback, boundary crossing alarms, and route optimization. Its accuracy directly affects the stability of system operation and the effectiveness of decision-making. However, affected by factors such as signal obstruction and multipath interference, GPS drift often occurs when vehicles are moving or stationary, leading to problems such as positioning jumps and abnormal trajectories, which in turn cause safety and management errors such as misjudged parking and false boundary crossings. Therefore, improving the accuracy of positioning information plays a crucial role in ensuring the safe driving and remote operation and maintenance management of fleets.

[0003] While existing filtering methods can correct GPS data drift, most still have limitations in practical applications. Some methods rely solely on fixed distance thresholds or simple filtering algorithms, making them difficult to adapt to real-world driving environments and prone to misjudgment or correction failure. For example, traditional models such as Kalman filtering rely heavily on fixed parameters when filtering GPS data, lacking the ability to dynamically respond to and incorporate factors like vehicle status. This results in unstable correction effects in complex scenarios, and the accuracy of the corrected data is difficult to guarantee. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention aims to provide a remote operation and maintenance management platform for the safe operation of intelligent vehicle fleets. The specific technical solution adopted is as follows:

[0005] One embodiment of the present invention provides a remote operation and maintenance management platform for safe driving of intelligent fleets, the platform comprising:

[0006] The line graph construction unit is used to collect driving data and GPS data for each vehicle; and to obtain driving line graphs and GPS line graphs based on the driving data and GPS data of a vehicle, respectively.

[0007] The target vehicle acquisition unit is used to acquire a first feature value sequence and a second feature value sequence based on a vehicle's driving line graph and GPS line graph, respectively; the absolute values ​​of the differences between the elements at corresponding times in the first and second feature value sequences form a difference feature value sequence; and the difference feature value sequence of a vehicle is used to determine whether the vehicle is the target vehicle.

[0008] The target time determination unit is used to obtain the time to be corrected of the target vehicle based on the difference feature value sequence of the target vehicle; and to determine whether the time to be corrected is the target time based on two adjacent times.

[0009] The correction unit is used to classify target times according to the distribution characteristics of a vehicle's target times, and to correct the GPS data of the target times based on the classification results.

[0010] Preferably, the driving data includes the vehicle's speed, acceleration, and direction of travel.

[0011] Preferably, the method involves obtaining a driving line graph and a GPS line graph based on a vehicle's driving data and GPS data, respectively, including:

[0012] Using the vehicle's direction of travel at the previous moment as the reference direction at that moment, the displacement between the two moments is obtained based on the vehicle speed at that moment and the acceleration at the next moment. The deflection angle of the vehicle's direction of travel at that moment relative to its reference direction is also obtained. This process is repeated to obtain the displacement between every two moments and the deflection angle of the vehicle's direction of travel at each moment relative to its reference direction. Then, a driving line graph is obtained with each moment as a node. The length of the line connecting two moments in the driving line graph represents the displacement distance between the two moments.

[0013] Based on GPS data of a vehicle at various times during its journey, the latitude and longitude coordinates of each time point are obtained, and adjacent latitude and longitude coordinate points are connected by line segments to obtain the GPS polyline map of the vehicle.

[0014] Preferably, obtaining the first feature value sequence and the second feature value sequence based on the driving line graph and GPS line graph of a vehicle, respectively, includes:

[0015] Using all nodes in the driving line graph except the first and last nodes as vertices of the angles; obtaining the lines connecting a vertex to its two adjacent nodes, taking the right-hand line as the starting edge and the left-hand line as the ending edge, calculating the angle from the starting edge to the ending edge in a counter-clockwise direction, and recording it as the angle value of the vertex at the corresponding time; the angle values ​​of each vertex in the driving line graph at the corresponding time form the first feature value sequence; similarly, obtaining the second feature value sequence based on the GPS line graph.

[0016] Preferably, determining whether a vehicle is the target vehicle using a sequence of differential feature values ​​includes:

[0017] Calculate the average value of the elements in the differential feature value sequence of a vehicle, and multiply it by the maximum value in the sequence to obtain the multiplication result. Normalize the multiplication result to obtain the vehicle correction necessity. If the vehicle correction necessity of the vehicle is greater than the judgment threshold, then the vehicle is the target vehicle.

[0018] Preferably, obtaining the time to be corrected for the target vehicle based on the difference feature value sequence of the target vehicle includes:

[0019] If the normalized value of the square of the differential feature value corresponding to a certain moment in the differential feature value sequence of the target vehicle is greater than the judgment threshold, then that moment is the moment to be corrected.

[0020] Preferably, determining whether a time to be corrected is a target time based on two adjacent times includes:

[0021] If at least one of the two adjacent times of a time to be corrected is not a time to be corrected, then the time to be corrected is not the target time; otherwise, the time to be corrected is the target time. All other times except the target time are normal times.

[0022] Preferably, the target time is classified according to the distribution characteristics of the target time of a vehicle, including:

[0023] The target time in which both adjacent times are normal times is categorized as the first type; the target time in which both ends of the continuous distribution are normal times is categorized as the second type; and the target time in which only one end of the continuous distribution is a normal time is categorized as the third type.

[0024] Preferably, the GPS data at the target time is corrected based on the classification results, including:

[0025] If the target time is of the first type, a two-dimensional coordinate system is established with the vehicle's travel direction at the previous adjacent time as the horizontal axis, the vehicle's latitude and longitude at the previous adjacent time as the origin, and the direction perpendicular to the horizontal axis as the vertical axis. In this two-dimensional coordinate system, the displacement between the previous adjacent time and the target time is decomposed into two component displacements. These two component displacements are then converted into longitude and latitude, denoted as the reference longitude and reference latitude corresponding to the previous adjacent time. Similarly, the reference longitude and reference latitude corresponding to the next adjacent time are obtained. An exponential function with a base of the natural constant is used to calculate the displacement between the previous adjacent time... The normalized squared values ​​of the difference eigenvalues ​​at each time step are negatively correlated to obtain the mapping value corresponding to the previous adjacent time step, and the mapping value corresponding to the next adjacent time step is obtained similarly. The weight corresponding to the previous adjacent time step is obtained by dividing the mapping value corresponding to the previous adjacent time step by the sum of the mapping values ​​corresponding to the previous and next adjacent time steps, and the weight corresponding to the next adjacent time step is obtained similarly. The corrected longitude corresponding to the target time step is obtained by weighting and summing the reference longitudes corresponding to the previous and next adjacent time steps using the weights corresponding to the previous and next adjacent time steps, and the corrected dimension corresponding to the target time step is obtained similarly.

[0026] For the third type of target time, the normal time at one end of the third type of target time is used as the reference time; the third type of target times are marked as the first target time, the second target time, and the third target time according to the order of the reference time toward the target time, until the last target time is marked as the last target time; the reference longitude and reference dimension corresponding to the reference time are obtained according to the vehicle's latitude and longitude at the reference time and the displacement between the first target time and the reference time, and are used as the corrected longitude and corrected dimension of the first target time; the reference longitude and reference dimension corresponding to the first target time are obtained according to the corrected longitude and corrected dimension of the first target time and its displacement with the second target time, and are used as the corrected longitude and corrected dimension of the second target time, and so on, to obtain the corrected longitude and corrected dimension of each target time in the third type of target time;

[0027] For the second type of target time, the normal times at both ends of the second type of target time are respectively denoted as the first reference time and the second reference time; the order from the first reference time toward the target time is denoted as the forward order, and the order from the second reference time toward the target time is denoted as the reverse order; using the method for obtaining the corrected longitude and corrected latitude of the target time using the third type of target time, starting from the first reference time and proceeding in the forward order, the forward corrected longitude and forward corrected latitude of each target time in the second type of target time are obtained sequentially, and similarly, starting from the second reference time and proceeding in the reverse order, the forward corrected longitude and forward corrected latitude of each target time in the second type of target time are obtained sequentially. The reverse-corrected longitude and reverse-corrected dimension of a time; based on the method for obtaining the weights corresponding to the previous adjacent time, obtain the weights corresponding to the forward and reverse adjacent target times of a target time in the second type of target time; use the weights corresponding to the forward and reverse adjacent target times to perform a weighted sum of the forward and reverse corrected longitude of a target time to obtain the corrected longitude of the target time; use the weights corresponding to the forward and reverse adjacent target times to perform a weighted sum of the forward and reverse corrected dimensions of a target time to obtain the corrected dimension of the target time.

[0028] The embodiments of the present invention have at least the following beneficial effects: This application collects driving data and GPS data for each vehicle, obtains corresponding driving line graphs and GPS line graphs respectively, and then obtains the difference feature value sequence of the vehicle based on the comparative analysis of the driving line graph and GPS line graph of the vehicle. It then determines whether the GPS data of the vehicle needs to be corrected, thereby identifying the vehicle that needs correction, i.e., the target vehicle, reducing the subsequent computational load. Furthermore, the time to be corrected of the target vehicle is obtained through the difference feature value sequence of the target vehicle, and the time to be corrected is determined based on the adjacent times of the target vehicle to determine whether the time to be corrected is the target time, further locking the time when the GPS data needs to be corrected, achieving the effect of accurate correction. Finally, the target vehicles of a vehicle are classified according to their distribution characteristics, and then the GPS data of the target time is corrected according to the classification results, thereby improving the accuracy of the vehicle GPS data, etc., so as to facilitate the subsequent remote management of vehicle data, etc. Attached Figure Description

[0029] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1A simplified architectural diagram of a remote operation and maintenance management platform for safe driving of an intelligent fleet, provided in an embodiment of the present invention;

[0031] Figure 2 A line graph diagram illustrating a remote operation and maintenance management platform for safe driving of an intelligent fleet provided in an embodiment of the present invention;

[0032] Figure 3 This is a line graph angle diagram of a remote operation and maintenance management platform for safe driving of an intelligent fleet provided in an embodiment of the present invention. Detailed Implementation

[0033] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a remote operation and maintenance management platform for safe driving of an intelligent fleet according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0034] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0035] The following description, in conjunction with the accompanying drawings, details the specific solution of the remote operation and maintenance management platform for safe driving of an intelligent fleet provided by the present invention.

[0036] In this embodiment, the main application scenario of the present invention is to build an adaptive and robust GPS drift correction system to improve the accuracy of positioning data and provide solid data support for safe driving and remote operation and maintenance management of fleets.

[0037] Please see Figure 1 This diagram illustrates a simplified architecture of a remote operation and maintenance management platform for safe driving of an intelligent fleet, provided by an embodiment of the present invention. The platform includes the following units:

[0038] The line graph construction unit is used to collect driving data and GPS data for each vehicle; based on the driving data and GPS data of a vehicle, a driving line graph and a GPS line graph are obtained respectively.

[0039] In the remote operation and maintenance management platform for safe driving of intelligent fleets, the platform comprehensively perceives the vehicle's operating status through vehicle terminals and various types of sensing devices, and uploads data to the cloud or dispatch center in real time through 4G / 5G communication networks to facilitate subsequent vehicle operation and maintenance management.

[0040] Furthermore, vehicle driving data needs to be collected, including vehicle speed, acceleration, and direction of travel. This allows us to obtain the vehicle speed, acceleration, and direction of travel (i.e., the vehicle's orientation) at each moment. Since the time interval between two moments is short, the acceleration at one moment is the average acceleration over the previous moment. Vehicle speed and acceleration are obtained from the vehicle's OBD interface or inertial measurement unit (IMU), while the vehicle's direction of travel is provided by a gyroscope or electronic compass. Additionally, GPS data for each moment needs to be acquired, which is obtained by a high-precision satellite positioning module. It should be noted that the driving data and GPS data are collected at the same frequency. The platform analyzes these collected data and processes the collected GPS data to improve the accuracy of vehicle positioning, facilitating subsequent remote operation and maintenance management of the vehicle.

[0041] In remote operation and maintenance management platforms for safe operation of intelligent fleets, GPS positioning data is fundamental to key functions such as vehicle trajectory monitoring, driving behavior analysis, and operation scheduling. However, during actual operation, the GPS data collected by vehicles may drift due to various factors, such as obstruction by tall buildings, electromagnetic interference, signal reflection under tunnels or overpasses, weak or insufficient satellite signals, etc. These interferences may cause momentary shifts, jumps, or discontinuous trajectories in positioning points, thus affecting the platform's accurate judgment of the vehicle's actual location and operating status. This can lead to problems such as false boundary violations, abnormal alarms, and errors in trajectory deviation analysis, reducing the platform's operational reliability and scheduling accuracy.

[0042] Therefore, in the remote operation and maintenance management of a fleet, improving the accuracy and stability of GPS data is of great significance for ensuring driving safety and efficient operation and maintenance.

[0043] To verify and correct the accuracy of vehicle GPS data, we will analyze data collected from any vehicle as an example.

[0044] Furthermore, a driving trajectory is represented by a driving line graph obtained from a vehicle's driving data. Specifically, the vehicle's direction of travel at the previous moment is used as the reference direction at that moment. Then, based on the vehicle's speed at that moment and the acceleration at the next moment, the displacement between the two moments is obtained, along with the deflection angle of the vehicle's direction of travel at that moment relative to its reference direction. This process is repeated to obtain the displacement between every two moments and the deflection angle of the vehicle's direction of travel at each moment relative to its reference direction. A driving line graph is then obtained with each moment as a node. The length of the line connecting two moments in the driving line graph represents the displacement distance between the two moments. The driving line graph is shown below. Figure 2 As shown.

[0045] In the diagram, tn represents the nth time, which is also the (n+1)th time. For example, t0, t1, t2, and t3 represent the initial time, the first time, the second time, and the third time in the vehicle data collection, which are also the first time, the second time, the third time, and the fourth time.

[0046] `atn` represents the deflection angle of the vehicle's direction of travel at time `tn` relative to its reference direction, with the vehicle's direction of travel preceding time `n` as the reference direction. `at1` represents the deflection angle of the vehicle's direction of travel at time `1` relative to its reference direction (i.e., the vehicle's direction of travel at the initial time). `X_tn` represents the displacement distance of the vehicle during the time interval between time `n` and its preceding time. `X_t1` represents the displacement distance of the vehicle during the time interval between time `1` and its preceding time (i.e., the initial time). `at2` represents the deflection angle of the vehicle's direction of travel at time `2` relative to its reference direction, with the vehicle's direction of travel at time `1` as the reference direction. `at3` represents the deflection angle of the vehicle's direction of travel at time `3` relative to its reference direction, with the vehicle's direction of travel at time `2` as the reference direction. This allows us to obtain a line graph showing the vehicle's movement during its journey.

[0047] Similarly, based on the GPS data of a vehicle at each sampling time during its journey, the corresponding latitude and longitude coordinates can be obtained. Connecting adjacent latitude and longitude coordinates with line segments also yields a line graph, which is denoted as a GPS line graph. In the two line graphs, each node corresponds to each time point.

[0048] The target vehicle acquisition unit is used to acquire a first feature value sequence and a second feature value sequence based on a vehicle's driving line graph and GPS line graph, respectively; the absolute values ​​of the differences between the elements at corresponding times in the first and second feature value sequences form a difference feature value sequence; and the difference feature value sequence of a vehicle is used to determine whether the vehicle is the target vehicle.

[0049] Based on the driving data and GPS data collected during each vehicle's journey, a driving line graph and a GPS line graph were obtained for each vehicle. If the GPS data collected during a vehicle's journey is relatively accurate, then theoretically, the shapes of the GPS line graph obtained from the GPS data and the driving line graph obtained during the vehicle's journey will have a high degree of similarity and consistency. This is because when the actual driving position of the vehicle deviates, the corresponding GPS data should also deviate. If the deviation in the actual driving position of the vehicle is inconsistent with the deviation reflected in the collected GPS data, it may indicate that the GPS data is drifting or inaccurate, and in this case, it may be necessary to correct the GPS data for that vehicle.

[0050] Because GPS data from not every vehicle experiences drift during operation, unnecessary corrections to GPS data that has no deviation or has deviations within a controllable range could increase the system's computational load and time. Therefore, corrections should only be performed on GPS data that may exhibit positioning drift.

[0051] Taking a vehicle's driving line graph and GPS line graph as examples, each node in the driving line graph is considered a vertex of an angle (excluding the node corresponding to the initial time and the node corresponding to the last time). A line connecting a vertex to its two adjacent nodes is obtained, with the right-hand line as the starting edge and the left-hand line as the ending edge. The angle from the starting edge to the ending edge is calculated counter-clockwise and recorded as the angle value at the corresponding time for that vertex. The angle values ​​at each vertex in the driving line graph at the corresponding time form the first feature value sequence. For example... Figure 3 As shown in the figure, T_2 represents the angle value of the vertex at time t2 in the driving line graph, which is used as the feature value of that vertex at that time. Similarly, the second feature value sequence is obtained based on the GPS line graph.

[0052] Then, the elements in the first feature value sequence and the second feature value sequence corresponding to a vehicle are aligned according to the sampling timestamp, and the absolute value of the difference between the elements at corresponding times in the first and second feature value sequences is obtained, which is the absolute value of the difference between the two feature values ​​at the same sampling time. Then, the absolute values ​​of the difference between the elements at corresponding times in the first and second feature value sequences are combined to form a difference feature value sequence, and each element in the sequence is a difference feature value.

[0053] Next, based on the differential feature value sequence corresponding to a vehicle, it is determined that the GPS data of that vehicle needs to be corrected.

[0054] Specifically, the average value of the elements in the differential feature value sequence is calculated and multiplied by the maximum value in the sequence to obtain the multiplication result. The multiplication result is then normalized to obtain the necessity of vehicle correction.

[0055] The specific calculation model for the necessity of vehicle correction is as follows:

[0056] ,

[0057] Where N represents the vehicle calibration necessity for a vehicle, indicating whether the GPS data collected by that vehicle needs correction; the larger the value, the greater the necessity for correction. `norm` represents the linear normalization function used for normalization, `m` represents the number of sampling times for the vehicle, and `m-2` represents the number of elements in the differential feature value sequence. This represents the i-th differential feature value in the differential feature value sequence corresponding to the vehicle. This represents the sequence of differential feature values ​​corresponding to the vehicle, and max() represents the maximum function. This represents the mean of all differential feature values ​​in the differential feature value sequence corresponding to the vehicle. The larger this value is, the greater the deviation between the vehicle's GPS data and its actual running trajectory may be, and the greater the possibility of GPS data drift. In other words, the greater the necessity for correcting the vehicle's GPS data. This represents the maximum value in the sequence of differential feature values. The larger the value, the higher the probability that there may be abnormal GPS data such as drift in the vehicle's GPS data, and the higher the necessity for correction.

[0058] Furthermore, a judgment threshold G is set, with a value of 0.2 (an empirical value, the specific value to be determined based on the actual needs of GPS data collected during vehicle operation and maintenance management; if the required positioning accuracy is higher, the threshold should be set smaller, otherwise it can be increased; that is, the positioning accuracy should be allowed to be within an acceptable range).

[0059] Finally, if N>G, that is, when the necessity of vehicle correction is greater than the judgment threshold, it indicates that the GPS data of the vehicle may be deviated and needs to be corrected. Otherwise, no correction operation will be performed in the subsequent analysis, and the vehicle whose necessity of vehicle correction is greater than the judgment threshold will be recorded as the target vehicle.

[0060] The target time determination unit is used to obtain the time to be corrected of the target vehicle based on the difference feature value sequence of the target vehicle; and to determine whether the time to be corrected is the target time based on two adjacent times.

[0061] Because the GPS data from the target vehicle may contain some offsets, while some data may be within the acceptable deviation range, it is only necessary to correct the GPS data that may exceed the acceptable deviation range.

[0062] To obtain GPS data for target vehicles that may exceed the deviation range for correction, and to obtain normal GPS data as a reference for the GPS data requiring correction, it is necessary to sequentially traverse the sequence of difference feature values ​​corresponding to each target vehicle to obtain the sampling time that needs correction.

[0063] Specifically, a ternary expression is constructed based on each differential feature value in the differential feature sequence, as follows: ,

[0064] in, The result of the judgment at time i in the differential feature sequence indicates whether the GPS data at time i needs to be corrected. If correction is needed, the result is True; otherwise, the result is False. Let Ti represent the square of the difference feature value corresponding to the i-th time in the difference feature sequence, and G represent the judgment threshold. Based on the above, the GPS data corresponding to the sampling time marked Ti is taken as the data time that the current target vehicle may need to be corrected, i.e., the time to be corrected. In other words, if the normalized value of the square of the difference feature value corresponding to a time in the difference feature value sequence of the target vehicle is greater than the judgment threshold, then that time is the time to be corrected.

[0065] Because a positioning deviation exists at one of the times to be corrected, it may affect the calculation of angular characteristic values ​​at times before and after it, thus causing the positioning at those times to appear as if there is a deviation. Therefore, the time points corresponding to the time to be corrected generally need to be at least two consecutive occurrences.

[0066] The system determines whether a time to be calibrated is a target time by comparing it with two adjacent times. If one or more of the two adjacent times are not target times, then the time to be calibrated is not a target time; otherwise, it is a target time. If the time to be calibrated is not a target time, then the GPS data corresponding to that time is considered normal GPS data and does not require calibration.

[0067] The time to be corrected that is not the target time is filtered out, and the remaining time to be corrected is the target time. For the sampling time, all times other than the target time are recorded as normal times.

[0068] The correction unit is used to classify target times according to the distribution characteristics of a vehicle's target times, and to correct the GPS data of the target times based on the classification results.

[0069] The distribution characteristics of target times vary throughout the entire time series, and therefore the correction methods also differ. Thus, target times need to be classified according to their distribution characteristics. Specifically, target times where both adjacent times are normal times are categorized as follows: Category 1; Category 2: Target times with continuous distribution where both ends are normal times; and Category 3: Target times with continuous distribution where only one end is normal. Here, "continuous distribution" refers to a continuous distribution in the time series.

[0070] Furthermore, the GPS data corresponding to the three types of target times are corrected separately. Specifically, if the target time is of the first type, a two-dimensional coordinate system is established with the vehicle's direction of travel at the previous adjacent time as the horizontal axis, the vehicle's latitude and longitude at the previous adjacent time as the origin, and the direction perpendicular to the horizontal axis as the vertical axis. In the two-dimensional coordinate system, the displacement between the previous adjacent time and the target time is decomposed to obtain two component displacements. These two component displacements are converted into longitude and latitude, and recorded as the reference longitude and reference latitude corresponding to the previous adjacent time. Similarly, the reference longitude and reference latitude corresponding to the next adjacent time are obtained. The previous phase is then corrected using an exponential function with a base of the natural constant. The normalized squared values ​​of the difference features at adjacent time points are negatively correlated to obtain the mapping value corresponding to the previous adjacent time point, and the mapping value corresponding to the next adjacent time point is obtained in the same way. The weight corresponding to the previous adjacent time point is obtained by dividing the mapping value corresponding to the previous adjacent time point by the sum of the mapping values ​​corresponding to the previous and next adjacent time points, and the weight corresponding to the next adjacent time point is obtained in the same way. The reference longitudes corresponding to the previous and next adjacent time points are weighted and summed using the weights corresponding to the previous and next adjacent time points to obtain the corrected longitude corresponding to the target time point, and the corrected dimension corresponding to the target time point is obtained in the same way.

[0071] The specific calculation model for the weight corresponding to the next adjacent time step is as follows:

[0072] ,

[0073] in, Let be the weight of the s-th time among the adjacent times before and after the target time t belonging to the first category, representing the reliability of the latitude and longitude information of the s-th time used to predict the target time; e represents the natural constant. The square of the difference characteristic value corresponding to the s-th time interval between consecutive time intervals is calculated. The smaller this value, the higher the accuracy of the GPS data at that time interval is likely to be. This may indirectly reflect that the accuracy of the GPS data at the normal time interval adjacent to the target time is likely to be higher. Therefore, the smaller this value, the higher the accuracy of the GPS data at that time interval is likely to be. This indicates that the higher the reliability of the GPS data corresponding to the selected reference time (adjacent times) for the target time, the better. The difference feature values ​​are obtained from the corresponding difference feature value sequence, where norm represents the normalization function. This represents the number of adjacent moments, which is 2. This indicates that the weight is obtained by weighting and normalizing the reliability of the predicted latitude and longitude information at the s-th time point among the selected preceding or following times corresponding to the target time t in the current analysis vehicle.

[0074] For the third type of target time, since only one end of the continuously distributed target time series has a normal time, meaning only the GPS data for the normal time at one end is accurate and does not require correction, the GPS data for this type of target time can only be corrected using the GPS data for the normal time at one end. The third type of target time can only be continuously distributed at the beginning or end of a time series. For example, if the starting time is a target time, and four consecutive times from the starting time are target times, and the fifth time is a normal time, then the first to the fifth target times constitute the third type of target time.

[0075] Specifically, the normal time at one end of the third type of target time is used as the reference time; the third type of target times are marked as the first target time, the second target time, and the third target time in order of reference time toward target time, until the last target time is marked as the last target time; the reference longitude and reference latitude of the vehicle at the reference time and the displacement between the first target time and the reference time are used to obtain the reference longitude and reference latitude, which are used as the corrected longitude and corrected latitude of the first target time; the reference longitude and reference latitude of the first target time are obtained based on the corrected longitude and corrected latitude of the first target time and its displacement with the second target time, which are used as the corrected longitude and corrected latitude of the second target time, and so on, to obtain the corrected longitude and corrected latitude of each target time in the third type of target time.

[0076] For example, if the distribution of the third type of target time is normal time, target time 1, target time 2, target time 3, and target time 4, then the normal time is the reference time, and target time 1, target time 2, target time 3, and target time 4 are the first target time, the second target time, the third target time, and the fourth target time. Based on the latitude and longitude of the vehicle at the reference time and the displacement between the first target time and the reference time, the reference longitude and reference latitude corresponding to the reference time are obtained, which are used as the corrected longitude and corrected latitude of the first target time. Similarly, the corrected longitude and corrected latitude of the first target time, the second target time, the third target time, and the fourth target time can be obtained.

[0077] For the second type of target time, since there are normal times at both ends of the continuously distributed target time, it is necessary to use these two normal times to correct the GPS data of the target time sandwiched in the middle from both ends, and then combine the correction results from both ends to obtain the final prediction result.

[0078] Specifically, the normal times at both ends of the second type of target time are designated as the first reference time and the second reference time, respectively. The order from the first reference time toward the target time is designated as the forward order, and the order from the second reference time toward the target time is designated as the reverse order. Using the method for obtaining the corrected longitude and corrected latitude of the target time from the third type of target time, starting from the first reference time and proceeding in the forward order, the forward corrected longitude and corrected latitude of each target time in the second type of target time are obtained sequentially. Similarly, starting from the second reference time and proceeding in the reverse order, the corrected longitude and corrected latitude of each target time in the second type of target time are obtained sequentially. The reverse-corrected longitude and reverse-corrected dimension are obtained; based on the method for obtaining the weights corresponding to the previous adjacent time, the weights corresponding to the forward and reverse adjacent target times of a target time in the second type of target time are obtained respectively; the forward-corrected longitude and reverse-corrected longitude of a target time are weighted and summed using the weights corresponding to the forward and reverse adjacent target times to obtain the corrected longitude of the target time; the forward-corrected dimension and reverse-corrected dimension of a target time are weighted and summed using the weights corresponding to the forward and reverse adjacent target times to obtain the corrected dimension of the target time.

[0079] In essence, the correction method for GPS data of the second type of target time is a combination of the correction methods for the first and second types of target times. This is because a continuous distribution of target times has GPS data from normal times at both ends that can be used as references for correction. For example, the distribution of the third type of target times is: Normal Time 1, Target Time 1, Target Time 2, Target Time 3, Normal Time 2. Normal Time 1 and Normal Time 2 are the first and second reference times, respectively. The order of the first reference time, Target Time 1, Target Time 2, and Target Time 3 is forward, while the order of Normal Time 2, Target Time 3, Target Time 2, and Target Time 1 is reverse. In this case, following the method used for obtaining the corrected longitude and corrected latitude of the target times in the third type of target times, starting from the first reference time, the forward-corrected longitude and corrected latitude of Target Time 1, Target Time 2, and Target Time 3 are obtained respectively. Then, the corrected longitude and corrected latitude of Target Time 1, Target Time 2, and Target Time 3 are obtained in reverse order. The reverse-corrected longitude and reverse-corrected dimension of target time 1, target time 2, and target time 3 are obtained. For target time 1, the forward adjacent time is the first reference time, and the reverse adjacent time is target time 2. According to the method of obtaining the weight of adjacent time in the first type of target time, the weights corresponding to the two forward and reverse adjacent times are obtained. Then, the forward and reverse corrected longitudes of target time 1 are weighted and summed, and the forward and reverse corrected dimensions of target time 1 are weighted and summed to obtain the corrected longitude and corrected dimension of target time 1.

[0080] This replaces the original latitude and longitude information (GPS data) of all target times with the corrected longitude and corrected latitude, thus completing the correction of the GPS data of the time that needs to be corrected during the vehicle's travel.

[0081] After calibrating the vehicle's GPS data, the platform can accurately monitor the fleet's operating status, identify abnormal behavior, trace driving routes, and optimize scheduling based on high-precision trajectories. This improves the security and intelligence of remote operation and maintenance management, and enables effective management of key aspects such as vehicle positioning, operating efficiency, driving behavior, and risk warning.

[0082] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0083] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0084] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A remote operation and maintenance management platform for the safe operation of an intelligent fleet, characterized in that, The platform includes: The line graph construction unit is used to collect driving data and GPS data for each vehicle; and to obtain driving line graphs and GPS line graphs based on the driving data and GPS data of a vehicle, respectively. The target vehicle acquisition unit is used to acquire a first feature value sequence and a second feature value sequence based on a vehicle's driving line graph and GPS line graph, respectively; the absolute values ​​of the differences between the elements at corresponding times in the first and second feature value sequences form a difference feature value sequence; and the difference feature value sequence of a vehicle is used to determine whether the vehicle is the target vehicle. The target time determination unit is used to obtain the time to be corrected of the target vehicle based on the difference feature value sequence of the target vehicle; and to determine whether the time to be corrected is the target time based on two adjacent times. The correction unit is used to classify the target time according to the distribution characteristics of the target time of a vehicle, and to correct the GPS data of the target time according to the classification results. The step of obtaining the first feature value sequence and the second feature value sequence based on the driving line graph and GPS line graph of a vehicle, respectively, includes: Using all nodes in the driving line graph except the first and last nodes as vertices of an angle; obtaining the line connecting a vertex to its two adjacent nodes, taking the right-hand line as the starting edge and the left-hand line as the ending edge, calculating the angle from the starting edge to the ending edge in a counter-clockwise direction, and recording it as the angle value of the vertex at the corresponding time; the angle values ​​of each vertex in the driving line graph at the corresponding time form the first feature value sequence; similarly, obtaining the second feature value sequence based on the GPS line graph; The method of determining whether a vehicle is the target vehicle using a sequence of differential feature values ​​of a vehicle includes: Calculate the average value of the elements in the differential feature value sequence of a vehicle, and multiply it with the maximum value in the sequence to obtain the multiplication result. Normalize the multiplication result to obtain the vehicle correction necessity. If the vehicle correction necessity of the vehicle is greater than the judgment threshold, then the vehicle is the target vehicle. The classification of target times based on the distribution characteristics of a vehicle's target time includes: The target time in which both adjacent times are normal times is categorized as the first type; the target time in which the two ends of the continuous distribution are both normal times is categorized as the second type; and the target time in which the two ends of the continuous distribution are only normal times is categorized as the third type. The step of correcting the GPS data at the target time based on the classification results includes: If the target time is of the first type, a two-dimensional coordinate system is established with the vehicle's travel direction at the previous adjacent time as the horizontal axis, the vehicle's latitude and longitude at the previous adjacent time as the origin, and the direction perpendicular to the horizontal axis as the vertical axis. In this two-dimensional coordinate system, the displacement between the previous adjacent time and the target time is decomposed into two component displacements. These two component displacements are then converted into longitude and latitude, denoted as the reference longitude and reference latitude corresponding to the previous adjacent time. Similarly, the reference longitude and reference latitude corresponding to the next adjacent time are obtained. An exponential function with a base of the natural constant is used to calculate the displacement between the previous adjacent time... The normalized squared values ​​of the difference eigenvalues ​​at each time step are negatively correlated to obtain the mapping value corresponding to the previous adjacent time step, and the mapping value corresponding to the next adjacent time step is obtained similarly. The weight corresponding to the previous adjacent time step is obtained by dividing the mapping value corresponding to the previous adjacent time step by the sum of the mapping values ​​corresponding to the previous and next adjacent time steps, and the weight corresponding to the next adjacent time step is obtained similarly. The corrected longitude corresponding to the target time step is obtained by weighting and summing the reference longitudes corresponding to the previous and next adjacent time steps using the weights corresponding to the previous and next adjacent time steps, and the corrected dimension corresponding to the target time step is obtained similarly. For the third type of target time, the normal time at one end of the third type of target time is used as the reference time; the third type of target times are marked as the first target time, the second target time, and the third target time according to the order of the reference time toward the target time, until the last target time is marked as the last target time; the reference longitude and reference dimension corresponding to the reference time are obtained according to the vehicle's latitude and longitude at the reference time and the displacement between the first target time and the reference time, and are used as the corrected longitude and corrected dimension of the first target time; the reference longitude and reference dimension corresponding to the first target time are obtained according to the corrected longitude and corrected dimension of the first target time and its displacement with the second target time, and are used as the corrected longitude and corrected dimension of the second target time, and so on, to obtain the corrected longitude and corrected dimension of each target time in the third type of target time; For the second type of target time, the normal times at both ends of the second type of target time are respectively denoted as the first reference time and the second reference time; the order from the first reference time toward the target time is denoted as the forward order, and the order from the second reference time toward the target time is denoted as the reverse order; using the method for obtaining the corrected longitude and corrected latitude of the target time using the third type of target time, starting from the first reference time and proceeding in the forward order, the forward corrected longitude and forward corrected latitude of each target time in the second type of target time are obtained sequentially, and similarly, starting from the second reference time and proceeding in the reverse order, the forward corrected longitude and forward corrected latitude of each target time in the second type of target time are obtained sequentially. The reverse-corrected longitude and reverse-corrected dimension of a time; based on the method for obtaining the weights corresponding to the previous adjacent time, obtain the weights corresponding to the forward and reverse adjacent target times of a target time in the second type of target time; use the weights corresponding to the forward and reverse adjacent target times to perform a weighted sum of the forward and reverse corrected longitude of a target time to obtain the corrected longitude of the target time; use the weights corresponding to the forward and reverse adjacent target times to perform a weighted sum of the forward and reverse corrected dimensions of a target time to obtain the corrected dimension of the target time.

2. The remote operation and maintenance management platform for safe driving of an intelligent fleet as described in claim 1, characterized in that, The driving data includes the vehicle's speed, acceleration, and direction of travel.

3. The remote operation and maintenance management platform for safe driving of an intelligent fleet as described in claim 1, characterized in that, The process of obtaining driving line graphs and GPS line graphs based on a vehicle's driving data and GPS data includes: Using the vehicle's direction of travel at the previous moment as the reference direction at that moment, the displacement between the two moments is obtained based on the vehicle speed at that moment and the acceleration at the next moment. The deflection angle of the vehicle's direction of travel at that moment relative to its reference direction is also obtained. This process is repeated to obtain the displacement between every two moments and the deflection angle of the vehicle's direction of travel at each moment relative to its reference direction. Then, a driving line graph is obtained with each moment as a node. The length of the line connecting two moments in the driving line graph represents the displacement distance between the two moments. Based on GPS data of a vehicle at various times during its journey, the latitude and longitude coordinates of each time point are obtained, and adjacent latitude and longitude coordinate points are connected by line segments to obtain the GPS polyline map of the vehicle.

4. The remote operation and maintenance management platform for safe driving of an intelligent fleet as described in claim 1, characterized in that, The step of obtaining the time to be corrected for the target vehicle based on the difference feature value sequence of the target vehicle includes: If the normalized value of the square of the differential feature value corresponding to a certain moment in the differential feature value sequence of the target vehicle is greater than the judgment threshold, then that moment is the moment to be corrected.

5. The remote operation and maintenance management platform for safe driving of an intelligent fleet as described in claim 1, characterized in that, The step of determining whether a time to be corrected is a target time based on two adjacent times includes: If at least one of the two adjacent times of a time to be corrected is not a time to be corrected, then the time to be corrected is not the target time; otherwise, the time to be corrected is the target time. All other times except the target time are normal times.

Images