Vehicle intelligent path matching system and method

Abstract

The invention discloses a vehicle intelligent path matching system and method. The method comprises the following steps: dividing vehicle driving data of a current path into a plurality of sub-segments; matching the vehicle driving data of the current path divided into the sub-segments with historical path driving data, and identifying the historical path with the highest matching degree; and calling the energy management strategy corresponding to the historical path with the highest matching degree as the energy management strategy of the vehicle. According to the method, the problem that the energy management strategy of the new energy automobile is difficult to implement under the condition of no navigation and no positioning information is solved, reliable basic data are provided for the energy management strategy by recording and matching the historical path and driving behavior information of the automobile, and the energy management capability of the automobile in various driving scenes is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of automobile intelligence and electrification, and specifically relates to a vehicle intelligent path matching system and method. Background Art

[0002] With the increasing popularity of new energy vehicles, energy management strategies are crucial for improving vehicle range and reducing energy consumption. Currently, the energy management strategies of most new energy vehicles rely on navigation information, optimizing energy usage through pre-set route planning. However, when driving on familiar routes, navigation is often not enabled, making energy management strategies difficult to implement. Some automakers use a map-free approach, predicting the vehicle's likely speed in the short term for limited energy management optimization. However, this approach can only achieve local optimization and is difficult to improve the overall energy management efficiency of the vehicle, resulting in limited success.

[0003] Although there are some energy management strategies based on driving habits in existing technologies, most of them focus on real-time data collection and simple historical data utilization, failing to fully explore the deep correlation between vehicle historical paths and driving behaviors. In addition, there are deficiencies in the accuracy and generalization capabilities of path matching, and it is impossible to effectively integrate multi-vehicle data for more comprehensive path and driving information matching. Summary of the Invention

[0004] In order to solve the problems described in the background art, the present invention proposes a vehicle intelligent path matching system and method.

[0005] A vehicle intelligent path matching system for achieving one of the objectives of the present invention includes:

[0006] Data processing module: used to divide the vehicle driving data of the current path into multiple sub-segments;

[0007] Data matching module: used to match the vehicle driving data of the current route after sub-segmentation with the driving data of the historical routes, and identify the historical route with the highest matching degree;

[0008] Energy management strategy determination module: used to call the energy management strategy corresponding to the historical path with the highest matching degree as the energy management strategy of the vehicle.

[0009] The technical effects of the above-mentioned vehicle intelligent path matching system include: optimizing energy management based on historical driving data without real-time navigation, solving the problem of strategy failure when the user does not turn on navigation, and getting rid of navigation dependence; realizing global strategy call through historical path matching, avoiding the local optimization limitations of the map-free mode, improving the energy management efficiency of the whole vehicle, and realizing globally optimized energy management; exploring the relationship between historical paths and driving behaviors, and using historical experience to improve the pertinence and effectiveness of energy management strategies.

[0010] Furthermore, the method of dividing the vehicle driving data of the current path into a plurality of sub-segments includes:

[0011] Sorting the vehicle driving data according to time or driving path to obtain sorted vehicle driving data;

[0012] Identify the marked data for the sorted vehicle driving data according to the preset significant event marking rules. When the vehicle driving data at a certain moment meets the significant event marking rules, the vehicle driving data at that moment is considered as marked data.

[0013] The sorted vehicle driving data is divided into sub-segments according to the marking data, and the vehicle driving data corresponding to the start time and the end time of each sub-segment are both marking data.

[0014] The technical effects of the above-mentioned sub-segmentation method include: breaking down continuous driving data into sub-segments with clear event boundaries (such as starting and stopping, turning, acceleration and deceleration, etc.), facilitating the refined matching of driving patterns to historical paths; sub-segments divided based on significant events are closer to actual driving behavior characteristics, avoiding the disorder of the original data, and improving the accuracy of path matching.

[0015] Furthermore, the method for identifying marked data from the sorted vehicle driving data according to the preset significant event marking rules includes:

[0016] When the absolute value of the vehicle's steering angle is greater than the steering angle threshold and the duration is greater than the first set time, the vehicle driving data at this moment is considered to be marked data;

[0017] When the absolute value of the acceleration is greater than the acceleration threshold and the duration is greater than the second set time, the vehicle driving data at this moment is considered to be marked data;

[0018] When the speed is equal to 0 and the duration is greater than the third set time length, the vehicle driving data at this moment is considered to be marked data.

[0019] The technical benefits of the aforementioned method for identifying labeled data from sorted vehicle driving data include: using "steering angle absolute value > threshold + duration" as labeled data; identifying typical driving events such as steering, distinguishing features such as curves and lane changes across different routes, and enhancing the behavioral correlation between sub-segments and historical routes; using steering event labeling to enable the system to more accurately identify similar driving scenarios (e.g., urban roads vs. highway curves), improving matching generalization capabilities, and optimizing route feature matching; using "acceleration absolute value > threshold + duration" as labeled data to distinguish driving styles, identifying aggressive driving behaviors such as sudden acceleration and deceleration, and facilitating matching of historical routes with similar driving habits (e.g., gentle driving vs. aggressive driving); invoking corresponding energy consumption optimization strategies based on acceleration characteristics (e.g., gentle acceleration to reduce power consumption), improving targeted energy management; using "speed = 0 + duration" as labeled data (e.g., stopping, waiting at traffic lights) to distinguish between driving and stationary states, optimizing energy management during stationary periods (e.g., idling energy consumption control); and using parking event labeling to more accurately match historical routes containing the same parking node, improving matching accuracy in complex scenarios such as urban roads.

[0020] Furthermore, if the data for a fourth set duration does not meet the preset significant event marking rules, the vehicle driving data is sub-segmented at fixed time intervals. This has the following technical benefits: when there are no significant events, the sub-segments are divided at fixed time intervals, avoiding segmentation interruptions caused by the absence of significant events, ensuring that all driving data is effectively utilized, and improving the comprehensiveness of the matching; and even on smooth roads without intense driving (such as high-speed cruising), reasonable segmentation can still be achieved, preventing strategic blind spots.

[0021] Furthermore, one method of identifying the historical path with the highest matching degree includes:

[0022] The temporal similarity measurement algorithm is used to calculate the path similarity between the current path and each historical path, and the historical paths with path similarity less than the set value are screened out to obtain the first candidate path set. The calculation method of path similarity is:

[0023]

[0024] d(a i ,b j ) is the sub-segment a of path A i With subsegment b of path B j The distance metric between them is calculated based on the feature differences of the sub-segments, such as steering angle differences, speed change differences, etc. M and N are the number of sub-segments of path A and path B respectively; min means taking the minimum value.

[0025] Calculate the temporal similarity between the driving data of each historical route in the first candidate route set and the driving data of the current route, sort the temporal similarity in ascending order, select the historical routes with a predetermined ratio, and obtain the second candidate route set; one method for calculating the temporal similarity includes: Δt = |t current -t history |;t current is the timestamp of the current path; t history The timestamp corresponding to the historical path;

[0026] The degree of matching between the driving intensity of the driving data of each historical route in the second candidate route set and the driving data of the current route is calculated, and the historical route with the smallest difference in driving intensity with the current route is the historical route with the highest matching degree finally identified.

[0027] The above method for identifying the most compatible historical routes utilizes a three-level process: "temporal similarity screening → temporal proximity sorting → driving intensity matching" to ensure similar driving patterns (e.g., acceleration-constant speed-deceleration sequences). This method prioritizes recent historical data to accommodate changes in driving habits over time, and matches similar driving styles (e.g., moderate / aggressive) to enhance strategy adaptability. By using set value screening and proportional sorting, historical routes with significant differences are eliminated, narrowing the matching scope and improving matching efficiency.

[0028] Furthermore, when using a temporal similarity measurement algorithm to calculate the path similarity between the current path and each historical path, the path similarity is also attenuated according to the timeliness of each path. The attenuation method includes: obtaining a path attenuation value based on the time difference between the current time and the current path, and the time difference between the current time and the historical path; and performing a weighted calculation on the path similarity between the current path and each historical path based on the path attenuation value to obtain a path similarity that takes into account the timeliness of the path. The path similarity calculation method includes:

[0029]

[0030] DTW(A,B) represents the path similarity between the current path A and the historical path B; w B is the path attenuation function of historical path B;

[0031] The technical benefits of attenuating path similarity based on the timeliness of each path include: introducing time-dependent decay (weighted by time differences) to path similarity. By increasing the weight of recent historical paths, this improves matching timeliness by avoiding the use of outdated data (e.g., using an old strategy after a user's driving style has changed). Furthermore, when integrating data from multiple vehicles, time-dependent decay can prioritize similar driving data from the same time period, enhancing the reliability of cross-vehicle matching.

[0032] Furthermore, the method further includes attenuating the time proximity by using the path attenuation degree. The calculation method of the time proximity degree includes:

[0033] Δt=|t current -t history |×w history ;t current is the timestamp of the current path; t history is the timestamp corresponding to the historical path; w history is the path attenuation function of the historical path.

[0034] The technical effects of using the path decay function to decay the degree of temporal proximity include: introducing timeliness decay (time difference weighting) to the degree of temporal proximity, and further avoiding calling outdated data and reducing the timeliness of matching by increasing the weight of recent historical paths.

[0035] Furthermore, the second method for identifying the historical path with the highest matching degree includes:

[0036] Traverse each historical path and use the temporal similarity measurement algorithm to calculate the path similarity between each historical path and the current path; calculate the temporal similarity between the driving data of the current path and the driving data of each historical path; calculate the matching degree of the driving intensity of the driving data of each historical path and the current path;

[0037] Perform a weighted calculation on the path similarity, time proximity, and driving intensity of each historical path and the current path to obtain the final weighted matching value of each historical path and the current path;

[0038] The historical path corresponding to the minimum final matching value is the identified historical path with the highest matching degree.

[0039] The technical effects of the above method for identifying the historical paths with the highest matching degree include: weighting the path similarity, temporal proximity, and driving intensity to calculate the final matching degree, integrating multi-dimensional parameters (rather than a single indicator) to calculate the matching degree, avoiding one-sidedness and ensuring that the historical path that best fits the current driving scenario is selected; at the same time, the strategy of flexibly configuring weights can adapt to different application scenarios (such as high-frequency commuting vs. occasional routes) by adjusting the weights of each dimension (such as focusing on temporal proximity or driving style), thereby improving generalization capabilities.

[0040] A vehicle intelligent path matching method for achieving the second objective of the present invention includes:

[0041] Divide the vehicle driving data of the current path into multiple sub-segments;

[0042] Match the vehicle driving data of the current route after the sub-segmentation with the driving data of the historical routes, and identify the historical route with the highest matching degree;

[0043] The energy management strategy corresponding to the historical path with the highest matching degree is called as the energy management strategy of the vehicle.

[0044] Furthermore, the method of dividing the vehicle driving data of the current path into a plurality of sub-segments includes:

[0045] Sorting the vehicle driving data according to time or driving path to obtain sorted vehicle driving data;

[0046] Identify the marked data of the sorted vehicle driving data according to the preset significant event marking rules;

[0047] The sorted vehicle driving data is divided into sub-segments according to the marking data, and the vehicle driving data corresponding to the start time and the end time of each sub-segment are both marking data.

[0048] Furthermore, one method of identifying marked data from the sorted vehicle driving data according to the preset significant event marking rules includes:

[0049] When the absolute value of the steering angle of the vehicle is greater than the steering angle threshold and the duration is greater than the first set time length, the vehicle driving data at this moment is considered to be marked data.

[0050] Furthermore, a second method of identifying marked data from the sorted vehicle driving data according to the preset significant event marking rules includes:

[0051] When the absolute value of the acceleration is greater than the acceleration threshold and the duration is greater than the second set time length, it is considered that the vehicle driving data at this moment is marked data.

[0052] Furthermore, a third method of identifying marked data from the sorted vehicle driving data according to the preset significant event marking rules includes:

[0053] When the speed is equal to 0 and the duration is greater than the third set time length, the vehicle driving data at this moment is considered to be marked data.

[0054] Furthermore, when the data within a fourth set time period do not meet the preset significant event marking rule, the vehicle driving data are divided into sub-segments at fixed time intervals.

[0055] Furthermore, the method of identifying the historical path with the highest matching degree includes:

[0056] A temporal similarity measurement algorithm is used to calculate the path similarity between the current path and each historical path, and the historical paths with path similarity less than the set value are screened out to obtain the first candidate path set;

[0057] Calculate the temporal similarity between the driving data of each historical route in the first candidate route set and the driving data of the current route, sort the temporal similarities in ascending order, and select the historical routes with the previously set ratio to obtain the second candidate route set;

[0058] The degree of matching between the driving intensity of the driving data of each historical route in the second candidate route set and the driving data of the current route is calculated, and the historical route with the smallest difference in driving intensity with the current route is the historical route with the highest matching degree finally identified.

[0059] Furthermore, when using a temporal similarity measurement algorithm to calculate the path similarity between the current path and each historical path, it also includes attenuating the path similarity according to the timeliness of each path. The attenuation method includes: obtaining a path attenuation value based on the time difference between the current time and the current path, and the time difference between the current time and the historical path; and performing a weighted calculation on the path similarity between the current path and each historical path based on the path attenuation value to obtain a path similarity that takes into account the timeliness of the path.

[0060] Furthermore, the method for identifying the historical path with the highest matching degree also includes:

[0061] Traverse each historical path and use a temporal similarity measurement algorithm to calculate the path similarity between each historical path and the current path; calculate the temporal similarity between the driving data of the current path and the driving data of each historical path; and calculate the matching degree of driving intensity between the driving data of each historical path and the current path.

[0062] Perform a weighted calculation on the path similarity, time proximity, and driving intensity of each historical path and the current path to obtain the final weighted matching value of each historical path and the current path;

[0063] The historical path corresponding to the minimum final matching value is the identified historical path with the highest matching degree.

[0064] Furthermore, it also includes a data acquisition module for collecting each driving path and driving information of the vehicle, and the driving information includes the vehicle's driving distance, speed, deceleration, steering angle, idling and parking data.

[0065] Furthermore, it also includes a data storage module for storing the information collected by the data collection module locally and / or in the cloud, and adding a timestamp to each collected information, the timestamp including the start time and end time of the trip.

[0066] Furthermore, in the data matching module, when a vehicle is matched, the driving data of other vehicles stored in the cloud is also matched with the driving data of the vehicle to be matched. The technical effects include: enabling cross-vehicle data sharing and matching, expanding the scope of matching data, and improving matching adaptability and accuracy.

[0067] A non-transitory computer-readable storage medium is provided to achieve the third objective of the present invention, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the vehicle intelligent path matching method are implemented.

[0068] A computer program product for achieving the fourth objective of the present invention includes a computer program / instruction, which, when executed by a processor, implements the steps of the vehicle intelligent path matching method.

[0069] The beneficial effects of the present invention include:

[0070] 1. The present invention solves the problem of traditional energy management strategies' strong dependence on the navigation system through deep matching of historical paths and driving behaviors. When the user is familiar with the path or actively turns off the navigation, the system can divide the current driving data into sub-segments with clear behavioral characteristics according to significant events (such as turning, acceleration, deceleration, parking, etc.) through the data processing module without the need for real-time positioning data, and then retrieve the path with the highest matching degree from the historical database through the data matching module. The implementation of energy management strategies is no longer limited by maps or real-time route planning, and is particularly suitable for complex urban road conditions, high-frequency commuting scenarios, or remote unmapped areas. It significantly improves the feasibility of energy management of vehicles in non-navigation environments, and fills the strategy gap of existing technologies in non-navigation scenarios.

[0071] 2. Different from the defects of the existing technology that simply relies on real-time data or rough segmentation, the present invention uses a three-level event marking rule (steering angle threshold + duration, acceleration threshold + duration, zero speed duration) to structure the driving data and divide the continuous driving process into sub-segments with clear behavioral characteristics (such as "sudden acceleration-constant speed-right turn-stop"). This segmentation method with physically significant events as boundaries effectively captures the key dynamic features of the driving process (such as driving style, road condition changes, node stops), so that the sub-segment data has higher scene recognition. For example, on urban roads, by identifying the sub-segment combination of "stop at a traffic light (zero speed mark) + sudden acceleration start (acceleration mark) + continuous turning (steering angle mark), the system can accurately match similar congested road driving patterns in historical data, avoiding the matching errors caused by insufficient data granularity of traditional methods, providing input conditions that are more in line with actual driving scenarios for energy management strategies, and making the optimization of control parameters such as braking energy recovery strategies and motor output power adjustment more targeted.

[0072] 3. The present invention constructs a three-level matching mechanism of "temporal similarity screening-time proximity sorting-driving intensity matching", and solves the problem of insufficient data matching generalization ability in the existing technology through the time-effect decay algorithm of path similarity and multi-parameter weighted calculation. Specifically, it includes: using algorithms such as dynamic time warping (DTW) to calculate the driving mode matching degree between the current path and the historical path, excluding invalid data with large differences in driving trajectories, and focusing on paths with similar acceleration-constant speed-deceleration sequences; by giving higher weights to recent historical data, adapting to the dynamic characteristics of driving habits that change with seasons and road conditions (such as the endurance decay strategy under low temperatures in winter needs to match recent similar data), avoiding the call of outdated and obsolete strategies; quantifying the driving intensity with parameters such as acceleration variance and steering frequency, ensuring that the optimal energy strategy is called separately in mild driving and aggressive driving scenarios (such as the former focuses on endurance optimization, and the latter focuses on power response). This multi-dimensional fusion matching algorithm can quickly locate the most similar historical experience in complex driving scenarios. Compared with traditional single indicator matching (such as relying only on speed curves), it improves the matching accuracy and significantly enhances the personalized adaptation ability of energy management strategies.

[0073] 4. The present invention breaks through the limitations of single vehicle data, realizes the shared matching of cross-vehicle historical routes and driving information through the cloud platform, and builds a closed-loop system of "single-vehicle data accumulation-multi-vehicle data collaboration-global strategy optimization". When a vehicle drives an unfamiliar route for the first time, the system can retrieve the historical data of other vehicles in similar time periods and similar driving styles in the cloud database to generate a temporary energy management strategy, solving the cold start problem of "no historical data available". For example, when driving on mountain roads, even if the vehicle has never passed through this section of road, it can optimize the battery output power and kinetic energy recovery intensity in advance by matching the energy consumption data of other vehicles in similar slope and curve scenarios, avoiding the energy waste caused by the lack of prior knowledge of traditional methods. In addition, the integration of multi-vehicle data upgrades the energy management strategy from "local optimization based on individual history" to "global optimization based on group data", which can reduce the energy consumption of the entire vehicle and improve the endurance and energy utilization efficiency of new energy vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 It is a structural diagram of an embodiment of the system of the present invention;

[0075] Figure 2 Schematic diagram of the definition of the driving sub-segment in the embodiment of the method of the present invention. DETAILED DESCRIPTION

[0076] The following specific embodiments are provided to explain the technical solutions of the present invention so that those skilled in the art can understand the present invention. The scope of protection of the present invention is not limited to the specific implementation structures described below. Any implementation schemes created by those skilled in the art that include the technical solutions of the present invention but differ from the following specific implementation schemes are also within the scope of protection of the present invention.

[0077] The embodiment of the present invention provides a vehicle intelligent path matching system, such as Figure 1 As shown, it includes the following modules:

[0078] Data acquisition module: uses sensors on the vehicle to record each driving path and driving information of the vehicle, including but not limited to the vehicle's driving distance, speed, deceleration, steering angle, idling and parking data.

[0079] Data storage module: used to store the data information collected by the data acquisition module. The storage is divided into local storage and cloud storage, and a timestamp is added to each record. The timestamp includes the start time and end time of the trip.

[0080] When the driving-related information of the historical route is uploaded to the cloud, when the vehicle is matched, it is not limited to the historical records of the vehicle itself, but also includes the records of other vehicles in the matching, realizing cross-vehicle data sharing and matching, expanding the scope of matching data, and improving matching adaptability and accuracy.

[0081] Data processing module: used to divide the vehicle driving data of the current path into several sub-segments, each sub-segment is marked with a significant event at the beginning and end, and the significant events include significant steering actions, significant driving actions, significant acceleration and deceleration actions, significant idling or parking actions, etc.

[0082] In one embodiment, the determination methods are as follows:

[0083] When the absolute value of the steering angle is greater than the steering angle threshold of 30° and the duration is greater than the first set time length of 1s, this action is considered to be a significant steering action and marked as a steering event;

[0084] When the absolute value of acceleration is greater than the acceleration threshold of 0.1g and the duration is greater than the second set time of 2s, the action is considered to be a significant acceleration or deceleration action and is marked as an acceleration or deceleration event;

[0085] When the speed=0 and the duration>the third set time length 10s, this action is considered to be a significant idling or parking action and is marked as a parking event.

[0086] If there are no significant events within a set continuous time period (such as 5 seconds), sub-segments are divided according to fixed time intervals (such as 30 seconds).

[0087] Taking the time of event occurrence as the endpoint, the driving behavior is divided into sub-segment sequence A=[a1,a2,...,a n ], each sub-segment contains: distance change Δs, speed curve v(t), acceleration average a_avg, steering angle peak θ_max, and braking intensity peak u_max.

[0088] Data matching module: used to match the vehicle driving data of the current path after sub-segmentation with the historical path driving data, and identify the historical path with the highest matching degree; every time the vehicle drives, the characteristics of each sub-segment during the driving process are continuously matched with the path sub-segment information recorded in the historical records from the moment the vehicle starts, and the most matching path is identified. The matching process is carried out according to the priority of the matching information. In this embodiment, the following are adopted: the highest priority is path matching, the second priority is time matching (traffic scenarios are likely to be similar within the same time, such as the congestion or smoothness of the same road section), and the third priority is the intensity of driving (i.e., matching with the user's driving habits). The historical path information with the highest matching degree is obtained, and as the current journey extends, the matching information is continuously updated, and the historical information matching target is dynamically adjusted.

[0089] Energy management strategy determination module: used to call the energy management strategy corresponding to the historical path with the highest matching degree as the energy management strategy of the vehicle; based on the historical path with the highest matching degree, implement the energy management strategy corresponding to the historical path.

[0090] The embodiment of the present invention also provides a vehicle intelligent path matching method, and the specific implementation steps are as follows:

[0091] 1. Data collection: Recording the driving information of each historical route of the vehicle: During the vehicle's driving process, the vehicle's driving distance, speed, acceleration, steering angle, braking intensity and other data are collected through on-board sensors and recorded according to time or route sequence to form multiple complete historical driving information records. These records can be stored in the vehicle's local storage device or uploaded to a cloud server.

[0092] 2. Data processing: Divide each driving behavior of the vehicle into several sub-segments: Divide the driving behavior into multiple sub-segments according to the preset significant event marking rules. Significant events include significant steering actions (such as steering angles exceeding a preset threshold), significant driving actions (such as acceleration or deceleration exceeding a preset amplitude and duration), significant deceleration actions (such as braking intensity exceeding a preset value), significant idling or parking actions (such as vehicle speed remaining at zero for more than a preset time), etc. The starting and ending points of each sub-segment are marked with these significant events, forming multiple driving behavior sub-segments with clear characteristics. The definition of the driving sub-segment in this embodiment is as follows: Figure 2 and as shown in Table 1.

[0093] Table 1 Definition of driving sub-segments

[0094]

[0095] 3. Data matching: Every time the vehicle starts to drive, from the moment the vehicle starts, the vehicle data of the current driving process is collected in real time, and the current driving process is divided into several sub-segments according to the same sub-segment division rules as the historical data. Then, the characteristics of the current driving sub-segment (including the vehicle distance change, speed curve characteristics, acceleration change characteristics, steering angle change characteristics, braking intensity change characteristics, etc. within the sub-segment) are continuously matched with the information in the historical records. The specific matching method is to match the characteristics of the above driving sub-segments according to priority. First, the path characteristics are the top priority matching items, that is, to find the historical path sub-segment that is most similar to the current driving sub-segment in terms of path shape, key node position, etc.; let the sub-segment sequence of the current driving path be A = [a1, a2, a3, ...], and the sub-segment sequence of the historical path be B = [b1, b2, b3, ...], and the degree of matching is determined by calculating the similarity of the two sequences. The similarity calculation can use the dynamic time warping (DTW) algorithm, and the formula is:

[0096]

[0097] Among them, w B is the path attenuation function of the historical path B, which is used to solve the problem of data timeliness and avoid the unreasonable impact of outdated data on the matching results; d(a i ,b j ) is sub-segment a i with b j The distance metric between them (which can be calculated based on the characteristic differences of the sub-segments, such as steering angle differences, speed changes, etc.), where M and N are the number of sub-segments in the current path A and the historical path B, respectively. This algorithm calculates the similarity between each sub-segment sequence of the historical path and the current driving path.

[0098] In some embodiments, a method for calculating a path attenuation function of a historical path includes:

[0099] w B =e -λΔTB ;

[0100] ΔTB is the time difference between the current time and the timestamp of historical path B, expressed in years. λ is the decay coefficient set based on actual conditions. For example, if λ is set to 1, the weight of data one year from the current time decays to 36.79%. This value can be optimized through real-vehicle data training to balance timeliness and historical data utilization.

[0101] The sub-segment sequences of the historical paths with DTW < a set value η1 (such as 0.4, not limited here) are screened to form a first candidate path set S1.

[0102] Secondly, the first candidate path set S1 is further screened with time as the second priority matching item, taking into account the similarity of traffic scenarios within the same time period. For example, during the same commuting time on weekdays, the congestion or smoothness of the same road section may be similar, and historical paths close to the current travel time are preferred. For example, if the current travel time is 9 am, if there are two historical paths with similar path matching, one of which is also around 9 am and the other is around 3 pm, the historical path around 9 am is preferred. The degree of temporal similarity can be measured by calculating the absolute value of the time difference, that is:

[0103] Δt=|t current -t history |×w history ………………(2)

[0104] t current is the timestamp of the current path; t history is the timestamp corresponding to the historical path; w history is the path attenuation function of the historical path; historical paths with smaller Δt are preferentially selected. The candidate path set S1 is sorted in ascending order by Δt, and the historical paths with a previously set ratio (eg, 20%) are retained to form a second candidate path set S2.

[0105] Next, the second candidate path set S2 is matched based on driving intensity. This means that the user's driving habits are considered and the driving intensity is evaluated by analyzing parameters such as vehicle acceleration and speed variation. Let the current driving intensity parameter be C (such as average acceleration, speed variation standard deviation, etc.) and the driving intensity parameter of a historical path be H. The Euclidean distance is used to calculate the difference between the current path A and the historical path B:

[0106]

[0107] Among them, n is the number of parameters involved in the calculation, k and H k The kth parameter values ​​of the driving data for the current route A and the driving data for the historical route are respectively selected. The historical route with the smallest difference in intensity from the current route (i.e., the smallest value of d(A, B)) is selected as the historical route that best matches the current route.

[0108] Through the above matching process, the historical route with the best matching degree and the corresponding driving information are found. This optimal historical route and driving information will serve as the basis for implementing the energy management strategy.

[0109] In some embodiments, one method of identifying the historical path with the highest matching degree includes:

[0110] Path feature screening: Set a similarity threshold η1 (e.g., similarity > 0.8), use the temporal similarity measurement algorithm to calculate the path similarity between the current path and each historical path, and screen out historical paths with path similarity less than the set value η1 to obtain the first candidate path set S1;

[0111] Time matching sorting: Calculate the temporal similarity (absolute value) between the driving data of each historical path in the first candidate path set S1 and the driving data of the current path, sort them in ascending order of temporal similarity, and retain the top k% (e.g., k = 20) of the paths to form the second candidate path set S2.

[0112] Driving intensity score: Calculate the Euclidean distance (e.g., the weighted sum of average acceleration and speed change standard deviation) of the paths in the second candidate path set S2, and select the historical path with the smallest difference from the current driving intensity (i.e., the smallest Euclidean distance d between the two paths in formula (3)) as the most matching historical path.

[0113] In some embodiments, a second method of identifying the historical path with the highest matching degree includes:

[0114] According to the above formula (1), the path matching degree P1 between the current driving path and the sub-segment sequence of each historical path is calculated;

[0115] According to the above formula (2), the time matching degree P2 between the current driving path and the sub-segment sequence of each historical path is calculated;

[0116] According to the above formula (3), the driving intensity matching degree P3 between the current driving path and the sub-segment sequence of each historical path is calculated;

[0117] The final matching value P between the current driving path and the sub-segment sequence of each historical path is calculated according to the following formula: total :

[0118] P total =P1×ω1+P2×ω2+P3×ω3

[0119] Where ω1, ω2, and ω3 are the path matching weight, time matching weight, and driving intensity matching weight set according to their importance, respectively. ω1, ω2, and ω3 are all positive values, and their sum is 1.

[0120] Final matching value P total The lowest historical path is the best matching historical path.

[0121] 4. Cloud Data Sharing and Matching: Once a vehicle's historical driving route information is uploaded to the cloud, it can leverage its own historical records when performing route matching. It can also access cloud servers to obtain driving characteristics of other vehicles on the same or similar road sections and incorporate these into the matching process. This allows the vehicle to match other vehicles' routes and driving information on the same road section, even if the vehicle has not driven on that road section before, for reference. This further expands the scope and diversity of matching data, improving matching accuracy and generalization capabilities.

[0122] In certain embodiments, the cloud server detects outliers in the uploaded data (e.g., filtering records with speeds greater than 200 km / h) and categorizes and stores data by vehicle type (new energy vehicle / fuel vehicle), region (city / highway), and time period (weekday / weekend) to improve matching efficiency. Furthermore, historical data older than a set period (e.g., more than one year) is automatically weighted down (e.g., multiplied by a factor of 0.8^t, where t is the year), prioritizing recent data.

[0123] Formulate energy management strategy: The energy management strategy is formulated in the following form:

[0124] When the historical route information of the vehicle is matched, the energy management strategy that matched the route in history can be called;

[0125] When the historical route information of other users of the same vehicle model is matched through the cloud server, the energy management strategy that matched the route in history can be called. If multiple pieces of the same route information are matched at the same time, the energy management strategy with the best energy consumption is selected. For hybrid vehicles, when the current oil price is much higher than the electricity price, the energy management strategy with the best energy consumption is the energy management strategy that reduces fuel consumption.

[0126] When the same historical route information of other users of other types of vehicles is matched through the cloud server, an energy management strategy corresponding to the route can be generated based on the global information of the matched route.

[0127] In some embodiments, when matching to another vehicle type (such as a fuel vehicle), the braking intensity data in the matched energy management strategy is converted. The conversion method includes:

[0128] According to the current battery status of the electric vehicle (SOC, temperature, etc.), the preset energy recovery efficiency coefficient table is consulted (for example, the coefficient is 0.6 when SOC ≥ 80%, and 0.9 when SOC ≤ 20%) to calculate the real-time conversion coefficient;

[0129] Calculate the regenerative torque; regenerative torque = braking intensity of a fuel vehicle × energy recovery efficiency coefficient. Combined with the maximum regenerative torque limit of the electric vehicle motor (to avoid exceeding physical limits), this generates effective control parameters.

[0130] The converted regenerative torque is integrated with the vehicle's historical energy management strategy and output to the motor through the vehicle controller (VCU) to achieve coordinated control of braking energy recovery and mechanical braking.

[0131] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0132] An embodiment of the present invention further provides a vehicle intelligent path matching system, comprising:

[0133] Data processing module: used to divide the vehicle driving data of the current path into multiple sub-segments;

[0134] Data matching module: used to match the vehicle driving data of the current route after sub-segmentation with the driving data of the historical routes, and identify the historical route with the highest matching degree;

[0135] Energy management strategy determination module: used to call the energy management strategy corresponding to the historical path with the highest matching degree as the energy management strategy of the vehicle.

[0136] In some embodiments, a method of dividing vehicle driving data of a current path into a plurality of sub-segments includes:

[0137] Sorting the vehicle driving data according to time or driving path to obtain sorted vehicle driving data;

[0138] The sorted vehicle driving data is identified as marked data according to a preset significant event marking rule. When the vehicle driving data at a certain moment meets the significant event marking rule, the vehicle driving data at that moment is considered as marked data. The preset significant event marking rule includes: the absolute value of the vehicle's steering angle is greater than the steering angle threshold and lasts for a first set time period; the absolute value of the vehicle's steering angle is greater than the steering angle threshold and lasts for a first set time period; the absolute value of the acceleration is greater than the acceleration threshold and lasts for a second set time period;

[0139] The sorted vehicle driving data is divided into sub-segments according to the label data, and the vehicle driving data corresponding to the start time and end time of each sub-segment are all label data, such as Figure 2 and as shown in Table 1.

[0140] In some embodiments, a method for identifying marked data from sorted vehicle driving data according to a preset significant event marking rule includes:

[0141] When the absolute value of the vehicle's steering angle is greater than a steering angle threshold (e.g., 30°) and lasts for a first set time (e.g., 1 second), the vehicle driving data at this moment is considered to be marked data;

[0142] When the absolute value of the acceleration is greater than the acceleration threshold (such as 0.1g) and lasts for a second set time (such as 2s), the vehicle driving data at this moment is considered to be marked data;

[0143] When the speed is equal to 0 and lasts for a third set time period (such as 10 seconds), the vehicle driving data at this moment is considered to be marked data.

[0144] In some embodiments, when the data within a fourth set time period (such as 5 seconds) does not meet the preset significant event marking rules, the vehicle driving data is divided into sub-segments according to fixed time intervals (such as 30 seconds).

[0145] In some embodiments, one method of identifying the historical path with the highest matching degree includes:

[0146] 1. Use the temporal similarity measurement algorithm to calculate the path similarity between the current path and each historical path, and filter out historical paths with path similarity less than the set value, or sort the path similarities from small to large, and take the historical paths corresponding to the path similarity of the previously set ratio to obtain the first candidate path set; one method for calculating path similarity is:

[0147]

[0148] d(a i ,b j ) is the sub-segment a of path A i With subsegment b of path B j The distance measurement between them (calculated based on the feature differences of the sub-segments, such as the difference in steering angles, the difference in speed changes, etc.), M and N are the number of sub-segments of path A and path B respectively; min means taking the minimum value;

[0149] For example: for sub-segment a i with b j When calculating path similarity based on steering angle differences and speed change differences, assuming that path A has 4 sub-segments, i = 4; path B has 2 sub-segments, j = 2, the path similarity between paths A and B is calculated as follows:

[0150] DTW(A,B)=min(d1(a1,b1),d1(a1,b2),d1(a2,b1),d1(a2,b2),d1(a3,b1),d1(a3,b2),d1(a4,b1),d1(a4,b 2), d2(a1,b1), d2(a1,b2), d2(a2,b1), d2(a2,b2), d2(a3,b1), d2(a3,b2), d2(a4,b1), d2(a4,b2)); among them, d1(a i ,b j ) represents the sub-segment a of path A i With subsegment b of path B j The distance measure of the steering angle between i ,b j ) represents the sub-segment a of path A i With subsegment b of path B j The speed between the distance measures.

[0151] In this embodiment, the distance metric is calculated using Euclidean distance, which is not limited in the present invention.

[0152] 2. Calculate the temporal similarity between the driving data of each historical route in the first candidate route set and the driving data of the current route, sort the temporal similarities in ascending order, and select the historical routes with the previously set ratio to obtain the second candidate route set;

[0153] In some embodiments, one method for calculating the degree of temporal proximity includes: Δt=|t current -t history |;t current is the timestamp of the current path; t history The timestamp corresponding to the historical path.

[0154] 3. Match the second candidate path set using driving intensity as the matching item. Calculate the degree of matching between the driving intensity of each historical path in the second candidate path set and the driving data of the current path. The historical path with the smallest difference in driving intensity with the current path is the historical path with the highest matching degree.

[0155] In some embodiments, the Euclidean distance is used to calculate the matching degree of driving intensity between the driving data of the historical route and the driving data of the current route;

[0156]

[0157] For example, the driving intensity of the current path and the historical path is calculated based on the acceleration and speed change of the vehicle; assuming that the current path A has 4 sub-segments, their speeds are C A,V1、C A,V2 、C A,V3 、C A,V4 , take its average speed as C A，ave_V =1 / 4(C A,V1 +C A,V2 +C A,V3 +C A,V4 ), whose accelerations are C A,a1 、C A,a2 、C A,a3 、C A,a4 , take its average acceleration as C A，ave_a =1 / 4(C A,a1 +C A,a2 +C A,a3 +C A,a4 ); Path B has two sub-segments, whose speeds are C B,V1 、C B,V2 , and their accelerations are C B,a1 、C B,a2 ; Take its average speed C B，ave_V =1 / 2(C B,V1 +C B,V2 ); take its average acceleration as C B，ave_a =1 / 2(C B,a1 +C B,a2 );but:

[0158]

[0159] In some embodiments, when using a temporal similarity measurement algorithm to calculate the path similarity between the current path and each historical path, the path similarity is also attenuated according to the timeliness of each path. The attenuation method includes: obtaining a path attenuation value based on the time difference between the current time and the current path, and the time difference between the current time and the historical path; and performing a weighted calculation on the path similarity between the current path and each historical path based on the path attenuation value to obtain a path similarity that takes into account the timeliness of the path. The second method for calculating path similarity includes:

[0160]

[0161] DTW(A,B) represents the path similarity between the current path A and the historical path B; w B is the path attenuation function of historical path B;

[0162] In certain embodiments, w B =e -λΔTB ; ΔTB is the time difference between the current time and the timestamp of historical path B, in years; λ is the attenuation coefficient set according to actual conditions.

[0163] If λ=1, the time difference between the timestamp of historical path B and the current time is 1 year, then ω B =e=0.3679, that is, the weight of data one year from the current time decays to 36.79%;

[0164] For example, if λ=0.5 and the time difference between the timestamp of historical path B and the current time is 1 year, then ω B =e -λ =0.6065, that is, the weight of data one year from the current time decays to 60.65%.

[0165] In some embodiments, the method for calculating the degree of temporal proximity includes: Δt=|t current -t history |×w history ;t current is the timestamp of the current path; t history is the timestamp corresponding to the historical path; w history is the path attenuation function of the historical path.

[0166] In some embodiments, a second method of identifying the historical path with the highest matching degree includes:

[0167] Traverse each historical path and use a temporal similarity measurement algorithm to calculate the path similarity between each historical path and the current path; calculate the temporal similarity between the driving data of the current path and the driving data of each historical path; and calculate the matching degree of driving intensity between the driving data of each historical path and the current path.

[0168] Perform a weighted calculation on the path similarity, time proximity, and driving intensity of each historical path and the current path to obtain the final weighted matching value of each historical path and the current path;

[0169] The historical path corresponding to the minimum final matching value is the identified historical path with the highest matching degree.

[0170] An embodiment of the present invention further provides a non-transitory computer-readable storage medium, which stores a computer program. The computer program includes program instructions, which implement the various steps of the method described in the present invention when executed by a processor, and will not be repeated here.

[0171] The computer-readable storage medium may be the data transmission device provided in any of the aforementioned embodiments or an internal storage unit of a computer device, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., provided on the computer device.

[0172] Furthermore, the computer-readable storage medium may include both an internal storage unit of the computer device and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by the computer device. The computer-readable storage medium may also be used to temporarily store data to be output or that has been output.

[0173] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0174] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0175] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0176] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0177] An embodiment of the present invention further provides a computer program product, comprising a computer program / instruction, which implements the steps of the vehicle intelligent path matching method when executed by a processor.

[0178] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

Claims

1. A vehicle intelligent path matching system, characterized in that: include: Data processing module: used to divide the vehicle driving data of the current path into multiple sub-segments; Data matching module: used to match the vehicle driving data of the current route after sub-segmentation with the driving data of the historical routes, and identify the historical route with the highest matching degree; Energy management strategy determination module: used to call the energy management strategy corresponding to the historical path with the highest matching degree as the energy management strategy of the vehicle.

2. The vehicle intelligent path matching system according to claim 1, characterized in that: The method of dividing the vehicle driving data of the current path into a plurality of sub-segments includes: Sorting the vehicle driving data according to time or driving path to obtain sorted vehicle driving data; Identify the marked data of the sorted vehicle driving data according to the preset significant event marking rules; The sorted vehicle driving data is divided into sub-segments according to the marking data, and the vehicle driving data corresponding to the start time and the end time of each sub-segment are both marking data.

3. The vehicle intelligent path matching system according to claim 2, characterized in that: The method for identifying marked data from the sorted vehicle driving data according to the preset significant event marking rules includes: When the absolute value of the steering angle of the vehicle is greater than the steering angle threshold and the duration is greater than the first set time length, the vehicle driving data at this moment is considered to be marked data.

4. The vehicle intelligent path matching system according to any one of claims 2 or 3, characterized in that: The method for identifying marked data from the sorted vehicle driving data according to the preset significant event marking rules includes: When the absolute value of the acceleration is greater than the acceleration threshold and the duration is greater than the second set time length, it is considered that the vehicle driving data at this moment is marked data.

5. The vehicle intelligent path matching system according to any one of claims 2 or 3, characterized in that: The method for identifying marked data from the sorted vehicle driving data according to the preset significant event marking rules includes: When the speed is equal to 0 and the duration is greater than the third set time length, the vehicle driving data at this moment is considered to be marked data.

6. The vehicle intelligent path matching system according to claim 1, characterized in that: Methods for identifying the most matching historical path include: A temporal similarity measurement algorithm is used to calculate the path similarity between the current path and each historical path, and the historical paths with path similarity less than the set value are screened out to obtain the first candidate path set; Calculate the temporal similarity between the driving data of each historical route in the first candidate route set and the driving data of the current route, sort the temporal similarities in ascending order, and select the historical routes with the previously set ratio to obtain the second candidate route set; The degree of matching between the driving intensity of the driving data of each historical route in the second candidate route set and the driving data of the current route is calculated, and the historical route with the smallest difference in driving intensity with the current route is the historical route with the highest matching degree finally identified.

7. The vehicle intelligent path matching system according to claim 6, characterized in that: When using a temporal similarity measurement algorithm to calculate the path similarity between the current path and each historical path, the path similarity is also attenuated according to the timeliness of each path. The attenuation method includes: obtaining a path attenuation value based on the time difference between the current time and the current path, and the time difference between the current time and the historical path; and performing a weighted calculation on the path similarity between the current path and each historical path based on the path attenuation value to obtain a path similarity that takes the timeliness of the path into account.

8. The vehicle intelligent path matching system according to claim 1, characterized in that: Methods for identifying the most matching historical path include: Traverse each historical path and use the temporal similarity measurement algorithm to calculate the path similarity between each historical path and the current path; calculate the temporal similarity between the driving data of the current path and the driving data of each historical path; calculate the matching degree of the driving intensity of the driving data of each historical path and the current path; Perform a weighted calculation on the path similarity, time proximity, and driving intensity of each historical path and the current path, and obtain the final weighted matching value of each historical path and the current path; The historical path corresponding to the minimum final matching value is the identified historical path with the highest matching degree.

9. The vehicle intelligent path matching system according to claim 1, characterized in that: It also includes uploading the driving data of the vehicle's historical path to the cloud; in the data matching module, when matching vehicles, it also includes matching the driving data of other vehicles stored in the cloud with the driving data of the vehicle to be matched.

10. A vehicle intelligent path matching method, characterized in that: include: Divide the vehicle driving data of the current path into multiple sub-segments; Match the vehicle driving data of the current route after the sub-segmentation with the driving data of the historical routes, and identify the historical route with the highest matching degree; The energy management strategy corresponding to the historical path with the highest matching degree is called as the energy management strategy of the vehicle.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the vehicle intelligent path matching method as claimed in claim 10 are implemented.

12. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the steps of the vehicle intelligent path matching method according to claim 10 are implemented.

Images