An ai-based electric vehicle positioning data assisted correction method and system

Abstract

The present application relates to the technical field of data processing, and particularly relates to an electric vehicle positioning data auxiliary correction method and system based on AI. The method comprises the following steps: acquiring positioning data sequences containing multiple dimensions of each historical moment in the historical journey of an electric vehicle; determining the movement coefficients of the positioning data sequences of each historical moment in each dimension; acquiring categories of multiple electric vehicle driving conditions, and determining the confidence of each category and each vehicle-mounted sensor; determining the correction value of the current moment in each vehicle-mounted sensor in the current journey; and based on the correction value, using an AI algorithm to realize the prediction of the positioning data of the current moment. The present application calculates the positioning data movement coefficient and clusters the driving conditions, and corrects the current sensor value in combination with the sensor confidence, thereby effectively inhibiting the trajectory deviation caused by the cumulative error of the vehicle-mounted sensor in the satellite signal shielding scene such as urban canyons and tunnels, and making the positioning closer to the real path.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to an AI-based electric vehicle positioning data auxiliary correction method and system. BACKGROUND

[0002] With the continuous improvement of the intelligence level of electric vehicles (new energy vehicles), accurate positioning has become a key technology to ensure the accuracy of navigation, the rationality of path planning, and the normal operation of safety auxiliary driving functions. Currently, electric vehicle positioning mainly relies on satellite positioning systems (such as GPS, Beidou, etc.) and data fusion schemes of vehicle-mounted sensors (such as inertial measurement units IMU, wheel speed sensors, etc.). This scheme can provide high-precision positioning results in open environments.

[0003] However, in urban canyons (streets lined with high-rise buildings), tunnels, underground parking lots, and other scenarios, satellite signals are easily blocked or interfered, resulting in signal weakening or even loss. In this case, the existing technology usually adopts a dead reckoning (DR) method, which is based on the positioning point before the satellite signal is blocked and uses vehicle-mounted sensor data (such as acceleration, angular velocity, etc.) to calculate the trajectory. However, this scheme also has significant defects: as time goes on, the measurement error of the sensor will continuously accumulate, leading to serious trajectory deviation of the positioning result, and even causing navigation failure, automatic driving decision errors, and other problems. In addition, the accumulation characteristics of sensor errors differ greatly under different driving scenarios (such as sudden acceleration, turning, constant speed driving, etc.), and traditional single calculation models are difficult to adapt to complex and variable driving conditions, further exacerbating positioning errors. Therefore, how to effectively suppress the error accumulation of vehicle-mounted sensors in satellite signal loss scenarios and achieve high-precision electric vehicle positioning has become a problem to be solved in the current electric vehicle positioning field. SUMMARY

[0004] To solve the problem of error accumulation in electric vehicle positioning caused by the existing electric vehicle positioning technology mainly relying on satellite positioning and data fusion of vehicle-mounted sensors, but in urban canyons, tunnels, and other satellite signal blocked scenarios, the positioning of the electric vehicle is the positioning point before the satellite signal is blocked, and the trajectory is calculated using vehicle-mounted sensor data, the present application provides an AI-based electric vehicle positioning data auxiliary correction method and system.

[0005] In a first aspect, the present application provides an AI-based electric vehicle positioning data auxiliary correction method, which adopts the following technical scheme:

[0006] An AI-based method for assisting in the correction of electric vehicle positioning data includes: acquiring a multi-dimensional positioning data sequence for each historical moment in the electric vehicle's historical journey; sliding a window of a preset length across the positioning data sequence; for any historical moment, when the center of the sliding window is located at that moment, determining the movement coefficient of the positioning data sequence in each dimension at that moment based on the standard deviation of the positioning data of all historical moments within the sliding window and the positioning data of historical moments at both ends of the sliding window in each dimension; clustering the historical moments based on the movement coefficient to obtain multiple categories of electric vehicle driving conditions; and determining the movement coefficient based on the standard deviation of the positioning data of all historical moments within each category. The standard deviation of the values ​​of each onboard sensor in the electric vehicle at historical moments, and the standard deviation of the movement coefficients of the positioning data sequences of all historical moments in each dimension within each category, are used to determine the confidence level of each category and each onboard sensor. Based on the current moment in this trip and the values ​​of each onboard sensor at all historical moments within each category, the reference category and the correction direction for the current moment are obtained. According to the confidence level of the reference category and each onboard sensor, and the correction direction, the values ​​of the current moment in each onboard sensor are corrected to obtain the corrected values ​​of the current moment in each onboard sensor. Based on the corrected values, an AI algorithm is used to predict the positioning data at the current moment.

[0007] The beneficial effects are as follows: By utilizing the statistical characteristics (such as standard deviation and motion coefficient) of historical positioning data and vehicle sensor data, combined with AI algorithms, the current positioning is corrected, effectively reducing the accumulation of errors caused by relying solely on vehicle sensor calculations after satellite signal blockage; by employing sliding window analysis and clustering methods, the confidence level of vehicle sensor data under different driving conditions of electric vehicles is assessed, making the correction more targeted, thereby improving positioning accuracy and navigation stability in complex environments (such as urban canyons and tunnels); through AI-assisted methods, the system can dynamically learn and adapt to various driving conditions and environmental changes, enhancing the system's ability to perceive vehicle status, improving the robustness of the positioning algorithm, and reducing the risk of positioning failure or significant trajectory deviation; based on the current correction value, positioning data is quickly predicted, providing more accurate location information for the electric vehicle navigation system, assisting in route planning and driving decisions, and improving the user's driving experience and safety.

[0008] Furthermore, the dimensions include longitude, latitude, and altitude.

[0009] Furthermore, the mobility coefficient satisfies:

[0010] In the formula, For the first The historical moment at the The movement coefficient of the positioning data sequence in each dimension. a standard deviation of the positioning data of all the historical time points in the first dimension, a standard deviation of the positioning data of all the historical time points in the first dimension, a standard deviation of the positioning data of all the historical time points in the first dimension, a standard deviation of the positioning data of all the historical time points in the first dimension, a standard deviation of the positioning data of all the historical time points in the first dimension, a standard deviation of the positioning data of all the historical time points in the first dimension, a standard deviation of the positioning data of all the historical time points in the first dimension, a standard deviation of the positioning data of all the historical time points in the first dimension, a standard deviation of the positioning data of all the historical time points in the first dimension, a standard deviation of the positioning data of all the historical time points in the first dimension,

[0011] The beneficial effects are that the movement coefficient effectively depicts the position change trend of each historical time point in different dimensions by combining the local standard deviation and the global standard deviation in the sliding window and the difference of the positioning data at both ends, and can reflect the dynamic fluctuation characteristics of the data; using the difference value of the positioning data at both ends as the weight, the historical time point with significant position change is given a higher movement coefficient, which helps to identify the possible mutation points or abnormal states in the trajectory, and improves the accuracy of subsequent clustering and classification.

[0012] Further, the clustering adopts DBSCAN clustering.

[0013] Further, the confidence degree satisfies:

[0014] ; in the formula, a confidence degree of the first category and the first vehicle-mounted sensor, a confidence degree of the first category and the first vehicle-mounted sensor, a confidence degree of the first category and the first vehicle-mounted sensor, a standard deviation of the values of the first vehicle-mounted sensor data of all the historical time points in the first category, a standard deviation of the values of the first vehicle-mounted sensor data of all the historical time points in the first category, a standard deviation of the values of the first vehicle-mounted sensor data of all the historical time points in the first category, a number of categories of the vehicle-mounted sensors, a standard deviation of the movement coefficients of the positioning data sequences in the first dimension of all the historical time points in the first category, a standard deviation of the movement coefficients of the positioning data sequences in the first dimension of all the historical time points in the first category, a number of dimensions, a number of categories, a number of categories, a maximum value function, a natural exponential function.

[0015] The beneficial effects are that: the confidence index combines the dispersion degree (standard deviation) of the vehicle-mounted sensor data and the fluctuation range (standard deviation of the moving coefficient) of the positioning data characteristics, is weighted by an exponential function, can quantify the reliability of each category (driving condition, such as straight driving, left turn, etc.) and each sensor, and comprehensively reflects the performance of the sensor in this category; the maximum function reflects the influence of the most significant fluctuation dimension in the positioning data on the confidence, which is helpful to identify the cases where the sensor data and the positioning change are least matched, so as to give reasonable weight to the sensor data; through continuous evaluation and updating of the confidence, the system can identify and adapt to the changes of sensor performance and complex environmental interference, enhance the resistance of the positioning algorithm to abnormal data, and improve the overall positioning accuracy and stability.

[0016] Further, the reference category of the current moment and the correction direction of the current moment include: based on the values of each vehicle-mounted sensor at the current moment in the current trip, a sequence of the current moment is constructed, based on the mean value of the values of each sensor at all historical moments in each category, a sequence of each category is constructed, the category with the largest cosine similarity between the sequence of the current moment and the sequence of each category is recorded as the reference category of the current moment; and the correction direction of the current moment in each vehicle-mounted sensor is determined according to the size of the values of each vehicle-mounted sensor at the current moment and the reference category.

[0017] The beneficial effects are that: by calculating the cosine similarity of the current vehicle-mounted sensor data and the mean value sequence of each category, the current driving state of the electric vehicle can be accurately matched, effective classification and attribution of the positioning data in complex driving environment can be realized; by comparing the sensor data of the similar historical moment in the current moment and the reference category, the offset trend of each sensor data can be scientifically determined, the correction direction can be guided, and the influence of sensor noise or abnormality on positioning accuracy can be reduced.

[0018] Further, the correction direction satisfies:

[0019] ; in the formula, is the correction direction of the current moment in the kth vehicle-mounted sensor, is the value of the current moment in the kth vehicle-mounted sensor, is the value of the current moment in the kth vehicle-mounted sensor, is the mean value of the values of the kth vehicle-mounted sensor at all historical moments in the reference category. Further, the correction value satisfies:

[0020]

[0021] ; in the formula, is the correction value of the current moment in the kth vehicle-mounted sensor, is the value of the current moment in the kth vehicle-mounted sensor, ​​The reference category at the current moment is in the 1st... Correction direction for vehicle-mounted sensors, The reference category for the current moment and the first Confidence level of vehicle-mounted sensors For the current moment at the th Values ​​of vehicle-mounted sensors, The previous time in the current time period is the time in the previous time period. Values ​​of vehicle-mounted sensors, It is a natural exponential function. It is the absolute value symbol.

[0022] The beneficial effects are as follows: The correction value, based on the ratio of sensor data between the current and previous moments, can capture abrupt changes in sensor data trends. Combined with exponential decay weights, it achieves a smooth transition during the correction process, avoiding positioning jumps caused by sudden data changes. The correction magnitude is adjusted through an exponential function of confidence. When confidence is high, the correction magnitude automatically decreases to avoid over-correction; when confidence is low, the correction magnitude increases, effectively addressing sensor data anomalies and improving the flexibility and accuracy of the correction. The correction direction plays a directional role, ensuring that the correction value is adjusted along a reasonable direction, effectively reducing positioning deviations caused by sensor errors and enhancing the physical rationality and practicality of the correction results. By dynamically balancing historical data with current observations, combined with confidence and correction direction, it achieves intelligent correction of sensor data, significantly reducing the risk of positioning error accumulation and ensuring navigation continuity and accuracy in complex environments.

[0023] Furthermore, the method of using AI algorithms to predict the location data at the current moment includes: the AI ​​algorithm adopts the LSTM algorithm, taking the location data of the previous moment and the correction values ​​of each sensor at the current moment as input to the LSTM algorithm to obtain the predicted value of the location data at the current moment, thus completing the AI-based auxiliary correction of electric vehicle location data.

[0024] Secondly, the present invention provides an AI-based electric vehicle positioning data-assisted correction system, which adopts the following technical solution:

[0025] An AI-based electric vehicle positioning data-assisted correction system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned AI-based electric vehicle positioning data-assisted correction method is implemented.

[0026] By adopting the above technical solution, a computer program is generated from the above-mentioned AI-based electric vehicle positioning data-assisted correction method and stored in a memory so that it can be loaded and executed by a processor. In this way, a terminal device can be made based on the memory and the processor for convenient use.

[0027] The present application has the following technical effects:

[0028] By calculating the positioning data movement coefficient and clustering the driving conditions, the current sensor value is corrected in combination with the sensor confidence, effectively suppressing the trajectory deviation caused by the accumulation of vehicle sensor errors in satellite signal shielding scenes such as urban canyons and tunnels, so that the positioning is closer to the real path; Fusion of multi-dimensional data, using AI algorithm to establish prediction model, dynamically adjusting sensor correction strategy according to different driving scenes, can automatically identify static, turning and other scenes, avoid the problem of insufficient adaptability of traditional methods, and ensure stable and reliable positioning; Real-time calibration of sensor error prevents "jumping points" or route planning failure in navigation, and can also provide accurate navigation guidance during signal shielding; At the same time, it can accumulate data and iteratively optimize, adapt to new environment and reduce maintenance cost. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 is a method flow chart of an AI-based electric vehicle positioning data auxiliary correction method. DETAILED DESCRIPTION

[0030] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0031] The embodiments of the present application disclose an AI-based electric vehicle positioning data auxiliary correction method, referring to Figure 1 , comprising steps S1-S5:

[0032] S1: Obtain the positioning data sequence containing multiple dimensions of each historical moment in the historical journey of the electric vehicle.

[0033] It should be noted that new energy vehicles are equipped with various sensors, such as global navigation satellite system (GNSS) receivers that can receive satellite signals to obtain real-time position information such as longitude, latitude, and altitude of the vehicle; inertial measurement units (IMU) that can measure acceleration, angular velocity, and other data of the vehicle to calculate the motion state and position change of the vehicle; wheel speed sensors that calculate the distance and speed of the vehicle by monitoring the wheel speed; these sensors work together to continuously collect data during vehicle travel, forming a positioning data sequence containing multiple dimensions. The collected positioning data is transmitted to the cloud server or the vehicle manufacturer's data center for storage through the vehicle's communication module, such as the in-vehicle infotainment system (IVI) or the telematics unit (T-Box), using mobile communication networks (such as 4G, 5G) or other wireless communication technologies; when performing this trip, the closest accurate historical trip positioning data to the starting position of this trip is obtained for subsequent analysis.

[0034] Specifically, the dimensions include longitude, latitude, and altitude.

[0035] S2: Determine the movement coefficient of the positioning data sequence in each dimension for each historical time.

[0036] It should be noted that the positioning of the electric vehicle is relatively accurate without satellite signal obstruction, but after the satellite signal is obstructed, the positioning of the electric vehicle will have some deviation without the support of satellite signals. Therefore, in order to measure the deviation of the positioning of the electric vehicle without the support of satellite signals, the movement coefficient is obtained based on the historical positioning data sequence.

[0037] The length of the preset length is the length of the sliding window, and the sliding window slides on the positioning data sequence. For any historical time, when the center of the sliding window is located at the time, the movement coefficient of the positioning data sequence in each dimension for the time is determined based on the standard deviation of the positioning data in each dimension for all historical times within the sliding window, and the positioning data in each dimension for the historical times located at both ends of the sliding window.

[0038] The length of the sliding window can be set by the implementer according to the specific implementation, for example, 11. For the first 10 times and the last 10 times in each historical trip, these times are used as auxiliary reference times and are not calculated.

[0039] Specifically, the movement coefficient satisfies:

[0040] ;

[0041] In the formula, is the positioning data in the i-th dimension for the j-th historical time. is the positioning data in the i-th dimension for the j-th historical time. The movement coefficient of the positioning data sequence in each dimension. The center of the sliding window is located at the first At the 1st historical moment, all historical moments in the sliding window are at the 1st historical moment. The standard deviation of the positioning data in each dimension. For all historical moments in the first The standard deviation of the positioning data in each dimension. The center of the sliding window is located at the first The location data of a historical moment at one end of a sliding window. The center of the sliding window is located at the first The location data of the historical moment at the other end of the sliding window at each historical moment. It is the absolute value symbol.

[0042] in, The larger the value, the more discrete the location data distribution within the sliding window is compared to the overall location data distribution, and the greater the movement range of the car at the corresponding moment in the sliding window. The larger the positioning movement coefficient in each dimension; The smaller the value, the more concentrated the location data distribution within the sliding window is compared to the overall location data distribution, and the smaller the movement range of the car at the corresponding moment in the sliding window. The smaller the positioning movement coefficient in each dimension. The larger the value, the higher the value of the electric vehicle in the corresponding time period of the sliding window. The greater the movement span in each dimension, the more... The larger the positioning movement coefficient in each dimension; The smaller the value, the more likely the electric vehicle is to be in the first position within the corresponding time period of the sliding window. The smaller the movement span in each dimension, the better. The smaller the positioning movement coefficient in each dimension.

[0043] S3: Obtain the categories of multiple electric vehicle driving conditions and determine the confidence level of each category with each on-board sensor.

[0044] It should be noted that various driving conditions occur during vehicle operation, such as going straight, turning left, and turning right. The vehicle's movement differs under these conditions. To ensure that the electric vehicle's positioning data correction can adapt to these different driving conditions, this step clusters historical time points based on the positioning movement coefficient. Then, based on the onboard sensor data corresponding to each category's historical time point, the confidence level between each category and its corresponding onboard sensor data is obtained.

[0045] Based on the mobility coefficient, clustering is performed on each historical moment to obtain multiple categories of electric vehicle driving conditions. The confidence level of each category and each on-board sensor is determined according to the standard deviation of the values ​​of each on-board sensor of the electric vehicle at all historical moments in each category, and the standard deviation of the mobility coefficient of the positioning data sequence in each dimension at all historical moments in each category.

[0046] Specifically, the clustering uses DBSCAN clustering.

[0047] In another embodiment, the clustering employs mean-shift clustering.

[0048] Specifically, the confidence level satisfies:

[0049] ;

[0050] In the formula, For the first The category and the first Confidence level of vehicle-mounted sensors For the first All historical moments within each category in the [number]th Standard deviation of values ​​for vehicle-mounted sensor data The number of types of vehicle-mounted sensors. For the first All historical moments within each category are in the [number]th [section]. The standard deviation of the movement coefficient of the location data sequence in each dimension. For the number of dimensions, For the number of categories, It is a function with maximum value. It is a natural exponential function.

[0051] in, Indicates the first The first corresponding sensor data at all historical moments within each category The relative magnitude of the dispersion of data from vehicle-mounted sensors; the larger the value, the more dispersed the data is in the first... Under the driving status corresponding to historical moments within each category, the first The greater the fluctuation of data from vehicle-mounted sensors compared to other sensors, the higher their instability and noise levels. Therefore, the first... The category and the first The lower the confidence level of vehicle-mounted sensors (vehicle-mounted sensor data), the better to avoid overcompensation or correction of positioning data deviations caused by their fluctuations; the smaller the value, the lower the confidence level. Vehicle-mounted sensor data in the first The more stable the driving state of the category is, the lower the noise level is, the better the reading consistency is, and the higher the confidence of the vehicle-mounted sensor in this state is, and the more the output of the vehicle-mounted sensor can be relied on for positioning correction. represents the maximum value of the relative proportion of the dispersion degree of the moving coefficient in all dimensions of the first category, and the greater the value is, the more complex the driving conditions corresponding to the historical moments in the first category are (for example, when straight driving, sometimes accelerating first and then decelerating, sometimes decelerating first and then accelerating, and sometimes constant speed), and the confidence of the sensor data in this condition should be lower to avoid blindly trusting the vehicle-mounted sensor data and causing deviation of the positioning data; the smaller the value is, the more single the driving conditions corresponding to the historical moments in the first category are (for example, when turning, deceleration is required), and the confidence of the sensor data in this condition should be higher.

[0052] S4: determining the correction value of the current moment in the current trip for each vehicle-mounted sensor.

[0053] It should be noted that due to the bumps in the driving process of the electric vehicle or the interference of the circuit in the electric vehicle, there are inevitably some inaccurate data or noise data in the data collected by the vehicle-mounted sensor, and the acquisition of the positioning data of the electric vehicle through these data will cause deviation of the positioning data of the electric vehicle. Therefore, in order to weaken the deviation of the positioning data of the electric vehicle, the positioning of the electric vehicle is assisted and corrected according to the confidence of each category and each vehicle-mounted sensor and the data of each vehicle-mounted sensor.

[0054] Based on the values of the current moment in the current trip and all historical moments in each category for each vehicle-mounted sensor, a reference category of the current moment and a correction direction of the current moment are obtained.

[0055] Specifically, the reference category of the current moment and the correction direction of the current moment are obtained by: constructing a sequence of the current moment based on the values of the current moment in the current trip for each vehicle-mounted sensor, constructing a sequence of each category based on the mean values of the values of all historical moments in each category for each sensor, and recording the category with the maximum cosine similarity between the sequence of the current moment and the sequence of each category as the reference category of the current moment.

[0056] The correction direction of the current moment for each vehicle-mounted sensor is determined according to the size of the values of the current moment and the reference category for each vehicle-mounted sensor.

[0057] Specifically, the correction direction satisfies:

[0058] ;

[0059] In the formula, V is the value of the current moment in the first category for each vehicle-mounted sensor, and V is the value of the reference category for each vehicle-mounted sensor. the correction direction of the vehicle-mounted sensor of the category, the value of the vehicle-mounted sensor of the category at the current time, the value of the vehicle-mounted sensor of the category at the current time, the average value of the values of the vehicle-mounted sensors of the category at all historical times within the reference category, the average value of the values of the vehicle-mounted sensors of the category at all historical times within the reference category.

[0060] wherein, the value of the vehicle-mounted sensor of the category at the current time is higher than the average level of the reference category, the more likely to be overestimated, the more should be corrected in the negative direction to be closer to the historical normal level; the value of the vehicle-mounted sensor of the category at the current time is lower than the average level of the reference category, the more likely to be underestimated, the more should be corrected in the positive direction to be closer to the historical normal level. In this example, is 1 or -1, and the implementer can also determine the value of according to the actual situation, for example, 0.9 or -0.9, etc.

[0061] According to the confidence of the reference category and each vehicle-mounted sensor, and the correction direction, the value of each vehicle-mounted sensor at the current time is corrected to obtain the corrected value of each vehicle-mounted sensor at the current time.

[0062] Specifically, the corrected value satisfies:

[0063] ;

[0064] wherein, the corrected value of the vehicle-mounted sensor of the category at the current time, the corrected value of the vehicle-mounted sensor of the category at the current time, the correction direction of the vehicle-mounted sensor of the category at the current time, the confidence of the reference category and the vehicle-mounted sensor of the category at the current time, the value of the vehicle-mounted sensor of the category at the current time, the value of the vehicle-mounted sensor of the category at the previous time at the current time, the value of the vehicle-mounted sensor of the category at the previous time at the current time, the value of the vehicle-mounted sensor of the category at the current time, the value of the vehicle-mounted sensor of the category at the previous time at the current time, the value of the vehicle-mounted sensor of the category at the previous time at the current time, the natural exponential function, the absolute value symbol.

[0065] wherein, represents the difference between the values of the vehicle-mounted sensor of the category at adjacent times, since the driving of the electric vehicle has continuity, the sensor data of adjacent times is usually similar, so the larger the difference is, the more the sensor data at the current time should be corrected, The smaller the value, the less need there is to correct the sensor data at the current moment. The larger the value, the more likely it is to be the first. The more reliable the data from vehicle-mounted sensors, the less need there is for... Make corrections; The smaller the value, the better. The more unreliable the data from vehicle-mounted sensors, the more necessary it is to... Make corrections.

[0066] S5: Based on the correction value, use AI algorithms to predict the location data at the current moment.

[0067] Specifically, the use of AI algorithms to predict the location data at the current moment includes:

[0068] The AI ​​algorithm uses the LSTM algorithm, taking the positioning data from the previous moment and the correction values ​​from each sensor at the current moment as input to the LSTM algorithm to obtain the predicted value of the positioning data at the current moment, thus completing the AI-based auxiliary correction of electric vehicle positioning data.

[0069] This invention also discloses an AI-based electric vehicle positioning data auxiliary correction system, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an AI-based electric vehicle positioning data auxiliary correction method according to the present invention.

[0070] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0071] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An AI-based method for assisting in the correction of electric vehicle positioning data, characterized in that, include: Obtain multi-dimensional location data sequences for each historical moment in the electric vehicle's historical journey; The sliding window is made to slide on the positioning data sequence with a preset length. For any historical moment, when the center of the sliding window is located at that moment, the movement coefficient of the positioning data sequence in each dimension at that moment is determined based on the standard deviation of the positioning data of all historical moments in each dimension within the sliding window, and the positioning data of the historical moments at both ends of the sliding window in each dimension. Based on the mobility coefficient, clustering is performed on each historical moment to obtain multiple categories of electric vehicle driving conditions. The confidence level of each category and each on-board sensor is determined according to the standard deviation of the values ​​of each on-board sensor of the electric vehicle at all historical moments in each category, and the standard deviation of the mobility coefficient of the positioning data sequence in each dimension at all historical moments in each category. Based on the current time in this trip and the values ​​of all historical times in each category at each on-board sensor, obtain the reference category for the current time and the correction direction for the current time; Based on the reference category, the confidence level of each vehicle sensor, and the correction direction, the values ​​of each vehicle sensor at the current time are corrected to obtain the corrected values ​​of each vehicle sensor at the current time. Based on the correction value, an AI algorithm is used to predict the location data at the current moment. This includes: the AI ​​algorithm uses the LSTM algorithm, taking the location data from the previous moment and the correction values ​​of each sensor at the current moment as input to the LSTM algorithm to obtain the predicted value of the location data at the current moment, thus completing the AI-based auxiliary correction of electric vehicle location data.

2. The AI-based electric vehicle positioning data-assisted correction method according to claim 1, characterized in that, The dimensions include longitude, latitude, and altitude.

3. The AI-based electric vehicle positioning data-assisted correction method according to claim 1, characterized in that, The mobility coefficient satisfies: ； In the formula, For the first The historical moment at the The movement coefficient of the positioning data sequence in each dimension. The center of the sliding window is located at the first At the 1st historical moment, all historical moments in the sliding window are at the 1st historical moment. The standard deviation of the positioning data in each dimension. For all historical moments in the first The standard deviation of the positioning data in each dimension. The center of the sliding window is located at the first The location data of a historical moment at one end of a sliding window. The center of the sliding window is located at the first The location data of the historical moment at the other end of the sliding window at each historical moment. It is the absolute value symbol.

4. The AI-based electric vehicle positioning data-assisted correction method according to claim 1, characterized in that, The clustering was performed using DBSCAN clustering.

5. The AI-based electric vehicle positioning data-assisted correction method according to claim 1, characterized in that, The confidence level satisfies: ； In the formula, For the first The category and the first Confidence level of vehicle-mounted sensors For the first All historical moments within each category in the [number]th Standard deviation of values ​​for vehicle-mounted sensor data The number of types of vehicle-mounted sensors. For the first All historical moments within each category are in the [number]th [section]. The standard deviation of the movement coefficient of the location data sequence in each dimension. For the number of dimensions, For the number of categories, It is a function with maximum value. It is a natural exponential function.

6. The AI-based electric vehicle positioning data-assisted correction method according to claim 1, characterized in that, The process of obtaining the reference category and the correction direction at the current moment includes: Based on the values ​​of each on-board sensor at the current moment during this trip, a sequence of the current moment is constructed. Based on the mean of the values ​​of each sensor at all historical moments within each category, a sequence of each category is constructed. The category with the highest cosine similarity between the sequence of the current moment and the sequence of each category is denoted as the reference category of the current moment. Based on the values ​​of the current time and reference category at each vehicle sensor, the correction direction at each vehicle sensor at the current time is determined.

7. The AI-based electric vehicle positioning data-assisted correction method according to claim 6, characterized in that, The correction direction satisfies: ； In the formula, For the current moment at the th Correction direction for vehicle-mounted sensors, For the current moment at the th Values ​​of vehicle-mounted sensors, For reference, all historical moments within the category are in the [number]th [year]. The average value of the vehicle-mounted sensor.

8. The AI-based electric vehicle positioning data-assisted correction method according to claim 7, characterized in that, The correction value satisfies: ； In the formula, For the current moment at the th Correction values ​​for vehicle-mounted sensors, The reference category at the current moment is in the 1st... Correction direction for vehicle-mounted sensors, The reference category for the current moment and the first Confidence level of vehicle-mounted sensors For the current moment at the th Values ​​of vehicle-mounted sensors, The previous time in the current time period is the time in the previous time period. Values ​​of vehicle-mounted sensors, It is a natural exponential function. It is the absolute value symbol.

9. An AI-based electric vehicle positioning data-assisted correction system, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement an AI-based electric vehicle positioning data-assisted correction method according to any one of claims 1-8.

Images