A method and system for accurate metering and analysis of energy consumption in commercial vehicles based on multi-source data fusion.

Abstract

This invention provides a method and system for accurate measurement and analysis of commercial vehicle energy consumption based on multi-source data fusion. The system includes: a vehicle-side multi-dimensional sensing module for collecting data on electrical parameters, load, high-precision positioning and slope, environmental conditions, and battery temperature; a data preprocessing and fusion module for initial screening based on commercial vehicle operating condition characteristic thresholds and the 3σ principle, and innovatively integrating load and slope data to identify the type of electrical parameter anomalies; an accurate energy consumption calculation module that uses a dynamic interval partitioning algorithm to calculate net energy consumption per kilometer using a differential method to avoid cumulative errors; and an adaptive transmission and cloud processing module that adjusts data transmission strategies according to network conditions and uses a distributed architecture to process massive amounts of data. This invention achieves contextualized, high-precision analysis, and deep attribution of commercial vehicle energy consumption.

Description

Technical Field

[0001] This invention relates to the field of automotive remote monitoring and big data analysis technology, and in particular to a method and system for accurate metering and analysis of commercial vehicle energy consumption based on multi-source data fusion. Background Technology

[0002] In the commercial vehicle and electric vehicle sectors, accurate energy consumption metering and analysis are crucial for improving energy efficiency and optimizing vehicle operation management. Currently, the industry commonly employs energy consumption monitoring methods based on vehicle operation data. These methods typically rely on basic electrical parameters such as current, voltage, and state of charge (SOC) for energy consumption estimation. However, in practical applications, existing methods still face several technical challenges:

[0003] First, existing energy consumption analysis solutions often focus on monitoring a single data dimension, making it difficult to comprehensively reflect the complex operating conditions of vehicles in actual operation, such as the impact of factors like load changes and road gradients on energy consumption. This results in energy consumption analysis results that often lack depth and cannot effectively explain the specific reasons for energy consumption fluctuations.

[0004] Secondly, due to objective factors such as sensor noise and signal interference, the raw data collected often contains outliers. If only general statistical methods are used for data cleaning, some reasonable high-power-consumption operating conditions (such as heavy-load uphill driving) may be misjudged as outliers, thus affecting the accuracy and reliability of the analysis.

[0005] In addition, traditional energy consumption calculation models often use the SOC integration method from the start to the end of the journey. This method is prone to cumulative errors due to sensor zero drift, battery management system (BMS) estimation deviations, etc. during long-term operation, and may even cause data overflow problems, resulting in reduced reliability of the calculation results.

[0006] On the other hand, existing data transmission mechanisms are usually based on fixed frequencies, which may lead to the loss of critical data in areas with poor network conditions, affecting the integrity and continuity of energy consumption analysis.

[0007] Although existing technical solutions have attempted to improve data coverage by increasing the number or type of sensors, most solutions remain at the level of simple data overlay, failing to achieve deep fusion and contextualized discrimination of multi-source data; at the same time, in terms of computational models, they have not fundamentally solved core problems in engineering practice such as cumulative errors. Summary of the Invention

[0008] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method and system for accurate measurement and analysis of commercial vehicle energy consumption based on multi-source data fusion.

[0009] To achieve the above objectives, the present invention adopts the following technical solution:

[0010] A method for accurate measurement and analysis of energy consumption in commercial vehicles based on multi-source data fusion includes the following steps:

[0011] S1: Real-time synchronous collection of vehicle electrical parameter data, load data, high-precision positioning and slope data, ambient temperature and battery temperature data through vehicle-side sensor clusters to construct a multi-dimensional data matrix of "electricity-operating condition-temperature";

[0012] S2: Preprocessing and fusion-based anomaly detection of the multi-dimensional data matrix, including the following sub-steps:

[0013] S21: Based on the preset threshold values ​​for commercial vehicle operating conditions, perform preliminary screening on various types of data and remove outliers that clearly exceed physical limits.

[0014] S22: Using a sliding window combined with the 3σ principle, the electrical parameter data {x1, x2, ..., xn} is smoothed to remove random noise;

[0015] S23: Integrate load data W and slope data θ to construct an anomaly type discrimination matrix based on the input combination of (F_I, F_V, W, θ), and output the judgment results including normal heavy-load uphill working conditions, sensor noise, vehicle mechanical faults and battery faults.

[0016] Where F_I represents the current abnormality indicator and F_V represents the voltage abnormality indicator;

[0017] S3: Based on high-precision GPS trajectory points, the Vincenti formula is used to continuously calculate the cumulative driving distance, and a calculation interval is dynamically defined every 1 kilometer of cumulative distance.

[0018] S4: Within each calculation interval, the differential calculation method is used to calculate the net power consumption E_net = |SOC_end - SOC_start| × C - Q_regen based on the difference between the start and end SOC_end of the interval and the recharge amount Q_regen.

[0019] S5: Associate the net power consumption with the average load, average slope, and temperature data within the corresponding range, store the data, and upload it to the cloud platform;

[0020] S6: The cloud platform generates detailed energy consumption analysis reports for vehicles, drivers, or routes based on data from all computing intervals.

[0021] Furthermore, the determination logic for the anomaly type discrimination matrix in step S23 is specifically as follows:

[0022] When (F_I=1∨F_V=1)∧(W>W)th )∧(θ>θ th When the condition is as described, it is determined to be a "normal heavy-load uphill condition";

[0023] When (F_I=1∨F_V=1)∧(W≤W) th )∧(θ≤θ th )∧(Δt <T max When this occurs, it is determined to be "sensor noise";

[0024] When (F_I=1∨F_V=1)∧(W≤W) th )∧(θ≤θ th )∧(Δt≥T max When this occurs, it is determined to be a "vehicle mechanical failure";

[0025] When (F_I=1∨F_V=1)∧(W≤W) th )∧(θ≤θ th )∧(ΔSOC>SOC max When this occurs, it is determined to be a "battery malfunction";

[0026] Among them, W th θ is the preset load threshold. th To preset the slope threshold, T max To preset the maximum transient time, SOC max This is the preset maximum SOC change threshold.

[0027] Furthermore, the specific formula for calculating the distance d between adjacent GPS points P1(lat1, lon1) and P2(lat2, lon2) using the Vincenti formula in step S3 is as follows:

[0028] Δσ = arctan( √( [cosφ2·sinΔλ]² + [cosφ1·sinφ2- sinφ1·cosφ2·cosΔλ]² ) / (sinφ1·sinφ2 + cosφ1·cosφ2·cosΔλ) )

[0029] d = R · Δσ

[0030] Where φ is latitude, λ is longitude, Δλ is the difference in longitude, and R is the Earth's radius.

[0031] Furthermore, during data transmission, an adaptive compression transmission strategy is implemented, including:

[0032] Real-time monitoring of network quality metrics Q (0≤Q≤1) between T-BOX and the cloud.

[0033] The transmission frequency f = f_max·Q and the compression ratio r = r_min + (r_max - r_min)·(1 - Q) are dynamically adjusted according to the Q value.

[0034] When Q>Q th Upload complete data in a timely manner. When Q ≤ Qth, prioritize uploading key statistical summary data and abnormal events.

[0035] A commercial vehicle energy consumption precision metering and analysis system based on multi-source data fusion includes:

[0036] The vehicle-mounted sensing and transmission unit includes:

[0037] CAN bus interface used for acquiring electrical parameters according to standard GB / T 32960;

[0038] A load sensor module for real-time vehicle mass measurement is a pressure sensor mounted on the vehicle's leaf springs or air suspension, and incorporates a calibration curve M = {k} based on piecewise linear fitting. ᵢ ·P + bᵢ} is used to convert sensor signal P into quality data M in real time;

[0039] A high-precision GPS and IMU fusion module is used to acquire position, velocity and calculate road slope. It fuses GPS elevation data with IMU triaxial accelerometer and gyroscope data through Kalman filtering algorithm to output a smooth road slope value.

[0040] A temperature sensor array used to monitor ambient and battery temperatures;

[0041] The intelligent T-BOX, which integrates adaptive compression transmission logic, is used to aggregate, process, and upload all sensor data.

[0042] The cloud-based data processing and analysis unit adopts a distributed architecture and includes:

[0043] The stream processing engine is used to perform data cleaning, fusion discrimination, and dynamic range calculation in real time.

[0044] Time-series databases and relational databases are used to store detailed data and calculation results;

[0045] The energy consumption report generation module is used to generate analysis reports based on the net power consumption within a range and its associated multidimensional data.

[0046] The client-side data display unit is used to provide users with a visual energy consumption traceability dashboard, driving behavior scores, and personalized comparison reports.

[0047] Furthermore, the high-precision GPS and IMU fusion module adopts the Kalman filter algorithm, and its system state vector is defined as X = [h, v, a, θ]ᵀ, where h is the elevation, v is the vertical velocity, a is the vertical acceleration, and θ is the slope angle;

[0048] The state equation is: X_k = A·X_{k-1} + W_k;

[0049] The observation equation is: Z_k = H·X_k + V_k;

[0050] Where A is the state transition matrix, H is the observation matrix, and W_k and V_k are the process noise and observation noise, respectively.

[0051] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for accurate metering and analysis of energy consumption in commercial vehicles based on multi-source data fusion.

[0052] Compared with the prior art, the beneficial effects of the present invention are as follows: the present application realizes the contextualization of energy consumption analysis through the "electricity-operating condition-temperature" multi-dimensional data matrix and anomaly discrimination matrix, and accurately distinguishes between normal high energy consumption and real anomalies;

[0053] This application adopts dynamic interval division and differential calculation method, which effectively avoids cumulative error and improves the reliability and accuracy of energy consumption measurement;

[0054] This application uses a secondary discrimination based on load and slope data, combined with a mathematical decision-making model, to reduce misjudgments and ensure data quality.

[0055] This application intelligently adjusts the data transmission strategy based on network conditions to ensure data integrity and real-time performance;

[0056] This application supports energy consumption monitoring and optimization for large-scale commercial vehicles through collaborative work between the vehicle, cloud, and client ends. Attached Figure Description

[0057] Figure 1 This is a flowchart illustrating the steps of the method for accurate metering and analysis of commercial vehicle energy consumption based on multi-source data fusion in Embodiment 1 of the present invention. Detailed Implementation

[0058] To provide a further understanding of the purpose, structure, features, and functions of the present invention, detailed descriptions are provided below with reference to specific embodiments.

[0059] The system deployment of this invention is divided into three parts: vehicle-side, cloud-side, and client-side.

[0060] A method for accurate measurement and analysis of energy consumption in commercial vehicles based on multi-source data fusion includes the following steps:

[0061] S1: Real-time synchronous collection of vehicle electrical parameter data, load data, high-precision positioning and slope data, ambient temperature and battery temperature data through vehicle-side sensor clusters to construct a multi-dimensional data matrix of "electricity-operating condition-temperature";

[0062] Electrical parameter data according to GB / T 32960 standard are acquired at a frequency of 1Hz via a CAN bus interface. A load cell collects pressure signals in real time, which are converted into mass data using a pre-stored calibration curve. A GPS / IMU fusion module outputs high-precision slope data through Kalman filtering. A temperature sensor group synchronously collects ambient and battery temperatures. Finally, the intelligent T-BOX aligns all the above data in time, constructing a multi-dimensional data matrix of "electricity-operating condition-temperature". This method overcomes the limitations of traditional single electrical parameter monitoring, providing comprehensive data support for energy consumption analysis. Timestamp synchronization ensures that data from different sources are correlated under the same operating conditions, laying a data foundation for subsequent fusion analysis and contextualized judgment.

[0063] S2: Preprocessing and fusion-based anomaly detection of the multi-dimensional data matrix, including the following sub-steps:

[0064] S21: Based on preset threshold values ​​for commercial vehicle operating conditions, perform initial screening of various types of data to remove outliers that clearly exceed physical limits; as the first data quality checkpoint, set hard thresholds based on the physical limits of commercial vehicles to quickly remove obviously abnormal data, reduce subsequent processing load, and improve system efficiency.

[0065] S22: A sliding window combined with the 3σ principle is used to smooth the electrical parameter data {x1, x2, ..., xn} and remove random noise. A sliding window with a window size of 10 is used to dynamically smooth the electrical parameter data, which can effectively filter out random noise and instantaneous interference and keep the data trend characteristics unaffected.

[0066] S23: Integrate load data W and slope data θ to construct an anomaly type discrimination matrix based on the input combination of (F_I, F_V, W, θ), and output the judgment results including normal heavy-load uphill working conditions, sensor noise, vehicle mechanical faults and battery faults.

[0067] Here, F_I represents the current abnormality indicator, and F_V represents the voltage abnormality indicator; this step avoids misjudging reasonable high-energy-consumption operating conditions as abnormal, realizes early warning of mechanical and battery faults, and upgrades from simple elimination to intelligent diagnosis.

[0068] S3: Based on high-precision GPS track points, the cumulative driving distance is continuously calculated using the Vincenti formula, and a calculation interval is dynamically defined every 1 kilometer accumulated. This method provides a unified comparison benchmark for different vehicles and routes, while overcoming the problem of speed changes affecting fixed time intervals.

[0069] S4: Within each calculation interval, the differential calculation method is used to calculate the net energy consumption E_net = |SOC_end - SOC_start| × C - Q_regen based on the difference between the start and end SOC_end of the interval and the recharge amount Q_regen. Compared with the traditional integral method, this method reduces the long-term metering error from >5% to <1%, and accurately reflects the net energy consumption by deducting the recharge amount.

[0070] S5: Associate the net power consumption with the average load, average slope, and temperature data within the corresponding range, store the data, and upload it to the cloud platform;

[0071] S6: The cloud platform generates detailed energy consumption analysis reports for vehicles, drivers, or routes based on data from all computing intervals. This step accurately identifies the root causes of high energy consumption.

[0072] Furthermore, the determination logic for the anomaly type discrimination matrix in step S23 is specifically as follows:

[0073] When (F_I=1∨F_V=1)∧(W>W) th )∧(θ>θ th When the condition is as described, it is determined to be a "normal heavy-load uphill condition";

[0074] When (F_I=1∨F_V=1)∧(W≤W) th )∧(θ≤θ th )∧(Δt <T max When this occurs, it is determined to be "sensor noise";

[0075] When (F_I=1∨F_V=1)∧(W≤W) th )∧(θ≤θ th )∧(Δt≥T max When this occurs, it is determined to be a "vehicle mechanical failure";

[0076] When (F_I=1∨F_V=1)∧(W≤W) th )∧(θ≤θ th )∧(ΔSOC>SOC max When this occurs, it is determined to be a "battery malfunction";

[0077] Among them, W th θ is the preset load threshold. th To preset the slope threshold, T max To preset the maximum transient time, SOC maxA preset maximum SOC change threshold is set. Traditional methods (such as the simple 3σ rule) typically discard abnormal data directly. This leads to the discarding of reasonable high-energy-consuming operating conditions such as heavy-load uphill driving, distorting energy consumption analysis. The discrimination matrix of this invention first introduces the discrimination result of "normal heavy-load uphill driving condition," thus achieving the retention of reasonable abnormal data.

[0078] Multi-source data fusion decision-making logically combines electrical parameter anomaly indicators (F_I, F_V), vehicle operating condition data (W, θ), and time series features (Δt, ΔSOC) to form a multi-dimensional decision-making system.

[0079] Furthermore, the specific formula for calculating the distance d between adjacent GPS points P1(lat1, lon1) and P2(lat2, lon2) using the Vincenti formula in step S3 is as follows:

[0080] Δσ = arctan( √( [cosφ2·sinΔλ]² + [cosφ1·sinφ2- sinφ1·cosφ2·cosΔλ]² ) / (sinφ1·sinφ2 + cosφ1·cosφ2·cosΔλ) )

[0081] d = R · Δσ

[0082] Where φ is latitude, λ is longitude, Δλ is the difference in longitude, and R is the Earth's radius. This provides a fair benchmark for comparing energy consumption data from different vehicles, routes, and times, making subsequent in-depth analyses (such as route optimization and driver behavior comparison) more scientific and reliable.

[0083] Furthermore, during data transmission, an adaptive compression transmission strategy is implemented, including:

[0084] Real-time monitoring of network quality metrics Q (0≤Q≤1) between T-BOX and the cloud.

[0085] The transmission frequency f = f_max·Q and the compression ratio r = r_min + (r_max - r_min)·(1 - Q) are dynamically adjusted according to the Q value. When the network is good, high-frequency transmission is used to ensure real-time performance; when the network is poor, low-frequency transmission is used to avoid congestion and packet loss. When the network is good, low compression or no compression is used to ensure data integrity; when the network is poor, high compression is used, sacrificing some details to preserve core information.

[0086] When Q>Q th Upload complete data when Q ≤ Q th Prioritize uploading key statistical summary data and abnormal events. When network quality is poor (Q ≤ Q...),... thPrioritizing the transmission of high-value information such as abnormal events and key statistics, this system implements an intelligent "sacrifice pawns to save the king" strategy, ensuring that core business functions are not interrupted.

[0087] A commercial vehicle energy consumption precision metering and analysis system based on multi-source data fusion includes:

[0088] The vehicle-mounted sensing and transmission unit includes:

[0089] CAN bus interface used for acquiring electrical parameters according to standard GB / T 32960;

[0090] A load sensor module for real-time vehicle mass measurement is a pressure sensor mounted on the vehicle's leaf springs or air suspension, and incorporates a calibration curve M = {k} based on piecewise linear fitting. ᵢ ·P + bᵢ} is used to convert the sensor signal P into mass data M in real time; the present invention adopts piecewise linear fitting of M = {kᵢ·P + bᵢ}, and approximates the curve with multiple straight lines, which specifically solves the measurement error caused by suspension nonlinearity and realizes high-precision perception of load data, which is the premise of "precise measurement".

[0091] A high-precision GPS and IMU fusion module is used to acquire position, velocity and calculate road slope. It fuses GPS elevation data with IMU triaxial accelerometer and gyroscope data through Kalman filtering algorithm to output a smooth road slope value.

[0092] A temperature sensor array used to monitor ambient and battery temperatures;

[0093] The intelligent T-BOX, which integrates adaptive compression transmission logic, is used to aggregate, process, and upload all sensor data.

[0094] The cloud-based data processing and analysis unit adopts a distributed architecture and includes:

[0095] The stream processing engine is used to perform data cleaning, fusion discrimination, and dynamic range calculation in real time.

[0096] Time-series databases and relational databases are used to store detailed data and calculation results;

[0097] The energy consumption report generation module is used to generate analysis reports based on the net power consumption within a range and its associated multidimensional data.

[0098] The client-side data display unit is used to provide users with a visual energy consumption traceability dashboard, driving behavior scores, and personalized comparison reports.

[0099] Furthermore, the high-precision GPS and IMU fusion module adopts the Kalman filter algorithm, and its system state vector is defined as X = [h, v, a, θ]ᵀ, where h is the elevation, v is the vertical velocity, a is the vertical acceleration, and θ is the slope angle. The slope angle θ is estimated as a direct state, rather than through conventional indirect calculation, which improves the estimation accuracy and smoothness.

[0100] The state equation is: X_k = A·X_{k-1} + W_k;

[0101] The observation equation is: Z_k = H·X_k + V_k;

[0102] Where A is the state transition matrix, H is the observation matrix, and W_k and V_k are the process noise and observation noise, respectively.

[0103] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for accurate metering and analysis of energy consumption in commercial vehicles based on multi-source data fusion.

[0104] The present invention has been described in the above-described embodiments; however, these embodiments are merely examples for implementing the present invention. It must be noted that the disclosed embodiments do not limit the scope of the present invention. Conversely, any modifications and refinements made without departing from the spirit and scope of the present invention are within the scope of patent protection of the present invention.

Claims

1. A commercial vehicle energy consumption precise metering and analysis method based on multi-source data fusion, characterized in that, Comprise the following steps: S1: Real-time synchronous acquisition of electric parameter data, load data, high-precision positioning and slope data, environmental temperature and battery temperature data of the vehicle through the vehicle-end sensor cluster, to construct a "electricity-working condition-temperature" multidimensional data matrix; S2: Preprocessing and fusion type abnormality discrimination of the multidimensional data matrix, comprising the following sub-steps: S21: Based on the preset commercial vehicle working condition characteristic threshold, screening each type of data, and eliminating abnormal values obviously exceeding the physical limit; S22: The electric parameter data {x1, x2, …, x n} are smoothed by using sliding window combined with 3σ principle to eliminate random noise; S23: fuse the load data W and the slope data θ, construct an abnormal type discrimination matrix based on the input combination (F_I, F_V, W, θ), the discrimination matrix is combined with the preset load threshold W th , the slope threshold θ th , the maximum transient time T max and the maximum SOC change threshold SOC max , and the electric parameter abnormal flag F_I and F_V, and distinguish the following four types through logical combination judgment rule: normal heavy load uphill working condition, sensor noise, vehicle mechanical fault and battery fault; Wherein, F_I represents the current abnormal flag, and F_V represents the voltage abnormal flag; The determination logic of the abnormal type discrimination matrix is specifically: When (F_I = 1 V F_V = 1) A (W > W th ) A (θ > θ th ) is satisfied, it is determined as "normal heavy load uphill working condition"; When (F_I = 1 V F_V = 1) A (W < W th ) A (θ < θ th ) A (Δt < T max ) is satisfied, it is determined as "sensor noise"; When (F_I = 1 V F_V = 1) A (W < W th ) A (θ < θ th ) A (Δt > T max ) is satisfied, it is determined as "vehicle mechanical failure"; When (F_I = 1 V F_V = 1) A (W < W th ) A (θ < θ th ) A (ΔSOC > SOC max ) is satisfied, it is determined as "battery failure"; wherein W th is a preset load threshold, θ th is a preset slope threshold, T max is a preset maximum transient time, SOC max is a preset maximum SOC change threshold S3: Based on the high-precision GPS track point, the Vincenty formula is used to continuously calculate the cumulative driving distance, and a calculation interval is dynamically defined every 1 kilometer; S4: In each calculation interval, the difference calculation method is used to calculate the net power consumption E_net=|SOC_end-SOC_start|×C-Q_regen according to the difference between the interval start SOC_start and end SOC_end and the regenerative power Q_regen; S5: The net power consumption is associated with the average load, average slope and temperature data in the corresponding interval and stored, and uploaded to the cloud platform; S6: The cloud platform generates a refined energy consumption analysis report for the vehicle, driver or line based on the data of all calculation intervals.

2. The commercial vehicle energy consumption precise metering and analyzing method based on multi-source data fusion according to claim 1, characterized in that: The specific formula for calculating the distance d between adjacent GPS points P1(lat1,lon1) and P2(lat2,lon2) in step S3 is: ; d = R · ; wherein, is the latitude, is the longitude, is the longitude difference, and R is the earth radius.

3. The commercial vehicle energy consumption precise metering and analyzing method based on multi-source data fusion according to claim 1, characterized in that: In the data transmission process, an adaptive compression transmission strategy is executed, including: Real-time monitoring of network quality indicators Q between T-BOX and the cloud; According to the Q value, dynamically adjusting the transmission frequency f=f_max·Q and the compression ratio r=r_min+(r_max-r_min)·(1-Q); Upload full data when Q>Q th Upload key statistical summary data and abnormal events first when Q≤Q th Upload full data when Q>Q th Upload key statistical summary data and abnormal events first when Q≤Q th Upload full data when Q>Q <000 4. A method for implementing the multi-source data fusion-based commercial vehicle energy consumption precise metering and analysis method system according to any one of claims 1-3, characterized in that: Comprise: Vehicle-end perception and transmission unit, comprising: CAN bus interface for collecting standard GB / T32960 electric parameters; A load sensor module for real-time measurement of vehicle mass, which is a pressure sensor mounted on a vehicle leaf spring or air suspension, and has a built-in calibration curve M={k i ·P+b i} for real-time conversion of the sensor signal P into mass data M; High-precision GPS and IMU fusion module for obtaining position, speed and calculating road slope, which fuses GPS elevation data and IMU three-axis accelerometer and gyroscope data through Kalman filtering algorithm to output smooth road slope value; Temperature sensor group for monitoring environmental and battery temperature; Intelligent T-BOX integrated with adaptive compression transmission logic, used for aggregating, processing and uploading all sensor data; Cloud data processing and analysis unit, adopting distributed architecture, comprising: Stream processing engine for real-time execution of data cleaning, fusion discrimination and dynamic interval calculation; Time series database and relational database for storing detailed data and calculation results; Energy consumption report generation module for generating analysis report based on interval net power consumption and its associated multidimensional data; Client data display unit for providing users with visual energy consumption traceability dashboard, driving behavior score and personalized comparison report.

5. The commercial vehicle energy consumption precise metering and analyzing system based on multi-source data fusion of claim 4, wherein: The high-precision GPS and IMU fusion module adopts Kalman filtering algorithm, and a system state vector is defined as X=[h, v, a, θ] T wherein h is an elevation, v is a vertical velocity, a is a vertical acceleration, and θ is a slope angle. State equation: X_k=A·X_{k-1}+W_k; Observation equation: Z_k=H·X_k+V_k; where A is a state transition matrix, H is an observation matrix, W_k and V_k are process noise and observation noise.

6. A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method of any one of claims 1-3.

Images