IMU calibration system and method based on intelligent driving navigation

Abstract

The invention belongs to the technical field of intelligent driving navigation, and particularly relates to an IMU calibration system and method based on intelligent driving navigation. According to the method, multi-source sensor data are collected and fused in real time, Kalman filtering and feedback correction mechanisms are utilized, online joint estimation of IMU error parameters is achieved in the dynamic driving process of a vehicle, the dependence of a traditional calibration method on a static environment and offline processing is broken through, and by constructing an extended state Kalman filter, the real-time calibration of the IMU error parameters is achieved. According to the method, the zero offset of the accelerometer, the scale factor error of the gyroscope and the installation angle deviation are brought into a unified estimation framework, and dynamic separation and real-time compensation of error parameters are realized by combining RTK high-precision positioning data and wheel speed information, so that the navigation precision and long-term stability of the vehicle-mounted IMU under a complex driving working condition are remarkably improved, and the vehicle-mounted IMU has a wide application prospect. The accumulated drift of the inertial error is effectively inhibited, and the reliable positioning capability of the intelligent driving system in a transient shielding or degradation scene of the GNSS signal is ensured.

Description

Technical Field

[0001] This invention belongs to the field of intelligent driving navigation technology, specifically relating to an IMU calibration system and method based on intelligent driving navigation. Background Technology

[0002] With the increasing demand for intelligent driving navigation in passenger vehicles, the IMU (Integrated Measurement Unit) is one of the core sensors in intelligent driving navigation systems, and its measurement accuracy directly affects the overall performance of the navigation system. However, in practical applications, the IMU is inevitably affected by various error factors, such as accelerometer zero bias, gyroscope scaling factor error, and installation angle deviation. These errors accumulate over time, leading to a gradual decrease in navigation and positioning accuracy. Therefore, how to accurately calibrate the IMU to improve its measurement accuracy and stability has become a key problem that urgently needs to be solved in the field of intelligent driving navigation. Traditional IMU calibration methods usually require specific environments and the calibration process is complex and time-consuming, making it difficult to meet the real-time and accuracy requirements of intelligent driving navigation systems. To address this, this invention proposes an IMU calibration system and method based on intelligent driving navigation to solve the above problems. Summary of the Invention

[0003] The purpose of this invention is to provide an IMU calibration system and method based on intelligent driving navigation, which can realize online joint estimation and real-time compensation of multiple error parameters of IMU during vehicle dynamic driving.

[0004] The specific technical solution adopted by this invention is as follows: An IMU calibration system and method based on intelligent driving navigation, comprising: Acquire raw measurement data, wheel speed pulse signals, and RTK-level positioning data from the vehicle-mounted IMU, and perform time synchronization and preprocessing on the raw measurement data, wheel speed pulse signals, and RTK positioning data to eliminate noise interference and timestamp deviation; The raw measurement data, wheel speed pulse signal and RTK level positioning data after time synchronization and preprocessing are fused and processed, and combined navigation calculation is performed through Kalman filter to output vehicle position, attitude and speed information in real time. A feedback correction loop is constructed with position error, attitude error and velocity error as observations. The integrated navigation solution is used as the observation value and compared with the dead reckoning result to estimate the initial installation angle between the IMU coordinate system and the vehicle coordinate system in real time. An extended state Kalman filter is constructed based on the initial installation angle of the line, including accelerometer zero bias, gyroscope scaling factor error, and installation angle residual error; Based on the extended state Kalman filter, the original measurement data of the on-board IMU and the RTK positioning information are fused during the dynamic driving process of the vehicle, and the residual installation angle error, accelerometer zero bias and gyroscope scaling factor error are jointly estimated and separated. Based on the separated residual installation angle error, accelerometer zero bias, gyroscope scaling factor error, and initial installation angle after shutdown, the raw IMU measurement data is compensated in real time, and the vehicle's final navigation status information is output synchronously.

[0005] In a preferred embodiment, the step of acquiring the raw measurement data from the vehicle-mounted IMU, wheel speed pulse signals, and RTK-level positioning data includes: The vehicle-mounted microelectromechanical inertial measurement unit (MEMS) collects raw measurement data from the triaxial accelerometer and gyroscope in real time. The wheel speed pulse signal is obtained from the vehicle's CAN bus, and the wheel speed is calculated by pulse counting conversion. RTK-level positioning data is acquired using a GNSS receiver. This positioning data includes longitude, latitude, altitude, and speed information.

[0006] In a preferred embodiment, the step of fusing the time-synchronized and preprocessed raw measurement data, wheel speed pulse signals, and RTK-level positioning data, performing combined navigation calculations through a Kalman filter, and outputting vehicle position, attitude, and speed information in real time includes: Construct a Kalman filter state vector that includes the vehicle's three-dimensional position, velocity, attitude angle, IMU accelerometer bias, gyroscope bias, and mounting angle deviation; Based on the gyroscope angular velocity and accelerometer force information in the raw IMU measurement data, the predicted values ​​of vehicle attitude, speed and position are calculated; The wheel rotation speed pulse signal is converted into vehicle speed information and compared with the predicted vehicle speed. RTK positioning data is directly used as position information as the observation value and input into the Kalman filter to update the state vector. The vehicle's position, velocity, and attitude angles are corrected based on the updated state vector, and the result is output as the navigation solution.

[0007] In a preferred embodiment, when the RTK signal is unavailable, dead reckoning is performed solely based on the onboard IMU and wheel speed pulse information.

[0008] In a preferred embodiment, the step of constructing a feedback correction loop with position error, attitude error, and velocity error as observations includes: The position, velocity, and attitude information in the integrated navigation solution are used as observations and compared with the dead reckoning position, velocity, and attitude predictions to form position error, velocity error, and attitude error. A nonlinear observation equation is constructed based on position error, velocity error and attitude error, and it is used as a feedback input to the Kalman filter for measurement update; The installation angle deviation is dynamically estimated based on the dead reckoning of the second-level Kalman filter, and closed-loop correction is performed at preset intervals during vehicle travel.

[0009] In a preferred embodiment, the step of using the integrated navigation solution as an observation value and comparing it with the dead reckoning result to estimate the initial installation angle between the IMU coordinate system and the vehicle coordinate system in real time includes: Define the angular deviation between the IMU coordinate system and the vehicle coordinate system as the installation angle, and establish the transformation relationship between the IMU coordinate system and the vehicle coordinate system; Attitude error data is generated by the difference between the attitude angles calculated by integrated navigation and the attitude angles calculated by dead reckoning. The attitude error data is correlated with the installation angle deviation to establish the correspondence between attitude error and installation angle deviation; Based on the attitude error sequence collected in real time during vehicle movement, the installation angle deviation is estimated online using a two-level Kalman filter dead reckoning, and the optimal estimate between the IMU coordinate system and the vehicle coordinate system is output as the initial installation angle for the lowering line.

[0010] In a preferred embodiment, the step of fusing raw measurement data from the onboard IMU and RTK positioning information during vehicle dynamic driving based on an extended state Kalman filter, and jointly estimating and separating the residual mounting angle error, accelerometer zero bias, and gyroscope scaling factor error, includes: The attitude error, velocity error, position error, acceleration bias, gyroscope bias, gyroscope scale factor error, and installation angle residual error are used as state variables to construct the state vector of the extended Kalman filter; The vehicle position and speed information from RTK positioning data are used as the reference benchmark for dynamic calibration, and the specific force information and angular velocity information output by the IMU are combined for state propagation. During the filter measurement update phase, the position, velocity, and attitude information calculated by the integrated navigation are used as observations to calculate the updated estimate of the state vector; Extract the optimal estimates of the remaining installation angle error, accelerometer zero bias, and gyroscope scaling factor error from the updated estimates.

[0011] In a preferred embodiment, the step of real-time compensation of the IMU raw measurement data based on the separated residual installation angle error, accelerometer zero bias, gyroscope scaling factor error, and initial installation angle after disconnection includes: Obtain a real-time estimate of the zero-bias acceleration and subtract it from the original acceleration measurement to obtain the corrected acceleration data; Based on the correction coefficients for each axis obtained from the calibration, the original angular velocity measurements are scaled proportionally to eliminate scaling factor errors; The initial installation angle and the remaining installation angle error are superimposed to generate complete installation angle correction parameters; Based on the superimposed installation angle correction parameters, the corrected acceleration and angular velocity data are transformed from the IMU coordinate system to the vehicle coordinate system; The corrected data transformed to the vehicle coordinate system is used for navigation calculations to generate the vehicle's position, speed, and attitude information in real time.

[0012] The present invention also provides an IMU calibration system based on intelligent driving navigation, using the above-mentioned IMU calibration method based on intelligent driving navigation, comprising: The data acquisition module is used to acquire raw measurement data, wheel speed pulse signals, and RTK-level positioning data from the vehicle-mounted IMU, and to perform time synchronization and preprocessing on the raw measurement data, wheel speed pulse signals, and RTK positioning data to eliminate noise interference and timestamp deviation. The navigation solution module is used to fuse the raw measurement data after time synchronization and preprocessing, wheel speed pulse signals and RTK-level positioning data, and perform combined navigation solution through Kalman filter to output vehicle position, attitude and speed information in real time. The comparison module is used to construct a feedback correction loop with position error, attitude error and velocity error as observations. It uses the integrated navigation solution as the observation value and compares it with the dead reckoning result to estimate the initial installation angle between the IMU coordinate system and the vehicle coordinate system in real time. The error separation module is used to construct an extended state Kalman filter based on the initial installation angle, including accelerometer zero bias, gyroscope scaling factor error and installation angle residual error. Based on the extended state Kalman filter, the module fuses the original measurement data of the on-board IMU and RTK positioning information during the vehicle's dynamic driving process, and jointly estimates and separates the residual installation angle error, accelerometer zero bias and gyroscope scaling factor error. The data compensation module is used to perform real-time compensation on the raw IMU measurement data based on the separated residual installation angle error, accelerometer zero bias, gyroscope scaling factor error, and initial installation angle after the line is closed, and simultaneously outputs the vehicle's final navigation status information.

[0013] And, an electronic device, the electronic device comprising: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the aforementioned IMU calibration method based on intelligent driving navigation.

[0014] The technical effects achieved by this invention are as follows: This invention achieves online joint estimation of IMU error parameters during vehicle dynamic driving by real-time acquisition and fusion of multi-source sensor data and utilizing Kalman filtering and feedback correction mechanisms. This overcomes the dependence of traditional calibration methods on static environment and offline processing. By constructing an extended state Kalman filter, accelerometer bias, gyroscope scaling factor error, and installation angle deviation are incorporated into a unified estimation framework. Combined with RTK high-precision positioning data and wheel speed information, dynamic separation and real-time compensation of error parameters are achieved. This significantly improves the navigation accuracy and long-term stability of the vehicle-mounted IMU under complex driving conditions, effectively suppresses the cumulative drift of inertial errors, and ensures the reliable positioning capability of the intelligent driving system in scenarios with temporary GNSS signal obstruction or degradation. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the system modules of the present invention; Figure 3 This is a schematic diagram of the electronic device structure of the present invention. Detailed Implementation

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0017] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0018] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that mutually excludes other embodiments.

[0019] Please see Figure 1 As shown, this invention provides an IMU calibration method based on intelligent driving navigation, comprising: S1. Acquire the raw measurement data, wheel speed pulse signal and RTK-level positioning data of the vehicle-mounted IMU, and perform time synchronization and preprocessing on the raw measurement data, wheel speed pulse signal and RTK positioning data to eliminate noise interference and timestamp deviation. In step S1, with the widespread adoption of intelligent driving navigation, the fusion positioning of vehicle-mounted IMU and RTK has become a core technology for high-precision navigation. In this embodiment, the raw measurement data and wheel speed pulse signals of the vehicle-mounted IMU are acquired through the vehicle's OBD interface. Simultaneously, RTK-level positioning data is acquired using a GNSS reference unit independent of the vehicle. Furthermore, the PPS signal from the satellite navigation system is used for hardware-level time synchronization of the raw measurement data of the IMU, wheel speed pulse signals, and RTK positioning data to ensure the time consistency of the data from each sensor. The steps of acquiring the raw measurement data of the vehicle-mounted IMU, wheel speed pulse signals, and RTK-level positioning data include: The vehicle-mounted microelectromechanical inertial measurement unit (MEMS) collects raw measurement data from the triaxial accelerometer and gyroscope in real time. The wheel speed pulse signal is obtained from the vehicle's CAN bus, and the wheel speed is calculated by pulse counting conversion. RTK-level positioning data is acquired using a GNSS receiver. The positioning data includes longitude, latitude, altitude, and speed information. Specifically, when collecting raw measurement data, an onboard microelectromechanical system (IMU) is used to achieve high-frequency data output. The sampling frequency can be set to above 100Hz to ensure data continuity and response accuracy during dynamic motion. When collecting wheel speed pulse signals, the speed pulse signals of each wheel are read from the vehicle's CAN bus, and the actual vehicle speed is calculated based on the number of pulses per unit time. Then, the vehicle's kinematic model is used for compensation processing to eliminate measurement errors caused by slippage or wheel spin. When collecting RTK positioning data, differential GNSS technology is used to obtain absolute position information with centimeter-level accuracy. Satellite signal errors are corrected in real time through network RTK or base station broadcasting to ensure the stability and reliability of the positioning results. Here, the RTK is built into the GNSS receiver, which can simultaneously receive satellite signals from multiple systems and multiple frequencies to enhance signal acquisition and tracking capabilities in complex environments. Compared with traditional vehicle-mounted solutions that do not have a separate RTK positioning module, this solution can maintain high-precision positioning output even in weak signal environments.

[0020] In the data preprocessing stage, the raw IMU measurements are... S2. The raw measurement data, wheel speed pulse signal and RTK level positioning data after time synchronization and preprocessing are fused and processed, and combined navigation calculation is performed through Kalman filter to output vehicle position, attitude and speed information in real time. In step S2, after the raw measurement data, wheel speed pulse signals, and RTK-level positioning data are acquired, corresponding preprocessing is performed simultaneously. The preprocessing methods may include data filtering, outlier removal, and timestamp alignment to ensure that the data from each sensor are aligned under a unified time reference. Then, the preprocessed multi-source data is input into an adaptive extended Kalman filter for fusion calculation, thereby achieving real-time estimation of the vehicle's position, speed, and attitude. The step of fusing the time-synchronized and preprocessed raw measurement data, wheel speed pulse signals, and RTK-level positioning data, performing combined navigation calculation through a Kalman filter, and outputting real-time vehicle position, attitude, and speed information includes: Construct a Kalman filter state vector that includes the vehicle's three-dimensional position, velocity, attitude angle, IMU accelerometer bias, gyroscope bias, and mounting angle deviation; Based on the gyroscope angular velocity and accelerometer force information in the raw IMU measurement data, the predicted values ​​of vehicle attitude, speed and position are calculated; The wheel rotation speed pulse signal is converted into vehicle speed information and compared with the predicted vehicle speed. RTK positioning data is directly used as position information as the observation value and input into the Kalman filter to update the state vector. The vehicle position, velocity, and attitude angles are corrected based on the updated state vector, and the result is output as the navigation solution. Specifically, after time synchronization and preprocessing of the raw measurement data, wheel speed pulse signals, and RTK positioning data, they are uniformly mapped to the same coordinate system to eliminate spatial installation deviations between sensors and construct a complete state vector including the vehicle's three-dimensional position, speed, attitude angles, and IMU error parameters. In the formula, Let be the inertial navigation position error vector. The inertial navigation velocity error vector, Let be the attitude error vector. This is the zero bias vector of the three-axis gyroscope. This is the zero bias vector of the triaxial accelerometer. This is the gyroscope scaling factor error vector. Let be the accelerometer scaling factor error vector, and let be the vector. Each component in the equation is a function of time. The IMU error parameters include accelerometer bias, gyroscope bias, and installation angle deviation. During the filter iteration process, the angular velocity and specific force information output by the IMU are first used to calculate the predicted values ​​of vehicle attitude, velocity, and position through kinematic equations as the state propagation stage. Then, the vehicle velocity information obtained by converting the wheel speed pulse signal through pulse counting is compared with the predicted velocity. At the same time, RTK positioning data is directly input into the Kalman filter as the position observation value. The error parameters in the state vector are corrected through the measurement update stage. In particular, for the estimation of installation angle deviation, a two-stage Kalman filter dead reckoning is used to dynamically adjust during vehicle movement. The optimal installation angle parameters are back-calculated through the continuously collected attitude error sequence. Finally, based on the updated state vector, the corrected vehicle position, velocity, and attitude information are output. It should be noted that when the RTK signal is unavailable, dead reckoning is performed only by the onboard IMU and wheel speed pulse information to maintain short-term navigation output. When the satellite signal becomes available again, the RTK positioning data is automatically reintroduced into the filter for state reset and correction, effectively suppressing integral drift accumulation.

[0021] S3. Construct a feedback correction loop with position error, attitude error and velocity error as observations, use the integrated navigation solution as the observation value, compare it with the dead reckoning result, and estimate the initial installation angle between the IMU coordinate system and the vehicle coordinate system in real time. In step S3, during the real-time output of vehicle position, velocity, and attitude, the integrated navigation solution result is used as an observation, and the difference is calculated with the dead reckoning result relying solely on IMU calculation. A feedback correction loop for position error, attitude error, and velocity error is constructed, and this is used as the measurement input for secondary correction of the Kalman filter. The initial installation angle between the vehicle coordinate system and the IMU coordinate system is estimated online. The step of constructing the feedback correction loop using position error, attitude error, and velocity error as observations includes: The position, velocity, and attitude information in the integrated navigation solution are used as observations and compared with the dead reckoning position, velocity, and attitude predictions to form position error, velocity error, and attitude error. A nonlinear observation equation is constructed based on position error, velocity error and attitude error, and it is used as a feedback input to the Kalman filter for measurement update; The installation angle deviation is dynamically estimated based on the dead reckoning of the second-level Kalman filter, and closed-loop correction is performed at preset intervals during vehicle travel. Specifically, when constructing the feedback correction loop, the position, velocity, and attitude information obtained from the integrated navigation solution are first used as benchmark observations. These are compared item by item with the predictions independently calculated by the dead reckoning system to generate a residual vector containing three-dimensional position error, velocity error, and attitude error. A nonlinear observation equation is constructed based on error propagation theory. After linearization through Taylor expansion, the equation is input into a Kalman filter for measurement updates. Simultaneously, a two-stage Kalman filter dead reckoning is used to dynamically estimate the installation angle deviation. During vehicle operation, a closed-loop correction mechanism is triggered according to preset time intervals or mileage. The installation angle parameters are iteratively optimized through continuously collected error sequences to ensure the real-time performance and accuracy of the initial deviation estimation.

[0022] Secondly, the step of using the integrated navigation solution results as observations and comparing them with dead reckoning results to estimate the initial installation angle between the IMU coordinate system and the vehicle coordinate system in real time includes: Define the angular deviation between the IMU coordinate system and the vehicle coordinate system as the installation angle, and establish the transformation relationship between the IMU coordinate system and the vehicle coordinate system; Attitude error data is generated by the difference between the attitude angles calculated by integrated navigation and the attitude angles calculated by dead reckoning. The attitude error data is correlated with the installation angle deviation to establish the correspondence between attitude error and installation angle deviation; Based on the attitude error sequence collected in real time during vehicle movement, the installation angle deviation is estimated online using two-level Kalman filter dead reckoning, and the optimal estimated value between the IMU coordinate system and the vehicle coordinate system is output as the initial installation angle for offline installation. In estimating the initial installation angle, it is first necessary to clarify the angular deviation between the IMU coordinate system and the vehicle coordinate system, i.e., the installation angle, and to establish an accurate transformation relationship between the two. Specifically, a coordinate transformation model can be established using the direction cosine matrix or quaternion method. The high-precision attitude angle output by the integrated navigation system is used as a reference, and the difference between it and the attitude angle calculated by the inertial navigation system is calculated to obtain the real-time attitude error sequence. This sequence is then correlated with the installation angle deviation through a linearized observation equation. Combining multiple sets of error data during vehicle motion, a second-level Kalman filter dead reckoning is used for parameter identification. Alternatively, methods such as recursive least squares can be used to achieve dynamic convergence and optimal estimation of the installation angle deviation, thereby improving the initial alignment accuracy. To improve navigation stability and effectively suppress attitude divergence and positioning drift caused by installation angle deviation, continuous iterative optimization ensures high-precision navigation performance in complex driving environments, further enhancing the system's robustness and reliability. It automatically calls historical best estimates as prior information each time the vehicle starts, shortening alignment time and improving dynamic response efficiency. Combined with multi-source sensor data, it performs corresponding joint optimization and introduces an adaptive filtering mechanism to dynamically adjust the noise covariance matrix according to the motion state, effectively suppressing estimation jitter under abrupt changes. Simultaneously, it integrates GNSS availability indicators and vehicle acceleration characteristics to construct observation validity criteria, avoiding misleading installation angle parameter identification by abnormal data.

[0023] S4. Construct an extended state Kalman filter based on the initial installation angle of the lower line, including accelerometer zero bias, gyroscope scaling factor error and installation angle residual error; In step S4, after the initial installation angle is output, an extended state vector is constructed based on it. The accelerometer zero bias, gyroscope scaling factor error and installation angle residual deviation are incorporated into the filter state variables, and a corresponding extended state Kalman filter is constructed to provide a basis for subsequent filter correction.

[0024] S5. Based on the extended state Kalman filter, the original measurement data of the on-board IMU and the RTK positioning information are fused during the dynamic driving process of the vehicle to jointly estimate and separate the residual installation angle error, accelerometer zero bias and gyroscope scaling factor error. In step S5, during the operation of the extended state Kalman filter, the raw force and angular velocity data output by the IMU and the position and velocity information provided by the RTK are fused in real time. Through the iterative process of state prediction and measurement update, the state estimate value inside the filter is continuously corrected. The high-precision position and velocity of the RTK are used as observations to construct a residual sequence, driving the Kalman gain to dynamically adjust the convergence rate of each error term. Even under complex conditions such as vehicle acceleration, steering, or bumps, it can still effectively separate the accelerometer zero bias and the gyroscope scaling factor error. The step of fusing the raw measurement data of the on-board IMU and the RTK positioning information during the dynamic driving process of the vehicle based on the extended state Kalman filter, and jointly estimating and separating the residual installation angle error, accelerometer zero bias, and gyroscope scaling factor error includes: The attitude error, velocity error, position error, acceleration bias, gyroscope bias, gyroscope scale factor error, and installation angle residual error are used as state variables to construct the state vector of the extended Kalman filter; The vehicle position and speed information from RTK positioning data are used as the reference benchmark for dynamic calibration, and the specific force information and angular velocity information output by the IMU are combined for state propagation. During the filter measurement update phase, the position, velocity, and attitude information calculated by the integrated navigation are used as observations to calculate the updated estimate of the state vector; Extract the optimal estimates of the remaining installation angle error, accelerometer zero bias, and gyroscope scaling factor error from the updated estimates; Specifically, in constructing the extended Kalman filter, the attitude error, velocity error, position error, and the IMU's acceleration bias, gyroscope bias, gyroscope scaling factor error, and residual installation angle error are collectively used to form the state vector of the extended Kalman filter. The state vector encompasses the vehicle's motion state and IMU error parameters. The vehicle position and velocity information provided by RTK positioning data is used as the reference benchmark for dynamic calibration. Simultaneously, the specific force information and angular velocity information output by the IMU in real time are combined, and state propagation is performed according to the vehicle's kinematic equations. During the state propagation process, the predicted values ​​of vehicle attitude, velocity, and position are calculated using IMU data. In the filter's measurement update phase, the position, velocity, and attitude information obtained from the integrated navigation solution are used as observations. The residuals are calculated by comparing them with the predicted values, and then the estimated values ​​of the state vector are updated. Through continuous iteration of this process, the various parameters in the state vector are continuously corrected. Finally, the optimal estimated values ​​of the residual installation angle error, accelerometer bias, and gyroscope scaling factor error are extracted from the updated estimated values. This achieves the separation and estimation of IMU error parameters, effectively improving the IMU's measurement accuracy and navigation performance.

[0025] S6. Based on the separated residual installation angle error, accelerometer zero bias, gyroscope scaling factor error, and initial installation angle after shutdown, the IMU raw measurement data is compensated in real time, and the vehicle's final navigation status information is output synchronously. In step S6, after the error parameters such as accelerometer bias, gyroscope scaling factor error, and residual installation angle error are separated, a compensation matrix is ​​constructed in conjunction with the initial installation angle to perform real-time compensation on the specific force and angular velocity data of the IMU's original output, eliminating the influence of systematic errors. The compensated data is used to update the vehicle's attitude, speed, and position information to ensure the accuracy of navigation calculation, thereby effectively suppressing error accumulation during long-term operation and improving the navigation stability and reliability of the vehicle in complex environments. The step of real-time compensation of the IMU's original measurement data based on the separated residual installation angle error, accelerometer bias, gyroscope scaling factor error, and initial installation angle includes: Obtain a real-time estimate of the zero-bias acceleration and subtract it from the original acceleration measurement to obtain the corrected acceleration data; Based on the correction coefficients for each axis obtained from dynamic calibration, the original angular velocity measurements are scaled proportionally to eliminate scaling factor errors. The initial installation angle deviation and the residual installation angle error are superimposed to generate complete installation angle correction parameters; Based on the superimposed installation angle correction parameters, the corrected acceleration and angular velocity data are transformed from the IMU coordinate system to the vehicle coordinate system; The corrected data transformed to the vehicle coordinate system is used for navigation calculation, and the vehicle's position, speed and attitude information are generated in real time. Specifically, when compensating for the raw IMU data, the first step is to obtain the real-time estimate of the accelerometer zero bias separated by the extended state Kalman filter. This zero bias value is then directly subtracted from the raw acceleration measurement data to obtain corrected acceleration data free from systematic bias. Subsequently, based on the gyroscope scaling factor error coefficients identified during dynamic calibration, the raw angular velocity measurement values ​​are scaled to eliminate range scaling bias caused by sensor manufacturing processes. Simultaneously, the initial installation angle estimate and the residual installation angle error separated during the extended filtering stage are algebraically superimposed to form a complete three-dimensional installation angle correction parameter matrix. Based on this correction parameter matrix, the corrected acceleration and angular velocity data are transformed from the IMU body coordinate system to the vehicle coordinate system using a coordinate transformation matrix. The corresponding coordinate transformation matrix is ​​as follows: The rotation matrix around the y-axis and the rotation matrix around the Z-axis are used to transform the IMU data to the vehicle coordinate system, eliminate the measurement value coupling error caused by the sensor installation posture, and finally perform integration and attitude calculation on the corrected measurement data transformed to the vehicle coordinate system to generate error-compensated vehicle position, speed and attitude navigation information in real time, ensuring that high-precision navigation performance can still be maintained under complex driving conditions. It should be noted that, to ensure the continuous stability of vehicle navigation performance, the initial installation angle between the IMU and the vehicle coordinate system is calibrated during the vehicle off-line stage (which includes the final testing stage before new vehicles leave the factory, the vehicle maintenance stage, and the vehicle's regular inspection stage). Specifically, the initial transformation matrix between the IMU coordinate system and the vehicle coordinate system can be obtained through a multi-position static calibration method. In practice, the vehicle is placed horizontally, and the vehicle is adjusted to multiple preset attitude angles and kept stationary. At each attitude angle, the specific force and angular velocity data output by the IMU are continuously collected. The initial installation angle between the IMU coordinate system and the vehicle coordinate system is fitted using a commonly used horizontal installation angle calibration algorithm, and this is used as the initial installation angle for subsequent navigation processes. The initial reference value is used, and during vehicle maintenance or periodic inspection, the IMU mounting angle is also re-measured using a multi-position static calibration method to detect changes in mounting posture caused by mechanical vibration or collision during long-term use. Based on this, it is automatically loaded as prior information when the vehicle is first started. Combined with the aforementioned second-level Kalman filter, the residual mounting angle error is estimated online, realizing the fusion of the initial reference and real-time correction, shortening the initial alignment time, improving the navigation convergence speed, and thus dealing with the possible deformation of the mounting structure during long-term use of the vehicle. This enables online estimation and compensation of mounting angle drift throughout the vehicle's entire life cycle, thereby ensuring the long-term stability of the mounting angle parameters and the continuous optimization of navigation accuracy.

[0026] Please see Figure 2 An IMU calibration system based on intelligent driving navigation, using the aforementioned IMU calibration method based on intelligent driving navigation, includes: The data acquisition module is used to acquire raw measurement data, wheel speed pulse signals, and RTK-level positioning data from the vehicle-mounted IMU, and to perform time synchronization and preprocessing on the raw measurement data, wheel speed pulse signals, and RTK positioning data to eliminate noise interference and timestamp deviation. The navigation solution module is used to fuse the raw measurement data after time synchronization and preprocessing, wheel speed pulse signals and RTK-level positioning data, and perform combined navigation solution through Kalman filter to output vehicle position, attitude and speed information in real time. The comparison module is used to construct a feedback correction loop with position error, attitude error and velocity error as observations. It uses the integrated navigation solution as the observation value and compares it with the dead reckoning result to estimate the initial installation angle between the IMU coordinate system and the vehicle coordinate system in real time. The error separation module is used to construct an extended state Kalman filter based on the initial installation angle, including accelerometer zero bias, gyroscope scaling factor error and installation angle residual error. Based on the extended state Kalman filter, the module fuses the original measurement data of the on-board IMU and RTK positioning information during the vehicle's dynamic driving process, and jointly estimates and separates the residual installation angle error, accelerometer zero bias and gyroscope scaling factor error. The data compensation module is used to compensate the raw IMU measurement data in real time based on the separated residual installation angle error, accelerometer zero bias, gyroscope scaling factor error and initial installation angle after the line is closed, and simultaneously output the vehicle's final navigation status information. The calibration system described above corresponds to the aforementioned calibration method, and will not be repeated here.

[0027] Please see Figure 3 An electronic device, the electronic device comprising: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the aforementioned IMU calibration method based on intelligent driving navigation.

[0028] The above description is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained in this invention are implemented according to conventional methods in the art unless otherwise specified or limited.

Claims

1. An IMU calibration method based on intelligent driving navigation, characterized in that: include: Acquire raw measurement data, wheel speed pulse signals, and RTK-level positioning data from the vehicle-mounted IMU, and perform time synchronization and preprocessing on the raw measurement data, wheel speed pulse signals, and RTK positioning data to eliminate noise interference and timestamp deviation; The raw measurement data, wheel speed pulse signal and RTK level positioning data after time synchronization and preprocessing are fused and processed, and combined navigation calculation is performed through Kalman filter to output vehicle position, attitude and speed information in real time. A feedback correction loop is constructed with position error, attitude error and velocity error as observations. The integrated navigation solution is used as the observation value and compared with the dead reckoning result to estimate the initial installation angle between the IMU coordinate system and the vehicle coordinate system in real time. An extended state Kalman filter is constructed based on the initial installation angle of the line, including accelerometer zero bias, gyroscope scaling factor error, and installation angle residual error; Based on the extended state Kalman filter, the original measurement data of the on-board IMU and the RTK positioning information are fused during the dynamic driving process of the vehicle, and the residual installation angle error, accelerometer zero bias and gyroscope scaling factor error are jointly estimated and separated. Based on the separated residual installation angle error, accelerometer zero bias, gyroscope scaling factor error, and initial installation angle after shutdown, the raw IMU measurement data is compensated in real time, and the vehicle's final navigation status information is output synchronously.

2. The IMU calibration method based on intelligent driving navigation according to claim 1, characterized in that: The steps for acquiring raw measurement data from the vehicle-mounted IMU, wheel speed pulse signals, and RTK-level positioning data include: The vehicle-mounted microelectromechanical inertial measurement unit (MEMS) collects raw measurement data from the triaxial accelerometer and gyroscope in real time. The wheel speed pulse signal is obtained from the vehicle's CAN bus, and the wheel speed is calculated by pulse counting conversion. RTK-level positioning data is acquired using a GNSS receiver. This positioning data includes longitude, latitude, altitude, and speed information.

3. The IMU calibration method based on intelligent driving navigation according to claim 1, characterized in that: The steps of fusing the time-synchronized and preprocessed raw measurement data, wheel speed pulse signals, and RTK-level positioning data, performing combined navigation calculations using a Kalman filter, and outputting vehicle position, attitude, and speed information in real time include: Construct a Kalman filter state vector that includes the vehicle's three-dimensional position, velocity, attitude angle, IMU accelerometer bias, gyroscope bias, and mounting angle deviation; Based on the gyroscope angular velocity and accelerometer force information in the raw IMU measurement data, the predicted values ​​of vehicle attitude, speed and position are calculated; The wheel rotation speed pulse signal is converted into vehicle speed information and compared with the predicted vehicle speed. RTK positioning data is directly used as position information as the observation value and input into the Kalman filter to update the state vector. The vehicle's position, velocity, and attitude angles are corrected based on the updated state vector, and the result is output as the navigation solution.

4. The IMU calibration method based on intelligent driving navigation according to claim 1, characterized in that: When the RTK signal is unavailable, dead reckoning is performed solely based on the onboard IMU and wheel speed pulse information.

5. The IMU calibration method based on intelligent driving navigation according to claim 1, characterized in that: The step of constructing a feedback correction loop with position error, attitude error, and velocity error as observations includes: The position, velocity, and attitude information in the integrated navigation solution are used as observations and compared with the dead reckoning position, velocity, and attitude predictions to form position error, velocity error, and attitude error. A nonlinear observation equation is constructed based on position error, velocity error and attitude error, and it is used as a feedback input to the Kalman filter for measurement update; The installation angle deviation is dynamically estimated based on the dead reckoning of the second-level Kalman filter, and closed-loop correction is performed at preset intervals during vehicle travel.

6. The IMU calibration method based on intelligent driving navigation according to claim 1, characterized in that: The step of using the integrated navigation solution results as observations and comparing them with dead reckoning results to estimate the initial installation angle between the IMU coordinate system and the vehicle coordinate system in real time includes: Define the angular deviation between the IMU coordinate system and the vehicle coordinate system as the installation angle, and establish the transformation relationship between the IMU coordinate system and the vehicle coordinate system; Attitude error data is generated by the difference between the attitude angles calculated by integrated navigation and the attitude angles calculated by dead reckoning. The attitude error data is correlated with the installation angle deviation to establish the correspondence between attitude error and installation angle deviation; Based on the attitude error sequence collected in real time during vehicle movement, the installation angle deviation is estimated online using a two-level Kalman filter dead reckoning, and the optimal estimate between the IMU coordinate system and the vehicle coordinate system is output as the initial installation angle for the lowering line.

7. The IMU calibration method based on intelligent driving navigation according to claim 1, characterized in that: The steps of fusing raw measurement data from the onboard IMU and RTK positioning information during vehicle dynamic driving based on the extended state Kalman filter, and jointly estimating and separating the residual installation angle error, accelerometer zero bias, and gyroscope scaling factor error, include: The attitude error, velocity error, position error, acceleration bias, gyroscope bias, gyroscope scale factor error, and installation angle residual error are used as state variables to construct the state vector of the extended Kalman filter; The vehicle position and speed information from RTK positioning data are used as the reference benchmark for dynamic calibration, and the specific force information and angular velocity information output by the IMU are combined for state propagation. During the filter measurement update phase, the position, velocity, and attitude information calculated by the integrated navigation are used as observations to calculate the updated estimate of the state vector; Extract the optimal estimates of the remaining installation angle error, accelerometer zero bias, and gyroscope scaling factor error from the updated estimates.

8. The IMU calibration method based on intelligent driving navigation according to claim 1, characterized in that: The step of real-time compensation of the IMU's raw measurement data based on the separated residual installation angle error, accelerometer zero bias, gyroscope scaling factor error, and initial installation angle after disconnection includes: Obtain a real-time estimate of the zero-bias acceleration and subtract it from the original acceleration measurement to obtain the corrected acceleration data; Based on the correction coefficients for each axis obtained from the calibration, the original angular velocity measurements are scaled proportionally to eliminate scaling factor errors; The initial installation angle and the remaining installation angle error are superimposed to generate complete installation angle correction parameters; Based on the superimposed installation angle correction parameters, the corrected acceleration and angular velocity data are transformed from the IMU coordinate system to the vehicle coordinate system; The corrected data transformed to the vehicle coordinate system is used for navigation calculations to generate the vehicle's position, speed, and attitude information in real time.

9. An IMU calibration system based on intelligent driving navigation, characterized in that: The IMU calibration method based on intelligent driving navigation according to any one of claims 1 to 8 includes: The data acquisition module is used to acquire raw measurement data, wheel speed pulse signals, and RTK-level positioning data from the vehicle-mounted IMU, and to perform time synchronization and preprocessing on the raw measurement data, wheel speed pulse signals, and RTK positioning data to eliminate noise interference and timestamp deviation. The navigation solution module is used to fuse the raw measurement data after time synchronization and preprocessing, wheel speed pulse signals and RTK-level positioning data, and perform combined navigation solution through Kalman filter to output vehicle position, attitude and speed information in real time. The comparison module is used to construct a feedback correction loop with position error, attitude error and velocity error as observations. It uses the integrated navigation solution as the observation value and compares it with the dead reckoning result to estimate the initial installation angle between the IMU coordinate system and the vehicle coordinate system in real time. The error separation module is used to construct an extended state Kalman filter based on the initial installation angle, including accelerometer zero bias, gyroscope scaling factor error and installation angle residual error. Based on the extended state Kalman filter, the module fuses the original measurement data of the on-board IMU and RTK positioning information during the vehicle's dynamic driving process, and jointly estimates and separates the residual installation angle error, accelerometer zero bias and gyroscope scaling factor error. The data compensation module is used to perform real-time compensation on the raw IMU measurement data based on the separated residual installation angle error, accelerometer zero bias, gyroscope scaling factor error, and initial installation angle after the line is closed, and simultaneously outputs the vehicle's final navigation status information.

10. An electronic device, characterized in that: The electronic device includes: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the IMU calibration method based on intelligent driving navigation as described in any one of claims 1 to 8.

Images