Vehicle pose initialization method, device and equipment, medium and vehicle

Abstract

The invention relates to the technical field of vehicle automatic driving, and discloses a vehicle pose initialization method, device and equipment, a medium and a vehicle, the method comprises the following steps: acquiring satellite positioning data, inertial navigation data and perception observation data of the vehicle, the satellite positioning data comprising state information, position data and pose data; determining the step length and range of the position data and the attitude data according to the state information, creating a first particle set, and performing pose prediction on each particle according to the inertial navigation data; screening the first particle set according to the perception observation data to obtain a second particle set; and performing state evaluation according to the second particle set, and determining the initial pose of the vehicle. According to the method, it can be guaranteed that the initial particle set has enough coverage range to offset the error influence caused by the poor state of the satellite positioning system, meanwhile, auxiliary positioning is carried out by means of front-end sensing equipment, pose initialization in various driving scenes is achieved, and the robustness, adaptability and accuracy of the positioning system are improved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle automatic driving technology, and in particular to a vehicle posture initialization method, device, equipment, medium and vehicle. Background Art

[0002] The positioning technology of autonomous driving vehicles is an important part of autonomous driving technology. At present, it is mostly based on multi-sensor fusion positioning. The sensors used include combined inertial navigation, cameras and lidar. The positioning system is the premise of path planning and decision control of autonomous driving vehicles. Among them, posture initialization is an important basis for the positioning system to output real-time, continuous and accurate posture during autonomous driving.

[0003] In the existing technology, pose initialization is highly dependent on the positioning status of the satellite positioning system. Position initialization can only be achieved when the positioning result is relatively accurate. When the satellite positioning system is in poor condition due to obstruction by tall buildings and trees in urban areas or in tunnels, the position may deviate from the lane, resulting in a long pose initialization time or even failure to complete initialization. Summary of the invention

[0004] In view of this, the present invention provides a vehicle posture initialization method, device, equipment, medium and vehicle to solve the problem of limited posture initialization.

[0005] In a first aspect, the present invention provides a vehicle posture initialization method, the method comprising:

[0006] Obtain the vehicle's satellite positioning data, inertial navigation data, and perception observation data. The satellite positioning data includes status information, position data, and attitude data;

[0007] Determine the step size and range of the position data and the attitude data according to the state information, create a first particle set based on the step size and the range, and predict the position and attitude of each particle in the first particle set according to the inertial navigation data;

[0008] The particles in the first particle set are screened according to the predicted particle position and the sensed observation data to obtain the second particle set with updated position and posture;

[0009] A state evaluation is performed according to the second particle set, and an initialization pose of the vehicle is determined according to the evaluation result.

[0010] The vehicle posture initialization method provided by the embodiment of the present invention determines the step size and range of the position data and posture data according to the status information in the satellite positioning data of the vehicle, creates a first particle set, and predicts the posture of each particle in the first particle set according to the inertial navigation data, screens the first particle set according to the predicted particle posture and the perception observation data, obtains the second particle set after the posture is updated, and performs state evaluation according to the second particle set to determine the initialization posture of the vehicle. The present invention sets different particle sets according to different satellite status information, which can ensure that the initial particle set has sufficient coverage to offset the error caused by the poor state of the satellite positioning system, and at the same time, assists in positioning with the help of the front-end sensing device, so as to complete the posture initialization in various driving scenarios and improve the robustness, adaptability and accuracy of the positioning system.

[0011] In an optional embodiment, satellite positioning data, inertial navigation data and perception observation data are pre-time-aligned, and state information includes a fixed solution state, a floating-point solution state, a differential positioning state and a single-point positioning state. The satellite positioning data, inertial navigation data and perception observation data of the vehicle are obtained, including: obtaining the satellite positioning data and inertial navigation data of the current frame, and determining whether there is perception observation data in the current frame; if there is perception observation data, obtaining a satellite positioning data set of a preset number of frames before the current frame, and determining the optimal satellite positioning state closest to the current frame in the satellite positioning data set according to the state priority, wherein the state priorities are, from high to low, fixed solution state, floating-point solution state, differential positioning state and single-point positioning state; taking the data frame corresponding to the optimal satellite positioning state as the starting frame, taking the satellite positioning data corresponding to the starting frame as the initial satellite positioning data, and obtaining the inertial navigation data set from the starting frame to the current frame.

[0012] By aligning the timestamps of the combined inertial navigation device and the front-end sensing device in advance, the present invention can use the front-end sensing device to assist in positioning during vehicle driving, complete posture initialization anytime and anywhere, and perform posture prediction based on satellite positioning data in the optimal satellite positioning state, which can improve the adaptability and accuracy of posture initialization.

[0013] In an optional embodiment, the position data includes a lateral position and a longitudinal position, the attitude data is a heading angle, the step size and range of the position data and the attitude data are determined according to the state information, and a first particle set is created based on the step size and range, including: determining a first range and a first step size of the heading angle, a second range and a second step size of the lateral position, and a third range and a third step size of the longitudinal position according to the state information of the initial satellite positioning data; determining the number of particles in the first particle set and the particle posture of each particle according to the first range, the first step size, the second range, the second step size, the third range and the third step size to obtain the first particle set; and assigning a normalized weight to each particle in the first particle set to determine the particle weight of each particle.

[0014] The present invention constructs different particle sets according to the status information of the satellite positioning system, which can ensure that the initial particle set has sufficient coverage to offset the error influence caused by the poor status of the satellite positioning system, so that posture initialization can be achieved under different satellite positioning states, and the posture initialization scenario is not restricted, thereby improving the reliability of the positioning system.

[0015] In an optional embodiment, the inertial navigation data includes wheel speed and angular velocity, and the position and posture of each particle in the first particle set is predicted according to the inertial navigation data, including: constructing a motion model, and taking the starting frame corresponding to the initial satellite positioning data as the prediction frame starting point of the motion model; based on the motion model, predicting the particle posture of each particle in the next prediction frame according to the wheel speed, angular velocity and particle posture of each particle in the current prediction frame in the inertial navigation data set, until the current frame is reached.

[0016] The present invention predicts and infers the posture of each particle in an ion set based on satellite positioning data and inertial navigation data under the optimal satellite positioning state, and can obtain the posture of each particle in the particle set at a subsequent moment, so that the posture of each particle can be evaluated, the most accurate particle posture can be determined, and the most accurate particle posture can be used as the initialization posture.

[0017] In an optional embodiment, each particle in the first particle set is screened according to the predicted particle posture and the sensed observation data to obtain a second particle set with updated posture, including: when the current predicted frame reaches the current frame, based on the particle posture of each particle corresponding to the current predicted frame, the sensed observation data is projected into a preset dynamic map, and the sensed observation data is matched with the preset dynamic map data; the particle weight of each particle is updated according to the matching result, and the updated particle weight is compared with the first preset weight threshold; if the updated particle weight is lower than the first preset weight threshold, the particle corresponding to the updated particle weight is deleted from the first particle set to obtain the second particle set.

[0018] The present invention can delete particles with large errors in the particle set by screening the particle set based on the perception observation data, realize the auxiliary positioning of the front-end perception device, and improve the positioning accuracy. In addition, the particle weight of each particle is updated at the same time according to the matching result, which is more efficient.

[0019] In an optional embodiment, a state evaluation is performed according to the second particle set, and an initialization posture of the vehicle is determined according to the evaluation result, including: determining the number of particles in the second particle set, and comparing the number of particles with a preset number threshold; if the number of particles is lower than the preset number threshold, comparing the particle weight of each particle in the second particle set with the second preset weight threshold, otherwise returning to the step of predicting the posture of each particle in the first particle set according to the inertial navigation data; if there is at least one particle weight greater than the second preset weight threshold, the particle posture of the particle corresponding to the maximum particle weight in the second particle set is used as the initialization posture; if not, judging the degree of similarity of the posture of each particle in the second particle set; if the degree of similarity of the posture meets the preset threshold, the particle posture of the particle corresponding to the maximum particle weight in the second particle set is used as the initialization posture, otherwise returning to the step of predicting the posture of each particle in the first particle set according to the inertial navigation data.

[0020] The present invention determines the most accurate initialization posture through step-by-step iteration by judging the multi-layer convergence conditions, thereby improving the accuracy of state assessment.

[0021] In a second aspect, the present invention provides a vehicle posture initialization device, the device comprising: a data acquisition module, used to acquire satellite positioning data, inertial navigation data and sensory observation data of the vehicle, the satellite positioning data comprising state information, position data and posture data; a posture prediction module, used to determine the step size and range of the position data and the posture data according to the state information, create a first particle set based on the step size and the range, and perform posture prediction on each particle in the first particle set according to the inertial navigation data; a particle update module, used to screen each particle in the first particle set according to the predicted particle posture and sensory observation data to obtain a second particle set after the posture is updated; a posture determination module, used to perform state evaluation according to the second particle set, and determine the initialization posture of the vehicle according to the evaluation result.

[0022] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor are communicatively connected to each other, computer instructions are stored in the memory, and the processor executes the vehicle posture initialization method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0023] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the vehicle posture initialization method of the first aspect or any corresponding embodiment thereof.

[0024] In a fifth aspect, the present invention provides a computer program product, comprising computer instructions, wherein the computer instructions are used to enable a computer to execute the vehicle posture initialization method of the above-mentioned first aspect or any corresponding embodiment thereof.

[0025] In a sixth aspect, the present invention provides a vehicle, comprising: a combined navigation system, the combined navigation system comprising: a satellite positioning system and an inertial measurement unit, respectively used to collect satellite positioning data and inertial navigation data; a front-end perception device, time-synchronized with the combined navigation system, for collecting perception observation data; a controller, connected to the combined navigation system and the front-end perception device, for executing the vehicle posture initialization method of the above-mentioned first aspect or any corresponding embodiment thereof.

[0026] Beneficial effects of the present invention:

[0027] (1) The present invention determines the step size and range of position data and attitude data through the status information of satellite positioning data, which can ensure that the initial particle set has sufficient coverage to offset the error caused by the poor status of the satellite positioning system, and realize the initialization of posture in various driving scenarios;

[0028] (2) The present invention uses front-end sensing equipment to assist in positioning, and can complete posture initialization anytime and anywhere during driving, thereby improving the robustness and adaptability of the positioning system;

[0029] (3) The present invention determines the most accurate initialization posture through step-by-step iteration by judging the multi-layer convergence conditions, thereby improving the accuracy of state assessment, and then obtaining an accurate initialization posture, thereby improving the accuracy of the positioning system. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 is a schematic flow chart of a vehicle posture initialization method according to an embodiment of the present invention;

[0032] Figure 2is a schematic diagram of an algorithm flow of a vehicle posture initialization method according to an embodiment of the present invention;

[0033] Figure 3 is a flow chart of another vehicle posture initialization system according to an embodiment of the present invention;

[0034] Figure 4 is a flow chart of another vehicle posture initialization method according to an embodiment of the present invention;

[0035] Figure 5 is a structural block diagram of a vehicle posture initialization device according to an embodiment of the present invention;

[0036] Figure 6 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0038] The embodiment of the present invention is applicable to the scenario of determining the initialization posture during the process of autonomous driving of a vehicle, so as to perform path planning or decision control. The embodiment of the present invention provides a vehicle posture initialization method, which achieves the effect of improving the adaptability and accuracy of posture initialization by constructing a particle set according to satellite status information and determining the initialization posture according to inertial navigation data and perception observation data.

[0039] According to an embodiment of the present invention, an embodiment of a vehicle posture initialization method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0040] In this embodiment, a vehicle posture initialization method is provided, which can be used for a controller of a vehicle. Figure 1 is a flow chart of a vehicle posture initialization method according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0041] Step S101, obtaining satellite positioning data, inertial navigation data and sensory observation data of the vehicle, where the satellite positioning data includes status information, position data and attitude data.

[0042] Specifically, in an embodiment of the present invention, the autonomous driving vehicle is pre-installed with an integrated navigation system (INS) and a front-end sensing device. The integrated navigation system INS includes: a satellite positioning system (e.g., a global positioning system, GPS) and an inertial measurement unit (IMU). Taking GPS as an example, the GPS signal contains a status bit and a posture data. The status bit represents the status information, which is determined by the number of GPS satellite searches. The GPS status bit includes RTK fixed solution, RTK floating point solution, differential positioning, single point positioning, etc. from best to worst; the posture data of the embodiment of the present invention includes position data and attitude data, the position data is the lateral position and the longitudinal position, and the attitude data is the heading angle, which is only used as an example and is not limited to this. The inertial measurement unit IMU includes three single-axis accelerometers and three single-axis gyroscopes. The accelerometer detects the acceleration signal of the object in the independent three-axis of the carrier coordinate system, and the gyroscope detects the angular velocity signal of the carrier relative to the navigation coordinate system, so the IMU can measure the angular velocity and acceleration of the object in three-dimensional space. The front-end sensing devices include a forward-looking camera and a laser radar, and the sensing observation data obtained are the forward-looking camera data and the laser radar data, respectively, for example only and not limited thereto. In addition, the combined inertial navigation device and the forward-looking camera and the laser radar corresponding to the combined navigation system have completed time synchronization in advance, that is, based on the same clock source for timing, thereby avoiding the limitation that the timestamps of the two data sources are not strictly aligned due to the different frequencies of the front-end sensing devices such as the laser radar and the combined inertial navigation device, resulting in the limitation that the posture initialization can only be completed when the vehicle is stationary.

[0043] In some optional implementations, the data collection frequency of INS is different from that of the front-end sensing device. Generally, the data collection frequency of INS is higher than that of the front-end sensing device. For example, the frequency of INS is 10 Hz and that of the front-end camera is 30 Hz. Therefore, two frames of front-end camera data correspond to multiple frames of satellite positioning data and inertial navigation data. The embodiment of the present invention obtains satellite positioning data, inertial navigation data, and sensing observation data with the same timestamp as the current frame data and as the data input of the embodiment of the present invention, such as Figure 2 shown.

[0044] Step S102, determining the step size and range of the position data and the attitude data according to the state information, creating a first particle set based on the step size and the range, and performing position and attitude prediction on each particle in the first particle set according to the inertial navigation data.

[0045] Specifically, in the embodiment of the present invention, different GPS status bits represent different positioning accuracy. If the status of the satellite positioning data corresponding to the current frame is poor, the posture formed by the corresponding position data and posture data may be inaccurate. Therefore, the optimal satellite positioning state closest to the current frame is searched within the range of the satellite navigation data of a preset number of frames before the current frame, such as the first 500 frames of INS data of the current frame, and the data frame corresponding to the optimal satellite positioning state is used as the starting frame. The step size and range of the position data and posture data in the starting frame are determined according to the status information of the starting frame, that is, the step size and range of the lateral position, longitudinal position and heading angle are determined respectively, and the lateral position, longitudinal position and heading angle are taken according to the step size within the corresponding range, so as to construct a first particle set centered on this posture, such as Figure 2 As shown in the figure. The step sizes and ranges corresponding to the lateral position, longitudinal position and heading angle under different GPS status information are different, so the coverage range and sparsity of the constructed first particle set are also different. When the accuracy of the GPS status bit representation is high, the coverage range is small and the number of particles is small. When the accuracy of the status bit is low, the coverage range is large and the number of particles is large, thereby ensuring that the coverage range of the initial state particle set can offset the errors under different GPS status bits, while taking into account the convergence speed of the particle set, with high robustness, and can meet the initialization requirements in scenarios with poor GPS status, such as high-rise buildings and trees, overpasses, tunnels, etc.

[0046] In some optional implementations, the first particle set includes particles in multiple posture states corresponding to the starting frame. Therefore, the embodiment of the present invention constructs a motion model, starting from the starting frame, and predicts and infers the posture state of the particles based on the angular velocity and wheel speed in the inertial navigation data, such as Figure 2 As shown, the position and posture of each particle in the particle set of the subsequent frame is obtained until the current frame is reached, and multiple particles with different position and posture states in the current frame are obtained.

[0047] Step S103, screening each particle in the first particle set according to the predicted particle posture and the sensed observation data to obtain a second particle set with updated posture.

[0048] Specifically, in an embodiment of the present invention, the first particle set includes multiple particles with different posture states in the current frame, and it is impossible to determine which particle's posture is more accurate. Therefore, based on the predicted particle posture, each particle in the first particle set is screened according to the perceived observation data, that is, auxiliary positioning is performed with the help of the front-end perception device. The embodiment of the present invention performs semantic recognition on the perceived observation data to obtain semantic information of the perceived observation data, such as lane lines, curbs, stop lines, road surface signs and other information, and then projects the semantic information into the dynamic map data based on the posture state of each particle, matches the semantic information in the perceived observation data with the semantic information in the dynamic map, screens each particle according to the matching result, deletes particles with larger posture state errors, and obtains a second particle set with relatively accurate particle posture, so as to achieve posture update, such as Figure 2 shown.

[0049] Step S104, performing state evaluation according to the second particle set, and determining the initialization posture of the vehicle according to the evaluation result.

[0050] Specifically, in the embodiment of the present invention, the second particle set after screening contains at least one particle, and when it contains multiple particles, it is necessary to further find the particle with the most accurate posture. Therefore, the state of the second particle set is evaluated, and the evaluation indicators include: the number of particles in the second particle set, the particle weight, and the degree of similarity between the postures of the particles, where the number of particles is used as the first-level convergence judgment condition, the particle weight is used as the second-level convergence judgment condition, and the degree of similarity between the postures of the particles is used as the third-level convergence judgment condition. Figure 2 As shown, if any convergence judgment condition is not met during the evaluation process, the posture prediction and posture update are repeated until the convergence judgment condition is met, the state evaluation is terminated, and the vehicle's initial posture is output according to the evaluation result. The embodiment of the present invention can also obtain an accurate initialization posture when the GPS status is poor in driving scenes such as urban high-rise buildings and trees or tunnels, and then the vehicle will perform path planning or decision control for autonomous driving according to the initialization posture, thereby reducing the positioning error to improve the overall performance of the autonomous driving system and enhance the user experience and satisfaction.

[0051] The vehicle posture initialization method provided by the embodiment of the present invention determines the step size and range of the position data and posture data according to the status information in the satellite positioning data of the vehicle, creates a first particle set, and predicts the posture of each particle in the first particle set according to the inertial navigation data, screens the first particle set according to the predicted particle posture and the perception observation data, obtains the second particle set after the posture is updated, and performs state evaluation according to the second particle set to determine the initialization posture of the vehicle. The present invention sets different particle sets according to different satellite status information, which can ensure that the initial particle set has sufficient coverage to offset the error caused by the poor state of the satellite positioning system, and at the same time, assists in positioning with the help of the front-end sensing device, so as to complete the posture initialization in various driving scenarios and improve the robustness, adaptability and accuracy of the positioning system.

[0052] In this embodiment, a vehicle posture initialization method is provided, which can be used for a controller of a vehicle. Figure 3 is a flow chart of a vehicle posture initialization method according to an embodiment of the present invention. Figure 3 As shown, the process includes the following steps:

[0053] Step S301, obtaining satellite positioning data, inertial navigation data and sensory observation data of the vehicle, where the satellite positioning data includes status information, position data and attitude data.

[0054] Specifically, the above step S301 includes:

[0055] Step S3011, obtain the satellite positioning data and inertial navigation data of the current frame, and determine whether there is perception observation data in the current frame.

[0056] Specifically, in the embodiment of the present invention, due to the different data sampling frequencies of INS and front-end sensing devices, a frame of data may contain only satellite positioning data and inertial navigation data, but no sensing observation data. In order to assist in positioning with the help of the front-end sensing device, the embodiment of the present invention needs to ensure that the data of the current frame contains satellite positioning data, inertial navigation data and sensing observation data. Therefore, after acquiring a frame of data, it is determined whether there is sensing observation data in the current frame data. If not, data collection continues.

[0057] Step S3012: If there is perception observation data, obtain a satellite positioning data set of a preset number of frames before the current frame, and determine the optimal satellite positioning state closest to the current frame in the satellite positioning data set according to the state priority, where the state priorities are from high to low: fixed solution state, floating point solution state, differential positioning state and single point positioning state.

[0058] Specifically, in an embodiment of the present invention, if the current frame contains satellite positioning data, inertial navigation data, and sensory observation data at the same time, a satellite positioning data set of a preset number of frames before the current frame is obtained, such as 500 frames of INS data within 5 seconds before the timestamp corresponding to the current frame, to obtain a 500-frame satellite positioning data set, which is used as an example only and is not limited to this. Each frame of satellite positioning data contains a corresponding state bit. Since the number of GPS satellite searches at different locations may change, the state bits corresponding to different frames may be different. In order to ensure the accuracy of the initialization posture, an embodiment of the present invention determines the optimal satellite positioning state closest to the current frame in the 500-frame satellite positioning data set according to the state priority, wherein the state priorities are from high to low: fixed solution state, floating point solution state, differential positioning state, and single point positioning state. At this time, the position data and attitude data in the satellite positioning data corresponding to the optimal satellite state are the most accurate within 5 seconds.

[0059] Step S3013, taking the data frame corresponding to the optimal satellite positioning state as the starting frame, taking the satellite positioning data corresponding to the starting frame as the initial satellite positioning data, and acquiring the inertial navigation data set from the starting frame to the current frame.

[0060] Specifically, in an embodiment of the present invention, each frame of INS data includes corresponding satellite positioning data and inertial navigation data, the INS data frame corresponding to the optimal satellite positioning state is used as the starting frame, the satellite positioning data corresponding to the starting frame is used as the initial satellite positioning data, and multiple frames of inertial navigation data between the starting frame and the current frame are used as the inertial navigation data set, wherein the starting frame may be the previous frame of the current frame, or may be separated from the current frame by multiple frames, which is specifically determined according to the satellite positioning state.

[0061] Step S302: Determine the step size and range of the position data and the attitude data according to the state information, create a first particle set based on the step size and the range, and perform position and attitude prediction on each particle in the first particle set according to the inertial navigation data. Figure 1 Step S102 of the illustrated embodiment will not be described in detail here.

[0062] Step S303, screening each particle in the first particle set according to the predicted particle position and the sensed observation data, to obtain a second particle set with updated position. Figure 1 Step S103 of the illustrated embodiment will not be described in detail here.

[0063] Step S304: perform state evaluation based on the second particle set, and determine the initial position and posture of the vehicle based on the evaluation result. Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0064] The vehicle posture initialization method provided by the embodiment of the present invention determines the step size and range of the position data and posture data according to the status information in the satellite positioning data of the vehicle, creates a first particle set, and predicts the posture of each particle in the first particle set according to the inertial navigation data, screens the first particle set according to the predicted particle posture and the perception observation data, obtains the second particle set after the posture is updated, and performs state evaluation according to the second particle set to determine the initialization posture of the vehicle. The present invention sets different particle sets according to different satellite status information, which can ensure that the initial particle set has sufficient coverage to offset the error caused by the poor state of the satellite positioning system, and at the same time, assists in positioning with the help of the front-end sensing device, so as to complete the posture initialization in various driving scenarios and improve the robustness, adaptability and accuracy of the positioning system.

[0065] In this embodiment, a vehicle posture initialization method is provided, which can be used for a controller of a vehicle. Figure 4 is a flow chart of a vehicle posture initialization method according to an embodiment of the present invention. Figure 4 As shown, the process includes the following steps:

[0066] Step S401, obtain the vehicle's satellite positioning data, inertial navigation data and sensor observation data, the satellite positioning data includes status information, position data and attitude data. For details, please refer to Figure 3 Step S301 of the illustrated embodiment will not be described in detail here.

[0067] Step S402, determining the step size and range of the position data and the attitude data according to the state information, creating a first particle set based on the step size and the range, and performing position and attitude prediction on each particle in the first particle set according to the inertial navigation data.

[0068] Specifically, the above step S402 includes:

[0069] Step S4021, determining a first range and a first step length of a heading angle, a second range and a second step length of a lateral position, and a third range and a third step length of a longitudinal position according to the status information of the initial satellite positioning data.

[0070] Specifically, in the embodiment of the present invention, the satellite positioning data includes attitude data represented by the heading angle, position data represented by the lateral position and the longitudinal position, and state information represented by the GPS state bit. The attitude data and position data corresponding to the initial satellite positioning data are the most accurate within 5 seconds, but there may be errors, and the worse the GPS state corresponding to the state information, the greater the error. Therefore, the step length and range of the heading angle, lateral position and longitudinal position are determined according to the state information. There are differences in the step length and range corresponding to different state information. Taking the RTK fixed solution and the RTK floating point solution as examples, the step length and range corresponding to the heading angle, lateral position and longitudinal position are shown in the following table, which is only used as an example and is not limited to this:

[0071]

[0072] Step S4022, determining the number of particles of the first particle set and the particle posture of each particle according to the first range, the first step length, the second range, the second step length, the third range and the third step length, to obtain the first particle set.

[0073] Specifically, in the embodiment of the present invention, the heading angle corresponding to the initial satellite positioning data is The horizontal position x and the vertical position y are used as the reference, and the value is taken according to the corresponding step size within the corresponding range. For example, when the status information is RTK fixed solution, the heading angle In the range of [-0.5, 0.5], the value is taken in steps of 0.5, and we get and Therefore, multiple heading angles, multiple lateral positions, and multiple longitudinal positions can be obtained under different state information, and different heading angles, different lateral positions, and different longitudinal positions constitute particles with different postures, thereby obtaining a first particle set containing multiple particles. In the embodiment of the present invention, when the accuracy of the GPS state bit representation is high, the coverage of the first particle set is small and the number of particles is small, and when the accuracy of the state bit representation is low, the coverage of the first particle set is large and the number of particles is large.

[0074] Step S4023, performing normalized weight assignment on each particle in the first particle set to determine the particle weight of each particle.

[0075] Specifically, in the embodiment of the present invention, in order to indicate the accuracy of each particle's position and posture, each particle is assigned a normalized weight, that is, the sum of the weights of each particle is 1. When the first particle set is initially created, the accuracy of each particle's position and posture is not clear, so each particle is assigned the same normalized weight. For example, if the first particle set contains 20 particles, the weight of each particle is 1 / 20, which is only an example and is not limited to this.

[0076] Step S4024, constructing a motion model, and taking the start frame corresponding to the initial satellite positioning data as the prediction frame start point of the motion model.

[0077] Specifically, in the embodiment of the present invention, the posture of each particle in the first particle set is the posture state corresponding to the starting frame, the starting frame is before the current frame, and the posture state accuracy of the satellite positioning data corresponding to the starting frame is the highest. Therefore, a motion model is constructed, and the starting frame is used as the prediction frame starting point of the motion model, and the motion model is used to predict the posture state of all particles in the first particle set in the next frame.

[0078] Step S4025, based on the motion model, predict the particle pose of each particle in the next prediction frame according to the wheel speed, angular velocity and particle pose of each particle in the current prediction frame in the inertial navigation data set in turn, until reaching the current frame.

[0079] Specifically, in an embodiment of the present invention, the posture state of the satellite positioning data corresponding to the starting frame has the highest accuracy within the last 5 seconds, that is, the GPS state of other frames between the starting frame and the current frame is poor, and the positioning accuracy is low. During the vehicle driving process, the inertial navigation data obtained by the inertial measurement unit IMU in the INS system is relatively accurate. Therefore, based on the motion model, the starting frame is used as the starting point of the prediction frame of the motion model, and the posture of each particle in the current prediction frame is predicted in turn according to the inertial navigation data corresponding to the current prediction frame, and the particle posture of each particle in the next prediction frame is obtained until the current frame is reached. Among them, the prediction process based on the motion model is shown in the following formula:

[0080] p next =p cur +R cur ·v·Δt

[0081] R next =R cur ·exp((ω·Δt)∧)

[0082] Among them, p next is the position of the next predicted frame, p cur is the position of the current prediction frame, which is the horizontal position and the vertical position; R next is the pose of the next predicted frame, R cur is the posture of the current prediction frame, where the posture is the heading angle; v is the wheel speed in the inertial navigation data corresponding to the current prediction frame, ω is the angular velocity in the inertial navigation data corresponding to the current prediction frame, and Δt is the time interval between the previous and next frames.

[0083] In some optional implementations, the embodiments of the present invention can use this motion model to infer the poses of all particles in the first particle set at subsequent moments, especially the poses of the current frame. Since the perception observation data itself does not have the pose information of the vehicle, such as the forward-looking camera data comes from an image, or the laser radar data comes from a cluster of point clouds, it is necessary to predict and infer the pose of the vehicle at the time point corresponding to the perception observation data, that is, the position of the current frame, through the above motion model, so as to provide a basis for updating the pose state weights for subsequent observations.

[0084] Step S403, screening each particle in the first particle set according to the predicted particle posture and the sensed observation data to obtain a second particle set with updated posture.

[0085] Specifically, the above step S403 includes:

[0086] Step S4031, when the current prediction frame reaches the current frame, based on the particle posture of each particle corresponding to the current prediction frame, the perception observation data is projected into the preset dynamic map, and the perception observation data is matched with the preset dynamic map data.

[0087] Specifically, in an embodiment of the present invention, semantic recognition is performed on the perceived observation data (forward-looking camera data and lidar data) to obtain semantic information of the perceived observation data, such as lane lines, curbs, stop lines, road surface signs and other information. Then, based on the predicted current frame posture state of each particle, the semantic information is projected into the local coordinate system of the dynamic map data, and the semantic information in the perceived observation data is matched with the semantic information in the dynamic map. At this time, the more accurate the particle posture state, the higher the matching degree of the semantic information in the perceived observation data and the semantic information in the dynamic map. Taking the lane line as an example, if the particle posture is accurate, the perceived lane line and the map lane line are overlapped.

[0088] Step S4032: update the particle weight of each particle according to the matching result, and compare the updated particle weight with the first preset weight threshold.

[0089] Specifically, in the embodiment of the present invention, the matching result represents the accuracy of the particle posture state, wherein the matching result includes the matching error, the number of matching points and the matching rate. The embodiment of the present invention adjusts the weight of each particle according to the matching result. If the matching degree is high, the particle weight is higher, and the matching degree is lower, the particle weight is lower. After the weight of each particle is adjusted according to the matching result, the updated particle weight is compared with the first preset weight threshold to determine whether the posture accuracy of each particle meets the requirements.

[0090] Step S4033: if the updated particle weight is lower than the first preset weight threshold, the particle corresponding to the updated particle weight is deleted from the first particle set to obtain a second particle set.

[0091] Specifically, in an embodiment of the present invention, if the updated particle weight is lower than the first preset weight threshold, such as 0.5, it is proved that the position error of this particle is large, and this particle is deleted from the first particle set. By comparing the updated weight of each particle with the first preset weight threshold, the particles with more accurate position states in the first particle set can be screened out, thereby obtaining the second particle set.

[0092] Step S404, performing state evaluation according to the second particle set, and determining the initialization posture of the vehicle according to the evaluation result.

[0093] Specifically, the above step S404 includes:

[0094] Step S4041, determining the number of particles in the second particle set, and comparing the number of particles with a preset number threshold.

[0095] Specifically, in the embodiment of the present invention, the number of particles in the second particle set is the first-level convergence judgment condition, that is, the number of particles is compared with a preset number threshold to determine whether there are too many particles in the second particle set and the convergence effect is not achieved.

[0096] Step S4042, if the number of particles is lower than the preset number threshold, the particle weight of each particle in the second particle set is compared with the second preset weight threshold, otherwise return to the step of predicting the position and posture of each particle in the first particle set according to the inertial navigation data.

[0097] Specifically, in an embodiment of the present invention, the preset number threshold is set according to actual conditions, for example, it is set to 20. If the number of particles is not less than 20, the iteration is continued based on the second particle set, and the pose prediction and pose update are repeated until the number meets the threshold requirement. If the number of particles in the second particle set is less than 20, the particle weight of each particle is compared with the second preset weight threshold. The second preset weight threshold can be set to be greater than the first preset weight threshold, for example, 0.7.

[0098] Step S4043: If there is at least one particle whose weight is greater than the second preset weight threshold, the particle posture corresponding to the particle with the maximum particle weight in the second particle set is used as the initialization posture.

[0099] Specifically, in an embodiment of the present invention, the particle weight is the second-layer convergence judgment condition. If there is a particle in the second particle set whose particle weight is greater than the second preset weight threshold, the posture of the particle is used as the initialization posture. If the weights of multiple particles are greater than the second preset weight threshold, the posture of the particle with the largest weight is selected as the initialization posture, and the initialization process ends at this time.

[0100] Step S4044: If it does not exist, the similarity of the positions of the particles in the second particle set is determined.

[0101] Specifically, in the embodiment of the present invention, if the weight of any particle in the second particle set is less than the second preset weight threshold, that is, there is no particle with a particularly large weight, the degree of similarity between the particles is used as the third-level convergence judgment condition. Among them, the judgment of the degree of similarity of the posture includes the judgment of the heading angle and the horizontal and vertical positions. By comparing the heading angle and the horizontal and vertical positions of each particle, the heading angle difference and the horizontal and vertical position difference of each particle are obtained.

[0102] Step S4045, if the degree of similarity of the postures meets the preset threshold, the particle posture of the particle corresponding to the maximum particle weight in the second particle set is used as the initialization posture, otherwise, the process returns to the step of predicting the posture of each particle in the first particle set according to the inertial navigation data.

[0103] Specifically, in an embodiment of the present invention, if the degree of similarity between the postures of the particles meets the preset threshold, for example, the maximum heading angle difference and the maximum horizontal and vertical position difference between the particles in the second particle set are both less than the corresponding preset threshold, it means that the postures of the particles in the second particle set are similar, which is equivalent to the same posture, so the posture of the particle corresponding to the maximum particle weight is used as the initialization posture, and the initialization process ends. However, if the judgment of the degree of similarity of the posture is not met, it will continue to iterate based on the second particle set, repeat the posture prediction and posture update, until the above-mentioned convergence judgment condition is met, the state evaluation is completed, and the initialization posture is output.

[0104] The vehicle posture initialization method provided by the embodiment of the present invention determines the step size and range of the position data and posture data according to the status information in the satellite positioning data of the vehicle, creates a first particle set, and predicts the posture of each particle in the first particle set according to the inertial navigation data, screens the first particle set according to the predicted particle posture and the perception observation data, obtains the second particle set after the posture is updated, and performs state evaluation according to the second particle set to determine the initialization posture of the vehicle. The present invention sets different particle sets according to different satellite status information, which can ensure that the initial particle set has sufficient coverage to offset the error caused by the poor state of the satellite positioning system, and at the same time, assists in positioning with the help of the front-end sensing device, so as to complete the posture initialization in various driving scenarios and improve the robustness, adaptability and accuracy of the positioning system.

[0105] In this embodiment, a vehicle posture initialization device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0106] This embodiment provides a vehicle posture initialization device, such as Figure 5 As shown, including:

[0107] The data acquisition module 501 is used to acquire the satellite positioning data, inertial navigation data and perception observation data of the vehicle. The satellite positioning data includes status information, position data and attitude data.

[0108] The posture prediction module 502 is used to determine the step size and range of the position data and the posture data according to the state information, create a first particle set based on the step size and the range, and perform posture prediction on each particle in the first particle set according to the inertial navigation data.

[0109] The particle updating module 503 is used to screen each particle in the first particle set according to the predicted particle posture and the sensed observation data to obtain a second particle set with updated posture.

[0110] The posture determination module 504 is used to perform state evaluation according to the second particle set and determine the initialization posture of the vehicle according to the evaluation result.

[0111] In some optional implementations, the satellite positioning data, the inertial navigation data, and the sensory observation data are pre-time-stamped, and the state information includes a fixed solution state, a floating point solution state, a differential positioning state, and a single point positioning state. The data acquisition module 501 includes:

[0112] The first data acquisition unit is used to acquire satellite positioning data and inertial navigation data of the current frame, and determine whether there is perception observation data in the current frame.

[0113] The state judgment unit is used to obtain a satellite positioning data set of a preset number of frames before the current frame if there is perception observation data, and determine the optimal satellite positioning state closest to the current frame in the satellite positioning data set according to the state priority, where the state priorities are from high to low: fixed solution state, floating point solution state, differential positioning state and single point positioning state.

[0114] The second data acquisition unit is used to use the data frame corresponding to the optimal satellite positioning state as the starting frame, the satellite positioning data corresponding to the starting frame as the initial satellite positioning data, and acquire the inertial navigation data set from the starting frame to the current frame.

[0115] In some optional implementations, the position data includes a lateral position and a longitudinal position, the posture data is a heading angle, and the posture prediction module 502 includes:

[0116] The step range determination unit is used to determine the first range and first step of the heading angle, the second range and second step of the lateral position, and the third range and third step of the longitudinal position according to the state information of the initial satellite positioning data.

[0117] The particle set construction unit is used to determine the number of particles in the first particle set and the particle posture of each particle according to the first range, the first step length, the second range, the second step length, the third range and the third step length to obtain the first particle set.

[0118] The particle weight determination unit is used to assign a normalized weight to each particle in the first particle set to determine the particle weight of each particle.

[0119] In some optional implementations, the posture prediction module 502 further includes:

[0120] The motion model building unit is used to build a motion model, and use the starting frame corresponding to the initial satellite positioning data as the prediction frame starting point of the motion model.

[0121] The particle posture prediction unit is used to predict the particle posture of each particle in the next prediction frame based on the motion model, in turn according to the wheel speed, angular velocity and particle posture of each particle in the current prediction frame in the inertial navigation data set, until the current frame is reached.

[0122] In some optional implementations, the particle updating module 503 includes:

[0123] The matching unit is used to project the sensed observation data into a preset dynamic map based on the particle pose of each particle corresponding to the current prediction frame when the current prediction frame reaches the current frame, and match the sensed observation data with the preset dynamic map data.

[0124] The weight updating unit is used to update the particle weight of each particle according to the matching result, and compare the updated particle weight with the first preset weight threshold.

[0125] The particle screening unit is used to delete the particle corresponding to the updated particle weight from the first particle set if the updated particle weight is lower than the first preset weight threshold, so as to obtain the second particle set.

[0126] In some optional implementations, the posture determination module 504 includes:

[0127] The particle quantity comparison unit is used to determine the particle quantity of the second particle set and compare the particle quantity with a preset quantity threshold.

[0128] The particle weight comparison unit is used to compare the particle weight of each particle in the second particle set with the second preset weight threshold if the number of particles is lower than the preset number threshold, otherwise return to the step of predicting the position and posture of each particle in the first particle set according to the inertial navigation data.

[0129] The first pose determination unit is used to use the particle pose of the particle corresponding to the maximum particle weight in the second particle set as the initialization pose if there is at least one particle weight greater than the second preset weight threshold.

[0130] The unit for judging the similarity of posture is used to judge the similarity of posture of each particle in the second particle set if it does not exist.

[0131] The second posture determination unit is used to use the particle posture of the particle corresponding to the maximum particle weight in the second particle set as the initialization posture if the posture similarity meets the preset threshold, otherwise return to the step of predicting the posture of each particle in the first particle set according to the inertial navigation data.

[0132] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0133] The vehicle posture initialization device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0134] The embodiment of the present invention also provides a computer device having the above Figure 5 The vehicle posture initialization device is shown.

[0135] See also Figure 6 , Figure 6 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 6 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 6 A processor 10 is taken as an example.

[0136] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0137] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0138] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0139] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0140] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0141] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0142] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0143] The embodiment of the present invention further provides a vehicle, comprising: an integrated navigation system, the integrated navigation system comprising: a satellite positioning system and an inertial measurement unit, respectively used to collect satellite positioning data and inertial navigation data; a front-end sensing device, time-synchronized with the integrated navigation system, used to collect sensing observation data; a controller, connected to the integrated navigation system and the front-end sensing device, used to execute the above Figure 1 , Figure 3 or Figure 4The vehicle pose initialization method shown.

[0144] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A vehicle posture initialization method, characterized in that: The method comprises: Acquire satellite positioning data, inertial navigation data and perception observation data of the vehicle, wherein the satellite positioning data includes status information, position data and attitude data; Determining the step size and range of the position data and the posture data according to the state information, creating a first particle set based on the step size and the range, and performing position and posture prediction for each particle in the first particle set according to the inertial navigation data; Screening each particle in the first particle set according to the predicted particle posture and the sensed observation data to obtain a second particle set with updated posture; A state evaluation is performed according to the second particle set, and an initialization posture of the vehicle is determined according to the evaluation result.

2. The method according to claim 1, characterized in that The satellite positioning data, the inertial navigation data and the sensory observation data are pre-time-stamped, and the state information includes a fixed solution state, a floating point solution state, a differential positioning state and a single point positioning state. The obtaining of satellite positioning data, inertial navigation data and perception observation data of the vehicle includes: Acquire the satellite positioning data and the inertial navigation data of the current frame, and determine whether the perception observation data exists in the current frame; If the perception observation data exists, a satellite positioning data set of a preset number of frames before the current frame is obtained, and the optimal satellite positioning state closest to the current frame is determined in the satellite positioning data set according to the state priority, wherein the state priorities are, from high to low, fixed solution state, floating point solution state, differential positioning state, and single point positioning state; The data frame corresponding to the optimal satellite positioning state is used as a starting frame, the satellite positioning data corresponding to the starting frame is used as initial satellite positioning data, and an inertial navigation data set from the starting frame to the current frame is acquired.

3. The method according to claim 2, characterized in that The position data includes a lateral position and a longitudinal position, the posture data is a heading angle, and the step length and range of the position data and the posture data are determined according to the state information, and a first particle set is created based on the step length and the range, including: Determine, according to the state information of the initial satellite positioning data, a first range and a first step length of the heading angle, a second range and a second step length of the lateral position, and a third range and a third step length of the longitudinal position; Determine the number of particles of the first particle set and the particle posture of each particle according to the first range, the first step length, the second range, the second step length, the third range, and the third step length to obtain the first particle set; Normalized weights are assigned to the particles in the first particle set to determine particle weights of the particles.

4. The method according to claim 3, characterized in that The inertial navigation data includes wheel speed and angular velocity, and performing posture prediction on each particle in the first particle set according to the inertial navigation data includes: Constructing a motion model, and taking the starting frame corresponding to the initial satellite positioning data as the prediction frame starting point of the motion model; Based on the motion model, the particle pose of each particle in the next prediction frame is predicted in turn according to the wheel speed, the angular velocity and the particle pose of each particle in the current prediction frame in the inertial navigation data set until the current frame is reached.

5. The method according to claim 4, characterized in that The step of screening each particle in the first particle set according to the predicted particle position and the sensed observation data to obtain a second particle set with updated position and posture includes: When the current prediction frame reaches the current frame, based on the particle posture of each particle corresponding to the current prediction frame, the sensed observation data is projected into a preset dynamic map, and the sensed observation data is matched with the preset dynamic map data; Updating the particle weight of each particle according to the matching result, and comparing the updated particle weight with a first preset weight threshold; If the updated particle weight is lower than the first preset weight threshold, the particle corresponding to the updated particle weight is deleted from the first particle set to obtain the second particle set.

6. The method according to claim 1, characterized in that The step of performing state evaluation according to the second particle set and determining the initialization posture of the vehicle according to the evaluation result includes: Determining the number of particles of the second particle set, and comparing the number of particles with a preset number threshold; If the number of particles is lower than the preset number threshold, the particle weight of each particle in the second particle set is compared with the second preset weight threshold, otherwise the process returns to the step of predicting the position and posture of each particle in the first particle set according to the inertial navigation data; If there is at least one particle whose weight is greater than the second preset weight threshold, taking the particle posture of the particle corresponding to the maximum particle weight in the second particle set as the initialization posture; If it does not exist, then the similarity of the positions of the particles in the second particle set is judged; If the degree of similarity of the postures meets a preset threshold, the particle posture of the particle corresponding to the maximum particle weight in the second particle set is used as the initialization posture, otherwise the process returns to the step of predicting the posture of each particle in the first particle set according to the inertial navigation data.

7. A vehicle posture initialization device, characterized in that: The device comprises: A data acquisition module is used to acquire satellite positioning data, inertial navigation data and perception observation data of the vehicle, wherein the satellite positioning data includes status information, position data and attitude data; a posture prediction module, used for determining a step size and a range of the position data and the posture data according to the state information, creating a first particle set based on the step size and the range, and performing posture prediction on each particle in the first particle set according to the inertial navigation data; A particle updating module, used for screening each particle in the first particle set according to the predicted particle posture and the sensed observation data to obtain a second particle set with updated posture; The posture determination module is used to perform state evaluation according to the second particle set and determine the initialization posture of the vehicle according to the evaluation result.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the vehicle posture initialization method according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the vehicle posture initialization method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises computer instructions for causing a computer to execute the vehicle posture initialization method according to any one of claims 1 to 6.

11. A vehicle, characterized in that: include: An integrated navigation system, the integrated navigation system comprising: a satellite positioning system and an inertial measurement unit, which are used to collect satellite positioning data and inertial navigation data respectively; A front-end sensing device is time-synchronized with the integrated navigation system and is used to collect sensing observation data; A controller is connected to the combined navigation system and the front-end sensing device, and is used to execute the vehicle posture initialization method according to any one of claims 1 to 6.

Images