Multi-scene three-dimensional trajectory reconstruction method and device based on pedestrian dead reckoning

Abstract

The invention discloses a multi-scene three-dimensional trajectory reconstruction method and device based on pedestrian dead reckoning, and relates to the technical field of motion tracking. The method comprises the steps that acceleration, angular velocity and magnetic field data are obtained through a chest sensor, track data are obtained through an improved step length estimation and track reconstruction algorithm after Kalman filtering, the motion direction is obtained through complementary filtering, the step length is obtained through acceleration double integration, and the motion direction and the track data are obtained. The Z-axis acceleration data is compensated by using the projection of gravity in the chest sensor coordinate system; quaternion data corresponding to the thighs and the shanks are obtained through two sensors arranged on the lower limbs on the two sides, then rotation of the shanks on the two sides relative to the thighs is calculated, and posture data of the lower limbs are obtained; and fusing the trajectory data with the posture data of the lower limbs to obtain three-dimensional trajectory data. The method achieves the construction of a complete motion model, accords with the real kinematics law of a human body, effectively reduces the problems of integral errors and drifting, and guarantees the positioning robustness and precision.

Description

Technical Field

[0001] The present invention relates to the field of motion tracking technology, and in particular to a multi-scene three-dimensional trajectory reconstruction method and device based on pedestrian dead reckoning. Background Art

[0002] Dead reckoning is the process of inferring a future destination, without any external reference, by using a known starting position combined with estimated speed and direction over time. Extending this concept to human motion, particularly walking or running, is pedestrian dead reckoning (PDR), widely used in indoor and outdoor navigation and virtual character motion control. This technology encompasses multiple technical areas, including sensor data fusion, filtering algorithms, distributed sensor network construction, and virtual reality.

[0003] Existing pedestrian dead reckoning uses multiple sensors (such as IMU, barometer, GPS or camera, etc.) to work together, and achieves positioning in multi-story or complex scenarios through map matching or environmental information assistance. The hardware complexity is high and it is greatly affected by environmental interference. Virtual character control usually requires wearing motion capture equipment (such as optical marker systems), which is relatively expensive, and the system deployment and data synchronization are difficult. At the same time, it often focuses on indoor applications and lacks adaptability to outdoor environments. At the same time, existing technologies install IMU sensors on the feet, collect acceleration and angular velocity data, and rely on zero-speed updates (ZUPT) for step length estimation and direction correction. It has poor adaptability to complex motion states (such as going up and down stairs and rapid turns) and is easily affected by the loss of short-term zero-speed states, resulting in reduced positioning accuracy. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a multi-scenario three-dimensional trajectory reconstruction method and device based on pedestrian dead reckoning, which uses the lower limb IMU nodes to collect joint posture information and fuses it with the chest sensor data to realize the construction of a complete motion model that conforms to the real kinematic laws of the human body, effectively reduces integration error and drift problems, and ensures positioning robustness and accuracy.

[0005] In a first aspect, the present invention provides a multi-scene three-dimensional trajectory reconstruction method based on pedestrian dead reckoning, comprising:

[0006] Trajectory data acquisition process: Acceleration, angular velocity, and magnetic field data are acquired through chest sensors. After Kalman filtering, trajectory data is obtained using an improved step-length estimation and trajectory reconstruction algorithm. The motion direction is obtained through complementary filtering, the step length is obtained through double integration of acceleration, and the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system.

[0007] Posture data acquisition process: The quaternion data corresponding to the thigh and calf are obtained through two sensors set up on both lower limbs. The rotation of each calf relative to the thigh is then calculated to obtain the posture data of the lower limb;

[0008] Data fusion process: The trajectory data is fused with the lower limb posture data to obtain three-dimensional trajectory data.

[0009] Furthermore, during the trajectory data acquisition process, the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system, specifically:

[0010]

[0011] Among them, a dynamic is the Z-axis acceleration after dynamic compensation, a raw is the Z-axis acceleration output by the chest sensor, g is gravity, is the real-time pitch angle.

[0012] Furthermore, in the process of acquiring the posture data, the quaternion obtained by the sensor attached to one side of the thigh is Q a =(w,x,y,z), the quaternion obtained by the calf sensor on the same side is Q b =(w',x',y',z'), then the rotation of the calf relative to the thigh is Q rel Calculated by the following formula:

[0013]

[0014] in, is the quaternion Q a The inverse, Represents quaternion multiplication.

[0015] Furthermore, the method also includes: dynamically adjusting the parameters of the complementary filter and the Kalman filter, specifically: calculating the acceleration modulus in real time, judging the pedestrian's motion state as stationary, slow walking or fast movement based on comparison with a preset threshold, and then adjusting the fusion weights of the high-frequency data and low-frequency data of the complementary filter according to the judgment result of the motion state, as well as adjusting the process noise covariance and measurement noise covariance parameters in the Kalman filter.

[0016] Furthermore, the fused three-dimensional trajectory data is transmitted to the Blender platform to drive the virtual human body model.

[0017] In a second aspect, the present invention provides a multi-scene three-dimensional trajectory reconstruction device based on pedestrian dead reckoning, characterized by comprising:

[0018] The trajectory data acquisition module is used to obtain acceleration, angular velocity, and magnetic field data from the chest sensor. After Kalman filtering, the trajectory data is obtained using an improved step-length estimation and trajectory reconstruction algorithm. The motion direction is obtained through complementary filtering, the step length is obtained through double integration of acceleration, and the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system.

[0019] The posture data acquisition module is used to obtain the quaternion data corresponding to the thigh and calf through two sensors set on both lower limbs, and then calculate the rotation of each calf relative to the thigh to obtain the posture data of the lower limb;

[0020] The data fusion module is used to fuse the trajectory data with the posture data of the lower limbs to obtain three-dimensional trajectory data.

[0021] Furthermore, in the trajectory data acquisition module, the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system, specifically:

[0022]

[0023] Among them, a dynamic is the Z-axis acceleration after dynamic compensation, a raw is the Z-axis acceleration output by the chest sensor, g is gravity, is the real-time pitch angle.

[0024] Furthermore, in the posture data acquisition module, the quaternion obtained by the sensor attached to one side of the thigh is Q a =(w,x,y,z), the quaternion obtained by the calf sensor on the same side is Q b =(w',x',y',z'), then the rotation of the calf relative to the thigh is Q rel Calculated by the following formula:

[0025]

[0026] in, is the quaternion Q a The inverse, Represents quaternion multiplication.

[0027] Furthermore, the trajectory data acquisition module is also used to dynamically adjust the parameters of the complementary filter and the Kalman filter. Specifically, it calculates the acceleration modulus in real time, judges the pedestrian's motion state as stationary, slow walking or fast movement based on the comparison with the preset threshold, and then adjusts the fusion weight of the high-frequency data and low-frequency data of the complementary filter according to the judgment result of the motion state, as well as adjusts the process noise covariance and measurement noise covariance parameters in the Kalman filter.

[0028] Furthermore, the data fusion process is also used to transmit the fused three-dimensional trajectory data to the Blender platform to drive the virtual human body model.

[0029] The technical solutions provided in the embodiments of the present invention have at least the following technical effects:

[0030] 1. Simplified hardware configuration: Only a single sensor (chest) and a small number of IMU nodes in the lower limbs are required to build a distributed sensor network. No additional complex sensors (such as barometers) are required, which greatly reduces system cost and deployment difficulty.

[0031] 2. Strong adaptability, enabling navigation both indoors and outdoors: Through improved step-length estimation and trajectory reconstruction algorithms, it achieves continuous motion data acquisition without relying on zero-speed state. At the same time, it performs layered suppression of noise and attitude drift to obtain high-precision step-length and direction information. It is suitable for a variety of motion environments such as flat ground, stairs, and outdoor scenes, achieving stable navigation.

[0032] 3. High-precision real-time virtual control: The fused motion data is transmitted to Blender in real time to drive the virtual human model, achieving natural and smooth virtual character control and enhancing the interactive experience.

[0033] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0035] Figure 1 Schematic diagram of the framework of the system of the present invention;

[0036] Figure 2 This is a flowchart of the method in Example 1 of the present invention;

[0037] Figure 3 Schematic diagram of the data processing process in the first embodiment of the present invention;

[0038] Figure 4 This is a schematic diagram of the structure of the device in Example 2 of the present invention. DETAILED DESCRIPTION

[0039] The embodiments of the present invention provide a multi-scenario three-dimensional trajectory reconstruction method and device based on pedestrian dead reckoning. The method uses the lower limb IMU nodes to collect joint posture information and fuses it with the chest sensor data to achieve the construction of a complete motion model that conforms to the real kinematic laws of the human body, effectively reduces integration errors and drift problems, and ensures positioning robustness and accuracy.

[0040] The technical solution in the embodiment of the present invention has the following general ideas:

[0041] A single sensor is mounted on the chest, and a step-length estimation model based on double acceleration integration is employed to achieve 3D trajectory reconstruction without relying on ZUPT. An improved Kalman filter and complementary filter collaborative architecture suppress sensor noise and attitude drift in different motion modes (such as flat ground and ascending and descending stairs), ensuring high-precision positioning. A distributed sensor network is constructed, and data fusion between the chest sensor and lower-limb IMU nodes (for lower-limb posture control) drives a virtual human model in Blender, enabling real-time motion control of virtual characters. This makes the system suitable for both indoor navigation and outdoor applications, improving overall environmental adaptability.

[0042] Before introducing the specific embodiments, the system framework corresponding to the embodiment method of the present application is first introduced. Figure 1 As shown, the system includes hardware and software parts:

[0043] (1) Hardware:

[0044] Chest single sensor module: A single IMU sensor with built-in high-precision accelerometer, gyroscope and magnetometer, fixed to the human chest to realize motion data collection.

[0045] Lower limb IMU node: Two IMU sensors are deployed on each side of the lower limbs (one for each leg) to capture lower limb posture information in real time

[0046] (2) Software:

[0047] Communication module: The sensor uses wireless transmission to receive data in real time on a computer through the SDK. The IMU sensor is used to collect acceleration, angular velocity, magnetic field, and quaternion.

[0048] Data preprocessing: Filter and reduce noise on sensor data, eliminating the need for zero-speed detection and enabling the acquisition of continuous motion data.

[0049] Example 1

[0050] This embodiment provides a multi-scene three-dimensional trajectory reconstruction method based on pedestrian dead reckoning, such as Figure 2 Shown, including:

[0051] S1. Trajectory data acquisition process: Acceleration, angular velocity, and magnetic field data are acquired through the chest sensor. After Kalman filtering, trajectory data is obtained using an improved step-length estimation and trajectory reconstruction algorithm. The motion direction is obtained through complementary filtering, the step length is obtained through double integration of acceleration, and the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system.

[0052] In traditional PDR algorithms, the integral calculation of the Z axis (vertical direction) typically uses a fixed gravity compensation value, which is prone to cumulative error and drift. This invention improves on the traditional PDR algorithm by utilizing a double-integration method for acceleration combined with a Kalman filter and complementary filter collaborative architecture. The Z axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system. This achieves layered suppression of sensor noise and attitude drift errors, thereby obtaining high-precision step length and direction information. This ensures positioning robustness and accuracy in a variety of motion scenarios (flat ground, climbing stairs, indoor and outdoor environments), enabling indoor and outdoor navigation.

[0053] The specific data processing process is as follows Figure 3 In this embodiment's PDR algorithm, the acceleration in the filtered IMU data is used to detect gait (determining whether a step is present or absent). When a step is detected, the filtered acceleration, angular velocity, and magnetic field data are passed through a complementary filter to output the direction of human motion. Simultaneously, the step length is calculated using a double integration of acceleration. Based on this step length and direction information, the human motion trajectory in three-dimensional space can be reconstructed in real time.

[0054] In this embodiment, the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system, specifically:

[0055]

[0056] Among them, a dynaic is the Z-axis acceleration after dynamic compensation, a raw is the Z-axis acceleration output by the chest sensor, g is gravity, is the real-time pitch angle (calculated by complementary filtering).

[0057] In calculating the Z-axis displacement, this method uses a dynamic gravity compensation algorithm based on quaternion attitude, which can accurately calculate the projection of gravity in the current sensor coordinate system, and use this information to compensate the Z-axis acceleration data. The dynamically compensated acceleration is used for integral calculation, thereby achieving accurate estimation of the Z-axis displacement and effectively reducing integration error and drift problems.

[0058] S2. Posture data acquisition process: The quaternion data corresponding to the thigh and calf are obtained through two sensors set on both lower limbs. Then, the rotation of each calf relative to the thigh is calculated to obtain the posture data of the lower limb.

[0059] Different from the existing technology that directly uses the data obtained by the sensor, in order to make the rotation of the lower limb calf and thigh conform to the principles of human kinematics, the present invention designs a quaternion transformation. According to the law of human movement, the movement of the calf changes based on the movement of the thigh. Therefore, when controlling the movement of the calf, it is necessary to calculate the relative rotation of the calf relative to the thigh. Taking the right leg as an example, let the quaternion obtained by the sensor attached to the right thigh be Q a =(w,x,y,z), the quaternion obtained by the right calf sensor is Q b =(w',x',y',z'), then the rotation of the calf relative to the thigh is Q rel Calculated by the following formula:

[0060]

[0061] because,

[0062] Will Substituting into formula (2) and (3), we get:

[0063]

[0064] in, is the quaternion Q a The inverse, Represents quaternion multiplication.

[0065] Through the above calculation, Q a Assign to the right thigh bone, Q rel By assigning values ​​to the right calf bone, we can control the movement of the right leg, thereby accurately and stably controlling the movement posture of the right leg to conform to the real kinematic laws of the human body. The left leg is handled in the same way as the right leg.

[0066] S3, Data Fusion Process: Fusing the trajectory data with the lower limb posture data to obtain 3D trajectory data. Steps S1 and S2 are performed simultaneously. During data fusion, the trajectory data and posture data are time-aligned and then fused.

[0067] In a preferred implementation, a dynamic parameter adjustment method is proposed in which a complementary filter and a Kalman filter work together. First, the modulus of the acceleration is calculated in real time.

[0068]

[0069] where a x is the x-axis acceleration, a y is the y-axis acceleration, a z is the z-axis acceleration.

[0070] The modulus of the acceleration is then compared with a set threshold to identify the different motion states of the human body, including stillness, slow walking, and fast movement.

[0071] Then, during the complementary filter data fusion process, the fusion weights of the high-frequency data (gyroscope) and the low-frequency data (accelerometer) are adjusted in real time according to the different motion states. The optimal weight combination for each motion state is determined based on experimental data. Similarly, based on the different motion states identified, the process noise covariance and measurement noise covariance parameters in the Kalman filter are determined for each corresponding state. The specific values ​​of these parameters are obtained based on statistical analysis of experimental data.

[0072] Through the above dynamic adjustment strategy, the filtering system can adaptively optimize the filtering parameters according to the real-time changing motion conditions, thereby achieving rapid response under different motion states, effectively improving the system's attitude estimation accuracy and significantly reducing the drift error, thereby enhancing the robustness and practicality of the filtering algorithm.

[0073] Existing technologies mostly focus on positioning and trajectory reconstruction, and fail to directly achieve real-time linkage with virtual platforms (such as Blender), affecting the naturalness of the virtual character's movement and interactive experience. Therefore, in one possible implementation method, the three-dimensional trajectory data obtained by fusing the method of this embodiment is transmitted to the Blender platform to drive the virtual human model, thereby realizing the natural motion control of real-time virtual characters and broadening the application of navigation technology in virtual reality and augmented reality. Using a distributed sensor network, the trajectory data of the chest sensor (to control bone movement) is fused with the posture data captured by the lower limb IMU (to control bone rotation), and the data format is JSON to build a complete motion model; using the TCP protocol, the fused data is transmitted to the Blender platform to drive the virtual human model to achieve natural motion control and interaction.

[0074] Based on the same inventive concept, this application also provides a device corresponding to the method in Example 1, see Example 2 for details.

[0075] Example 2

[0076] In this embodiment, a multi-scene three-dimensional trajectory reconstruction device based on pedestrian dead reckoning is provided. Figure 4 Shown, including:

[0077] The trajectory data acquisition module is used to obtain acceleration, angular velocity, and magnetic field data from the chest sensor. After Kalman filtering, the trajectory data is obtained using an improved step-length estimation and trajectory reconstruction algorithm. The motion direction is obtained through complementary filtering, the step length is obtained through double integration of acceleration, and the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system.

[0078] The posture data acquisition module is used to obtain the quaternion data corresponding to the thigh and calf through two sensors set on both lower limbs, and then calculate the rotation of each calf relative to the thigh to obtain the posture data of the lower limb;

[0079] The data fusion module is used to fuse the trajectory data with the posture data of the lower limbs to obtain three-dimensional trajectory data.

[0080] Preferably, in the trajectory data acquisition module, the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system, specifically:

[0081]

[0082] Among them, a dynamic is the Z-axis acceleration after dynamic compensation, a raw is the Z-axis acceleration output by the chest sensor, g is gravity, is the real-time pitch angle.

[0083] Specifically, in the posture data acquisition module, the quaternion obtained by the sensor attached to one side of the thigh is Q a =(w,x,y,z), the quaternion obtained by the calf sensor on the same side is Q b =(w',x',y',z'), then the rotation of the calf relative to the thigh is Q rel Calculated by the following formula:

[0084]

[0085] in, is the quaternion Q a The inverse, Represents quaternion multiplication.

[0086] Preferably, the trajectory data acquisition module is also used to dynamically adjust the parameters of the complementary filter and the Kalman filter, specifically: calculate the acceleration modulus in real time, judge the pedestrian's motion state as stationary, slow walking or fast movement based on the comparison with the preset threshold, and then adjust the fusion weight of the high-frequency data and low-frequency data of the complementary filter according to the judgment result of the motion state, as well as adjust the process noise covariance and measurement noise covariance parameters in the Kalman filter.

[0087] In a possible implementation, the data fusion process is further used to transmit the fused three-dimensional trajectory data to a Blender platform to drive the virtual human body model.

[0088] Since the device described in the second embodiment of the present invention is used to implement the method of the first embodiment of the present invention, those skilled in the art will be able to understand the specific structure and variations of the device based on the method described in the first embodiment of the present invention, and therefore will not be described in detail here. All devices used in the method of the first embodiment of the present invention fall within the scope of protection of the present invention.

[0089] The present invention only requires a single sensor (chest) in conjunction with a small number of IMU nodes in the lower limbs, without the need for additional complex sensors (such as barometers) to build a distributed sensor network. It utilizes the trajectory data of the chest sensor (to control bone movement) and the data fusion of the lower limb IMU nodes (for lower limb posture control) to significantly reduce system cost and deployment difficulty. Through improved step length estimation and trajectory reconstruction algorithms, it is not dependent on the zero-speed state, and the acquisition of continuous motion data is achieved. At the same time, noise and posture drift are suppressed in layers to obtain high-precision step length and direction information. It is suitable for a variety of motion environment scenarios such as flat ground, stairs and outdoors. It has strong adaptability and can achieve stable navigation both indoors and outdoors. The fused motion data is transmitted to Blender in real time to drive the virtual human model, which can achieve high-precision and natural and smooth virtual character control and enhance the interactive experience.

[0090] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0091] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0092] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0093] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0094] Although the specific embodiments of the present invention are described above, those skilled in the art should understand that the specific embodiments described are merely illustrative and are not intended to limit the scope of the present invention. Equivalent modifications and changes made by those skilled in the art in accordance with the spirit of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A multi-scene 3D trajectory reconstruction method based on pedestrian dead reckoning, characterized in that: include: Trajectory data acquisition process: Acceleration, angular velocity, and magnetic field data are acquired through chest sensors. After Kalman filtering, trajectory data is obtained using an improved step-length estimation and trajectory reconstruction algorithm. The motion direction is obtained through complementary filtering, the step length is obtained through double integration of acceleration, and the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system. Posture data acquisition process: The quaternion data corresponding to the thigh and calf are obtained through two sensors set up on both lower limbs. The rotation of each calf relative to the thigh is then calculated to obtain the posture data of the lower limb; Data fusion process: The trajectory data is fused with the lower limb posture data to obtain three-dimensional trajectory data.

2. The method according to claim 1, wherein: During the trajectory data acquisition process, the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system, specifically: Among them, a dynamic is the Z-axis acceleration after dynamic compensation, a raw is the Z-axis acceleration output by the chest sensor, g is gravity, is the real-time pitch angle.

3. The method according to claim 1, wherein: In the process of acquiring the posture data, the quaternion obtained by the sensor attached to one side of the thigh is Q a =(w,x,y,z), the quaternion obtained by the calf sensor on the same side is Q b =(w',x',y',z'), then the rotation of the calf relative to the thigh is Q rel Calculated by the following formula: in, is the quaternion Q a The inverse, Represents quaternion multiplication.

4. The method according to claim 1, wherein The method also includes: dynamically adjusting the parameters of the complementary filter and the Kalman filter, specifically: calculating the acceleration modulus in real time, judging the pedestrian's motion state as stationary, slow walking or fast movement based on comparison with a preset threshold, and then adjusting the fusion weights of the high-frequency data and low-frequency data of the complementary filter according to the motion state judgment result, as well as adjusting the process noise covariance and measurement noise covariance parameters in the Kalman filter.

5. The method according to claim 1, wherein: The fused 3D trajectory data is transmitted to the Blender platform to drive the virtual human body model.

6. A multi-scene three-dimensional trajectory reconstruction device based on pedestrian dead reckoning, characterized in that: include: The trajectory data acquisition module is used to obtain acceleration, angular velocity, and magnetic field data from the chest sensor. After Kalman filtering, the trajectory data is obtained using an improved step-length estimation and trajectory reconstruction algorithm. The motion direction is obtained through complementary filtering, the step length is obtained through double integration of acceleration, and the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system. The posture data acquisition module is used to obtain the quaternion data corresponding to the thigh and calf through two sensors set on both lower limbs, and then calculate the rotation of each calf relative to the thigh to obtain the posture data of the lower limb; The data fusion module is used to fuse the trajectory data with the posture data of the lower limbs to obtain three-dimensional trajectory data.

7. The device according to claim 6, characterized in that: In the trajectory data acquisition module, the Z-axis acceleration data is compensated using the projection of gravity in the chest sensor coordinate system, specifically: Among them, a dynamic is the Z-axis acceleration after dynamic compensation, a raw is the Z-axis acceleration output by the chest sensor, g is gravity, is the real-time pitch angle.

8. The device according to claim 6, characterized in that: In the posture data acquisition module, the quaternion obtained by the sensor attached to one side of the thigh is Q a =(w,x,y,z), the quaternion obtained by the calf sensor on the same side is Q b =(w',x',y',z'), then the rotation of the calf relative to the thigh is Q rel Calculated by the following formula: in, is the quaternion Q a The inverse, Represents quaternion multiplication.

9. The device according to claim 6, characterized in that The trajectory data acquisition module is also used to dynamically adjust the parameters of the complementary filter and the Kalman filter. Specifically, it calculates the acceleration modulus in real time, determines the pedestrian's motion state as stationary, slow walking, or fast movement based on a comparison with a preset threshold, and then adjusts the fusion weights of the high-frequency data and low-frequency data of the complementary filter based on the motion state judgment result, as well as the process noise covariance and measurement noise covariance parameters in the Kalman filter.

10. The device according to claim 6, characterized in that: The data fusion process is also used to transmit the fused three-dimensional trajectory data to the Blender platform to drive the virtual human body model.

Images