Indoor pedestrian trajectory prediction method based on multiple sensors

Abstract

The invention provides an indoor pedestrian trajectory prediction method based on multiple sensors, and relates to the technical field of data processing, and the method comprises the steps: measuring the magnetic field intensity of an environment where a pedestrian is located through two magnetometers disposed in front of and behind a safety helmet; according to an accelerometer, a gyroscope and a magnetometer at the front end of the safety helmet, measuring postures of the safety helmet in real time, including a roll angle, a pitch angle and a yaw angle, correcting magnetic field data of a Z axis of the magnetometer, and converting the magnetic field data to a Z axis of a navigation coordinate system; preprocessing the corrected Z-axis magnetic field intensity, including filtering and normalization, so as to obtain preprocessed data; and matching magnetic field waveform data in the processed data by using a dynamic time warping algorithm to obtain a magnetometer matching result. According to the invention, through multi-sensor fusion, accurate tracking and real-time updating of the pedestrian motion trail are realized, and the accuracy of position positioning is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method for predicting indoor pedestrian trajectories based on multiple sensors. Background Art

[0002] In indoor and outdoor environments, accurate prediction of pedestrian trajectories plays a key role in many applications, including but not limited to security monitoring, crowd analysis, and navigation assistance. However, existing pedestrian speed estimation and trajectory prediction technologies encounter various challenges in actual operations, especially when facing complex and changing indoor environments.

[0003] Traditional positioning technologies, such as satellite positioning, perform well in outdoor environments, but their performance will be greatly reduced or even fail to work indoors or in enclosed spaces because the signals are blocked by buildings. This results in services that rely on satellite positioning becoming unstable or completely unavailable in indoor environments.

[0004] In addition, although wireless technologies such as RFID, Wi-Fi positioning and Bluetooth beacons can provide effective positioning information in some scenarios, the reliability of these technologies will be seriously affected in complex environments such as construction sites, where the line of sight is often blocked by buildings, mechanical equipment or other obstacles. Blockage not only causes signal attenuation, but also may cause multipath effects, further reducing positioning accuracy.

[0005] At the same time, cost is also an issue that cannot be ignored. High-precision positioning equipment, such as those using advanced UWB technology, can provide excellent positioning performance, but its cost is high. This cost issue is particularly prominent in large-scale deployments, especially in scenarios with limited budgets. For RFID and Bluetooth beacon technologies, in order to achieve comprehensive coverage and high-precision positioning, a large number of readers or beacons need to be deployed, which also increases the overall cost. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide an indoor pedestrian trajectory prediction method based on multiple sensors, which can provide stable positioning services in complex indoor environments.

[0007] In order to solve the above technical problems, the technical solution of the present invention is as follows: In a first aspect, a method for indoor pedestrian trajectory prediction based on multiple sensors is provided, the method comprising: The magnetic field strength of the pedestrian's environment is measured by two magnetometers configured on the front and back of the helmet; According to the accelerometer, gyroscope and magnetometer at the front end of the helmet, the attitude of the helmet is measured in real time, including roll angle, pitch angle and yaw angle, and the magnetometer is The magnetic field data of the axis is corrected and converted to the navigation coordinate system. Axis-down; preprocessing the corrected magnetic field intensity, including filtering and normalization, to obtain preprocessed data; Using the dynamic time warping algorithm, the magnetic field waveform data in the processed data is matched to obtain the magnetometer matching result; The horizontal distance between two magnetometers passing the same position and the walking speed of the pedestrian are calculated based on the magnetometer matching results; According to the pedestrian's walking speed, yaw angle and initial position information, the pedestrian's real-time position is updated to obtain the pedestrian's motion trajectory.

[0008] Furthermore, a micro motor is used to drive the magnetometer to rotate in three-dimensional space to obtain three-axis magnetometer data, and the measured values ​​of the magnetometer data are compensated by ellipsoid fitting.

[0009] Furthermore, the dynamic time warping algorithm is used to match the magnetic field waveform data in the processed data to obtain the magnetometer matching result, including: Create a two-dimensional matrix with the number of rows equal to the length of magnetic field waveform data sequence A plus 1, and the number of columns equal to the length of magnetic field waveform data sequence B plus 1; Initialize the first row and the first column of the two-dimensional matrix to infinity, and traverse the remaining grids of the two-dimensional matrix to calculate the matching cost between the corresponding data points in the two sequences; Update the value of the grid in the two-dimensional matrix, start from the end point of the two-dimensional matrix, trace back to the starting point, extract the corresponding matching points in sequence A and sequence B, and obtain the magnetometer matching result.

[0010] Furthermore, the attitude of the helmet is measured in real time based on the accelerometer, gyroscope and magnetometer, including roll angle, pitch angle and yaw angle, including: The acceleration changes of the helmet in three axes are measured by an accelerometer, and the angular velocity of the helmet is measured by a gyroscope to predict the rotation of the helmet. The accelerometer, gyroscope, and magnetometer continuously collect data to reflect the helmet's motion status; The data from the accelerometer, gyroscope and magnetometer are combined to obtain the helmet's attitude information, including pitch, roll and heading angle data.

[0011] Furthermore, the calculation formula for pedestrian walking speed is: ;

[0012] in, Indicates the walking speed of pedestrians; Indicates the installation distance between magnetometers; It represents the cosine value of the angle between the horizontal direction and the direction of the line connecting the front and rear magnetometers, that is, the cosine value of the pitch angle; Represents the time difference between two matching waveforms.

[0013] Furthermore, the horizontal distance calculation formula of two magnetometers passing the same position is: ; in, Indicates the horizontal distance between two magnetometers passing the same position

[0014] Furthermore, the calculation formula of the pedestrian's motion trajectory is: ; in, Indicates that the pedestrian is at time point The position vector of Indicates that the pedestrian is at the initial time point The position vector of Indicates that the pedestrian is at time point The instantaneous velocity vector of Indicates the current time point; Indicates the initial time point when the pedestrian starts moving.

[0015] In a second aspect, a multi-sensor-based indoor pedestrian trajectory prediction system includes: The acquisition module is used to measure the magnetic field strength of the pedestrian's environment and the pedestrian's dynamic posture through two magnetometers, accelerometers and gyroscopes configured on the front and back of the helmet. The magnetic field data of the axis is corrected and converted to the navigation coordinate system. Below the axis; after correction The axial magnetic field intensity is preprocessed, including filtering and normalization, to obtain preprocessed data; the magnetic field waveform data in the processed data is matched using a dynamic time warping algorithm to obtain a magnetometer matching result; The processing module is used to calculate the horizontal distance between the two magnetometers passing the same position and the walking speed of the pedestrian according to the magnetometer matching results; update the real-time position of the pedestrian according to the walking speed, yaw angle and initial position information of the pedestrian to obtain the movement trajectory of the pedestrian.

[0016] According to a third aspect, a computing device includes: one or more processors; The storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method described.

[0017] In a fourth aspect, a computer-readable storage medium stores a program, and when the program is executed by a processor, the method described is implemented.

[0018] The above solution of the present invention includes at least the following beneficial effects: By combining the data from magnetometers, accelerometers, and gyroscopes, high-precision pedestrian trajectory prediction can be achieved in indoor environments. The magnetometer is used to sense the ambient magnetic field strength and calculate the yaw angle, while the accelerometer and gyroscope provide attitude information of the pedestrian's pitch and roll angles. After processing, these data can accurately reflect the pedestrian's movement status and position changes, thereby improving positioning accuracy. It is not affected by closed indoor environments or obstructions, because magnetometers, accelerometers, and gyroscopes are all measured based on physical principles and do not rely on external signal sources. Therefore, even when GPS signals are unavailable or wireless signals are interfered with, they can still work stably and provide accurate trajectory prediction for pedestrians.

[0019] It can measure and update the posture and position information of pedestrians in real time. By matching the magnetic field waveform data through the dynamic time warping algorithm and using the accelerometer, gyroscope and magnetometer to measure the helmet posture in real time, the horizontal distance between the two magnetometers passing the same position and the walking speed of the pedestrian can be quickly calculated, and the position trajectory of the pedestrian can be updated in real time. The sensors used are low-cost and do not require large-scale deployment of external equipment or beacons. It is not only suitable for complex environments such as construction sites, but can also be promoted for pedestrian positioning and navigation in indoor places such as shopping malls, airports, and exhibition halls. At the same time, it can also provide accurate position and posture information for technologies such as virtual reality (VR) and augmented reality (AR). BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a flow chart of a method for predicting indoor pedestrian trajectories based on multiple sensors provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of an indoor pedestrian trajectory prediction system based on multiple sensors provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0022] like Figure 1As shown, an embodiment of the present invention proposes a method for predicting indoor pedestrian trajectories based on multiple sensors, and the method comprises the following steps: Step 1, measuring the magnetic field strength of the pedestrian's environment by configuring two magnetometers on the front and back of the helmet; Step 2: Based on the accelerometer, gyroscope and magnetometer at the front end of the helmet, the helmet's posture is measured in real time, including roll angle, pitch angle and yaw angle. The magnetic field data of the axis is corrected and converted to the navigation coordinate system. Below the axis; after correction The axial magnetic field intensity is preprocessed, including filtering and normalization, to obtain preprocessed data; Step 3, using a dynamic time warping algorithm to match the magnetic field waveform data in the processed data to obtain a magnetometer matching result; Step 4, calculating the horizontal distance between the two magnetometers passing the same position and the walking speed of the pedestrian according to the magnetometer matching result; Step 5: Update the real-time position of the pedestrian according to the pedestrian's walking speed, yaw angle and initial position information to obtain the pedestrian's motion trajectory.

[0023] In the embodiment of the present invention, the magnetic field strength of the pedestrian's environment can be measured by the magnetometer configured on the helmet. This feature enables the system to operate stably in a variety of environments, especially those areas that are difficult to cover or where the signal is unstable with traditional positioning technology. The use of the magnetometer enhances the system's adaptability to different environments. By preprocessing the magnetic field strength, including filtering and normalization, the system can eliminate noise and outliers, thereby improving the accuracy and reliability of the data.

[0024] By using accelerometers, gyroscopes and magnetometers to measure the helmet's posture in real time, including roll, pitch and yaw angles, the system can accurately capture the dynamic posture changes of pedestrians. This not only helps to more accurately calculate walking speed and distance, but also provides precise posture input for other applications such as virtual reality or augmented reality.

[0025] Combined with the magnetometer matching results, the horizontal distance between the two magnetometers passing the same position and the pedestrian's walking speed can be accurately calculated. According to the pedestrian's walking speed, yaw angle and initial position information, the system can update the pedestrian's position in real time, thereby generating an accurate motion trajectory.

[0026] In a preferred embodiment of the present invention, the above step 1, measuring the magnetic field strength of the pedestrian's environment by two magnetometers arranged in front and behind the helmet, may include: The magnetometer contains components that can sense changes in the magnetic field, such as Hall effect sensors, magnetoresistive sensors, or fluxgate sensors. When these sensors are placed in a magnetic field, they generate electrical signals proportional to the strength of the magnetic field. The magnetometer converts these electrical signals into analog signals that can be processed and analyzed. These signals represent the strength of the magnetic field components in the three axes (X, Y, Z) and are called three-axis magnetometer data.

[0027] In a preferred embodiment of the present invention, in step 2, the posture of the helmet is measured in real time according to the accelerometer, gyroscope and magnetometer at the front end of the helmet, including the roll angle, pitch angle and yaw angle, and the magnetometer is used to measure the posture of the helmet. The magnetic field data of the axis is corrected and converted to the navigation coordinate system. Below the axis; after correction The axial magnetic field strength is preprocessed, including filtering and normalization, to obtain preprocessed data, which may include: Before preprocessing the magnetic field data and matching the magnetic field waveform data, the device's posture change will cause the magnetometer data to shift and rotate, and this change will directly affect the accuracy of data matching. In order to ensure the reliability of the matching results, first The axis data is converted from the device coordinate system to the navigation coordinate system, and aligned to the z-axis of the navigation coordinate system. The formula is as follows: ;in, Indicated in The z-axis value of the system; , , They are the three-axis data of the magnetometer, represents the roll angle, Indicates the pitch angle. Filter the corrected magnetic field strength data, select a low-pass filter based on the type and characteristics of the noise, and process the magnetic field strength data through the filter. The filter will smooth the data to make it smoother and more stable. The specific formula is: ( 1); among which, Indicates the output value of this filtering; Indicates the sampling rate; represents the cut-off frequency; Indicates the sampling value of this time; ( 1) Represents the last filtered output value. The data is scaled to a specific range through normalization operation. The specific operations are as follows: ;in, represents the normalized data, represents the minimum value in the data set. Indicates the maximum value in the data set; Represents a dataset.

[0028] In another preferred embodiment of the present invention, a micro motor is used to drive a magnetometer to rotate in three-dimensional space, to obtain three-axis magnetometer data, and to compensate the measured value of the magnetometer data by ellipsoidal surface fitting, which may include: The magnetometer is driven by a micro motor to rotate inside the helmet to simulate the spatial position of the worker's head in different postures. Due to the magnetic interference in the environment, the actual measurement data will form an ellipsoidal shape. In order to correct the measurement deviations caused by these interferences, the least squares method is used to fit the magnetometer measurement data to the ellipsoidal surface. The goal of fitting is to find an ellipsoid model that matches the measurement data, so as to obtain the fitting parameters of the ellipsoid. These parameters include the center coordinates of the ellipsoid and the lengths of the three main axes, which describe the shape and position of the ellipsoid. Once these parameters are obtained, the system can compensate the original measurement value to eliminate the influence of interference and improve the measurement accuracy of the magnetometer.

[0029] Standard equations for an axis-aligned ellipsoid: is used to describe the fitting process, where , , The center of the ellipsoid is x, y, z Coordinates on the axis; , , The ellipsoid is x, y, z The radius in the axial direction. This equation is expanded and rearranged to multiply by After that, it is transformed into a linear equation system: . Rewrite the above formula as The form is: , the linear equations are solved by the least squares method to obtain the solution of the ellipsoid fitting parameters. The calculation formula of the least squares method is: , in summary, , ; The coordinates of the center of the ellipsoid are: ; The lengths of the three axes of the ellipsoid are: This automated calibration process not only avoids the inconvenience of workers having to manually shake the helmet for calibration, but also improves the efficiency and accuracy of calibration. In complex construction site environments, this calibration method ensures that the magnetometer can maintain high-precision measurements in various postures, thereby improving the overall performance and ease of use of the smart helmet.

[0030] In a preferred embodiment of the present invention, the above step 3, using a dynamic time warping algorithm to match the magnetic field waveform data in the processed data to obtain a magnetometer matching result, may include: Step 33, creating a two-dimensional matrix, the number of rows is the length of the magnetic field waveform data sequence A plus 1, and the number of columns is the length of the magnetic field waveform data sequence B plus 1; Step 34, initializing the first row and the first column of the two-dimensional matrix to infinity, and traversing the remaining grids of the two-dimensional matrix to calculate the matching cost between the corresponding data points in the two sequences; Step 35, updating the value of the grid in the two-dimensional matrix, starting from the end point of the two-dimensional matrix, backtracking to the starting point, extracting the corresponding matching points in sequence A and sequence B to obtain the magnetometer matching result.

[0031] In an embodiment of the present invention, a two-dimensional matrix is ​​created, the number of rows and columns of which are 1 more than the length of the two magnetic field waveform data sequences A and B. This matrix will be used to store the matching costs between the data points in the sequence. The first row and the first column of the two-dimensional matrix are initialized to infinity. This is to ensure that during the backtracking process, the algorithm will not choose to match from the edge of the matrix, because there are no actual data points corresponding to these positions. Traverse the remaining grids of the two-dimensional matrix, and calculate the matching cost between the corresponding data points in sequence A and sequence B as: ; in, Represents the matching cost between corresponding data points in sequence A and sequence B; Represents data points To data point The Euclidean distance between the two grids is calculated; for each grid in the matrix, the value of the current grid is updated according to the values ​​of the three adjacent grids on the left, top and top left and the matching cost of the current grid. Starting from the lower right corner (end point) of the two-dimensional matrix, trace back to the upper left corner (starting point) along the minimum value path of each grid source. In the process of backtracking, the corresponding matching points in sequence A and sequence B are extracted. These matching points are the matching results between the two magnetic field waveform data sequences.

[0032] The algorithm can process time series data of different lengths and speeds, which makes it highly flexible when processing magnetometer data collected in real environments. By calculating the matching cost between each data point and finding the path with the smallest overall matching cost, the similarity between two magnetic field waveform data sequences can be more accurately identified. The matching points extracted through the backtracking process can clearly show the matching relationship between the two sequences.

[0033] In another preferred embodiment of the present invention, the attitude of the helmet is measured in real time according to the accelerometer, gyroscope and magnetometer, including the roll angle, pitch angle and yaw angle, which may include: Step 61, using an accelerometer to measure the acceleration change of the helmet in three axes, and using a gyroscope to measure the angular velocity of the helmet to predict the rotation movement of the helmet; Step 62, the accelerometer, gyroscope and magnetometer continuously collect data to reflect the motion state of the helmet; Step 63 , combining the data of the accelerometer, the gyroscope and the magnetometer to obtain the helmet posture information, including the data of the pitch angle, the roll angle and the yaw angle.

[0034] In an embodiment of the present invention, the accelerometer is capable of measuring the helmet's X, Y, Z ). These acceleration data include static acceleration caused by the earth's gravity and dynamic acceleration caused by the movement of the helmet. In static or slow motion state, the accelerometer is mainly used to determine the attitude of the helmet relative to the direction of gravity. By analyzing the acceleration values ​​on the three axes, the pitch angle and roll angle of the helmet can be calculated. For example, when the helmet is placed horizontally and stationary, Z The acceleration value on the axis will be close to the acceleration due to gravity g ,and X Axis and Y The values ​​on the axes will be close to zero. When the helmet is tilted, these values ​​will change, so that the tilt angle can be solved. In dynamic motion, accelerometer data can also be used to assist the gyroscope to provide additional attitude information, especially when moving at high speed or turning quickly. The gyroscope is able to measure the angular velocity of the helmet around three axes (X, Y, Z). These angular velocity data directly reflect the rotational movement of the helmet. The main advantage of the gyroscope is that it can quickly respond to changes in the rotation of the helmet and provide angular velocity data in real time. This allows the system to quickly track and update the attitude of the helmet, especially in the case of rapid movement or rotation. By integrating the angular velocity data output by the gyroscope, the change in the rotation angle of the helmet can be calculated, and then the attitude angle (roll angle, pitch angle and yaw angle) can be updated.

[0035] Specifically, the representation method of the helmet posture is the Euler angle. The rotation of one coordinate system relative to another coordinate system can be represented by rotation around three orthogonal axes. These three rotation angles are called a set of Euler angles. The following is the rotation matrix from the n system to the b system: ; in, Represents the direction cosine matrix from the b coordinate system (i.e., the carrier coordinate system) to the n coordinate system (the navigation coordinate system); represents the roll angle; Indicates the pitch angle; Indicates the yaw angle. In static state, the roll angle and pitch angle can be used to obtain the attitude of the carrier using the three-axis observation values ​​of the accelerometer. When the gyroscope and accelerometer are stationary in the navigation coordinate system (n system), the output of the accelerometer is: ;in, Represents the acceleration vector in the navigation coordinate system; Represents the acceleration due to gravity. At this time, the output of the accelerometer in the body coordinate system (b system) is: ;in, Represents the acceleration vector in the body coordinate system; , , express x, y, z Acceleration components on the axes; have to: ; Since the rotation axis of the yaw angle is parallel to the direction of gravity, the yaw angle cannot be obtained using an accelerometer, and the data from the magnetometer is needed to calculate it.

[0036] In the northern hemisphere, the horizontal component of the Earth's magnetic field vector points to magnetic north, and the vertical component points to the ground. Because magnetic north and geographic north do not coincide, the angle between the line connecting any point and magnetic north and the line connecting that point and geographic north is Let it be the magnetic declination. is the yaw angle of the carrier relative to magnetic north, , , are the magnetic east, magnetic north and celestial components, then: ; True heading angle The magnetic declination angle minus the magnetic north yaw angle gives: . Get the Euler angle from the angular velocity using the quaternion. Use the first-order Runge-Kutta method to find the quaternion ; in, , , In the carrier coordinate system X axis, Y axis, Z Axis angular velocity; , , , is a quaternion, the final roll angle, pitch angle and yaw angle calculation formula is .

[0037] The complementary filtering algorithm is used to fuse the data of the accelerometer and gyroscope to obtain a more accurate and stable attitude estimate. The weights of the accelerometer and gyroscope data are dynamically adjusted according to the motion state of the helmet. Specifically: In this case, the accelerometer can accurately measure the acceleration of gravity to determine the attitude of the helmet. Therefore, the weight of the accelerometer data in the fusion process can be increased. At the same time, since the gyroscope may be affected by the integration error under static conditions, the weight of its data can be reduced accordingly. When the helmet is in a fast motion or rotation state, the gyroscope data becomes more important because it can quickly capture the change of angular velocity. Therefore, in a high dynamic state, the weight of the gyroscope data should be increased. The data of the accelerometer may be disturbed by large dynamic acceleration under high dynamic conditions, so its weight should be reduced accordingly. The gyroscope has a high dynamic response capability and can quickly capture the change of angular velocity. However, long-term integration may cause error accumulation. Therefore, it is necessary to set the integration time window reasonably in the fusion process to reduce the impact of the integration error. The dynamic response of the accelerometer is relatively slow, but it provides more accurate data under static or low dynamic conditions. During fusion, the weight and utilization of accelerometer data should be dynamically adjusted according to the actual motion state of the helmet.

[0038] By combining the data of the accelerometer and the gyroscope, the advantages of both can be fully utilized and their deficiencies can be complemented. The accelerometer provides accurate attitude information under static or low-dynamic conditions, while the gyroscope can respond quickly to dynamic changes. The accuracy and stability of attitude measurement are improved. Since both the accelerometer and the gyroscope collect data in real time, this method can quickly capture the attitude changes of the helmet, providing strong support for real-time monitoring and safety warning. It is not only suitable for attitude measurement in static or slow motion states, but also can maintain high measurement accuracy under high dynamic conditions. By monitoring the attitude of the helmet in real time, the wearer's abnormal movements or dangerous situations can be detected in time, so that corresponding safety measures can be taken.

[0039] In a preferred embodiment of the present invention, the walking speed calculation formula of the pedestrian is: ;in, Indicates the walking speed of pedestrians; Indicates the installation distance between magnetometers; It represents the cosine value of the angle between the horizontal direction and the direction of the line connecting the front and rear magnetometers, that is, the cosine value of the pitch angle; Represents the time difference between two matching waveforms.

[0040] In the embodiment of the present invention, two magnetometers are installed at the front and rear ends of the pedestrian's helmet, and the distance between the two magnetometers is This mounting distance is fixed and remains constant during the measurement. As the pedestrian walks, the magnetometer records changes in the magnetic field that correspond to the movement of the head. Represents the angle between the walking direction and the direction of the magnetometer connection. By measuring or estimating this angle and using the cosine function To ensure that the horizontal distance calculated is the same as the distance between the two magnetometers passing through the same position. Represents the time difference between two matching waveforms. It is determined by analyzing the signal waveforms recorded by the magnetometer. When a pedestrian walks, the movement of the head causes the magnetometer to record periodic signal changes. By identifying the matching waveforms in these signals and measuring the time difference between them, we can get Finally, use the formula Calculate the walking speed of pedestrians.

[0041] By introducing the installation distance between the magnetometers , the cosine value of the angle between the horizontal direction and the direction of the line connecting the front and rear magnetometers , and the time difference between the two matching waveforms , which can more accurately calculate the walking speed of pedestrians. This formula is applicable to different walking directions and speeds because it takes into account the angle between the horizontal direction and the direction of the line connecting the front and rear magnetometers, that is, the change in the pitch angle.

[0042] In a preferred embodiment of the present invention, the horizontal distance calculation formula of two magnetometers passing through the same position is: ;in, Indicates the horizontal distance between two magnetometers passing the same position

[0043] In an embodiment of the present invention, Represents the straight-line distance between two measurement points (e.g., two magnetometers). This distance is fixed and serves as a reference for the measurement. The pitch angle of the helmet is a key parameter. In practical applications, the pitch angle of the helmet is determined by a sensor. It represents the angle between the horizontal direction and the direction of the line connecting the front and rear magnetometers. It is used to calculate the horizontal distance between two magnetometers passing the same position. The cosine function can be used according to the angle Adjust straight line distance , thus obtaining the effective distance in the horizontal direction. The cosine of the angle Multiplying them together, the result is the horizontal distance between the two magnetometers passing the same position. .

[0044] By introducing the angle The cosine value is used to adjust the straight-line distance , so that the horizontal distance between two magnetometers passing the same position can be calculated more accurately. The formula can quickly calculate the horizontal distance between two magnetometers passing the same position, simplifying the calculation process and improving efficiency. This formula is applicable to different walking paths and angles, as long as the straight-line distance can be determined and angle , we can accurately calculate the horizontal distance between the two magnetometers passing the same location. By continuously collecting and analyzing the horizontal distance between the two magnetometers passing the same location, we can get the speed of the pedestrians, which can help decision makers better understand the behavior patterns of pedestrians and make decisions.

[0045] In a preferred embodiment of the present invention, the formula for calculating the motion trajectory of a pedestrian is: ; in, Indicates that the pedestrian is at time point The position vector of Indicates that the pedestrian is at the initial time point The position vector of Indicates that the pedestrian is at time point The instantaneous velocity vector of Indicates the current time point; Indicates the initial time point when the pedestrian starts moving.

[0046] In an embodiment of the present invention, Indicates that the pedestrian is at the initial time point This is the starting point for calculating the motion trajectory and provides the specific position of the pedestrian when he starts moving. Indicates that pedestrians at any time The instantaneous velocity vector of the pedestrian. This velocity changes with time and reflects the speed and direction of the pedestrian at different time points. The integral part of the formula From the initial time point To the current time The total distance and direction of the pedestrian's movement during this period. By integrating the instantaneous velocity, the pedestrian's position change during this period can be calculated. Adding it to the integral result, we get the pedestrian's current time point The position vector This position vector describes the overall motion trajectory of the pedestrian from the starting point to the current time point. As time goes by, the position of the pedestrian will continue to change. By continuously updating the instantaneous velocity And recalculate the integral to track the movement trajectory of pedestrians in real time.

[0047] This formula can accurately calculate the position of pedestrians at any point in time, thereby accurately tracking the movement trajectory of pedestrians. The position of pedestrians can be calculated in real time, so the current position information of pedestrians can be updated instantly. By analyzing the past movement trajectory data of pedestrians, their future movement trends and possible locations can be predicted.

[0048] like Figure 2 As shown, an embodiment of the present invention further provides a multi-sensor based indoor pedestrian trajectory prediction system 20, comprising: The acquisition module 21 is used to measure the magnetic field strength of the pedestrian's environment and the pedestrian's dynamic posture through two magnetometers, accelerometers and gyroscopes configured on the front and back of the helmet; the posture of the helmet is measured in real time according to the accelerometer, gyroscope and magnetometer, including the roll angle, pitch angle and yaw angle, and the magnetometer is used to measure the magnetic field strength of the pedestrian's environment and the pedestrian's dynamic posture; The magnetic field data of the axis is corrected and converted to the navigation coordinate system. Below the axis; after correction The axial magnetic field intensity is preprocessed, including filtering and normalization, to obtain preprocessed data; the magnetic field waveform data in the processed data is matched using a dynamic time warping algorithm to obtain a magnetometer matching result; The processing module 22 is used to calculate the horizontal distance between the two magnetometers passing the same position and the walking speed of the pedestrian according to the magnetometer matching result; update the real-time position of the pedestrian according to the walking speed, yaw angle and initial position information of the pedestrian to obtain the movement trajectory of the pedestrian.

[0049] It should be noted that the system is a system corresponding to the above method, and all implementation methods in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0050] The embodiment of the present invention further provides a computing device, comprising: a processor, a memory storing a computer program, wherein when the computer program is executed by the processor, the method described above is executed. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0051] The embodiment of the present invention also provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the method described above. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0052] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A multi-sensor based indoor pedestrian trajectory prediction method, characterized in that: The method comprises: The magnetic field strength of the pedestrian's environment is measured by two magnetometers configured on the front and back of the helmet; According to the accelerometer, gyroscope and magnetometer at the front end of the helmet, the attitude of the helmet is measured in real time, including roll angle, pitch angle and yaw angle, and the magnetometer is The magnetic field data of the axis is corrected and converted to the navigation coordinate system. axis; after correction The axial magnetic field intensity is preprocessed, including filtering and normalization, to obtain preprocessed data; Using the dynamic time warping algorithm, the magnetic field waveform data in the processed data is matched to obtain the magnetometer matching result; The horizontal distance between two magnetometers passing the same position and the walking speed of the pedestrian are calculated based on the magnetometer matching results; According to the pedestrian's walking speed, yaw angle and initial position information, the pedestrian's real-time position is updated to obtain the pedestrian's motion trajectory.

2. The method for indoor pedestrian trajectory prediction based on multiple sensors according to claim 1, characterized in that: A micro motor is used to drive the magnetometer to rotate in three-dimensional space to obtain three-axis magnetometer data, and the measured values ​​of the magnetometer data are compensated by ellipsoid fitting.

3. The method for indoor pedestrian trajectory prediction based on multiple sensors according to claim 2 is characterized in that: The dynamic time warping algorithm is used to match the magnetic field waveform data in the processed data to obtain the magnetometer matching results, including: Create a two-dimensional matrix with the number of rows equal to the length of magnetic field waveform data sequence A plus 1, and the number of columns equal to the length of magnetic field waveform data sequence B plus 1; Initialize the first row and the first column of the two-dimensional matrix to infinity, and traverse the remaining grids of the two-dimensional matrix to calculate the matching cost between the corresponding data points in the two sequences; Update the value of the grid in the two-dimensional matrix, start from the end point of the two-dimensional matrix, trace back to the starting point, extract the corresponding matching points in sequence A and sequence B, and obtain the magnetometer matching result.

4. The method for indoor pedestrian trajectory prediction based on multiple sensors according to claim 3 is characterized in that: According to the accelerometer, gyroscope and magnetometer at the front end of the helmet, the attitude of the helmet is measured in real time, including roll angle, pitch angle and yaw angle, including: The acceleration changes of the helmet in three axes are measured by an accelerometer, and the angular velocity of the helmet is measured by a gyroscope to predict the rotation of the helmet. The accelerometer, gyroscope, and magnetometer continuously collect data to reflect the helmet's motion status; The data from the accelerometer, gyroscope and magnetometer are combined to obtain the helmet's attitude information, including pitch, roll and heading angle data.

5. The method for indoor pedestrian trajectory prediction based on multiple sensors according to claim 4 is characterized in that: The formula for calculating the walking speed of pedestrians is: ； in, Indicates the walking speed of pedestrians; Indicates the installation distance between magnetometers; It represents the cosine value of the angle between the horizontal direction and the direction of the line connecting the front and rear magnetometers, that is, the cosine value of the pitch angle; Represents the time difference between two matching waveforms.

6. The method for indoor pedestrian trajectory prediction based on multiple sensors according to claim 5, characterized in that: The calculation formula for the horizontal distance between two magnetometers passing the same position is: ； in, Indicates the horizontal distance between two magnetometers passing the same position.

7. The method for indoor pedestrian trajectory prediction based on multiple sensors according to claim 6, characterized in that: The calculation formula of the pedestrian's motion trajectory is: ； in, Indicates that the pedestrian is at time point The position vector of Indicates that the pedestrian is at the initial time point The position vector of Indicates that the pedestrian is at time point The instantaneous velocity vector of Indicates the current time point; Indicates the initial time point when the pedestrian starts moving.

8. A multi-sensor based indoor pedestrian trajectory prediction system, characterized in that: Applied to the method according to any one of claims 1 to 7, comprising: The acquisition module is used to measure the magnetic field strength of the pedestrian's environment through two magnetometers configured in front and behind the helmet; the accelerometer, gyroscope and magnetometer at the front of the helmet are used to measure the helmet's posture in real time, including roll angle, pitch angle and yaw angle, and the magnetometer is used to measure the magnetic field strength of the pedestrian's environment ... The magnetic field data of the axis is corrected and converted to the navigation coordinate system. axis; after correction The axial magnetic field intensity is preprocessed, including filtering and normalization, to obtain preprocessed data; the magnetic field waveform data in the processed data is matched using a dynamic time warping algorithm to obtain a magnetometer matching result; The processing module is used to calculate the horizontal distance between the two magnetometers passing the same position and the walking speed of the pedestrian according to the magnetometer matching results; update the real-time position of the pedestrian according to the walking speed, yaw angle and initial position information of the pedestrian to obtain the movement trajectory of the pedestrian.

9. A computing device, characterized in that include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, which, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images