Wearable device mutual positioning method and system based on approach-departure principle

Abstract

The present application relates to wearable device mutual search positioning technical field, specifically wearable device mutual search positioning method and system based on approach and retreat principle, including setting base station utilizes multi antenna array to obtain angle and distance information, time slot positioning is carried out simultaneously;Utilize device posture information to construct shielding model, identify line-of-sight communication and signal shielding blind area;Adopt dynamic sector confidence gating strategy to adjust Kalman filter noise weight, combine geometric constraint to solve relative pose vector and generate guide instruction;In near field verification stage, transmit coded vibration sequence, calculate time domain wave packet cross correlation to confirm physical contact, realize automatic exit search.The present application effectively suppresses the positioning drift caused by shielding through shielding model and dynamic gating strategy, improves the guidance robustness in complex environment;Utilize coded vibration and cross correlation verification mechanism, accurately distinguish partition wall misjudgment and real contact, realize intelligent closed loop of approach and guidance, contact and retreat.

Description

Technical Field

[0001] This invention relates to the field of wearable device mutual location technology, specifically to a wearable device mutual location method and system based on the principle of entry and exit. Background Technology

[0002] With the expansion of the application of radio direction finding and ranging technology in the field of miniaturized mobile terminals, target search and positioning technology based on radio signal characteristic parameters has become the core function of wearable device interaction. Existing short-range mutual search schemes mainly use Bluetooth or ultra-wideband radio waves as measurement carriers. By calculating observed physical quantities such as signal received strength indication, angle of arrival or time of flight, the relative distance and azimuth vector of the target source are estimated to achieve guided approach to the lost terminal.

[0003] However, existing solutions primarily rely on a single radio frequency (RF) ranging threshold as the basis for target acquisition, which faces core technical problems such as ranging ambiguity and lack of physical state awareness in practical applications. In complex indoor electromagnetic environments, multipath effects and non-line-of-sight propagation cause severe amplitude fading and time delay drift in radio signals, resulting in a nonlinear deviation between the estimated distance based on the signal propagation model and the actual physical distance. This distortion in the mapping between the RF measurement domain and the physical space domain directly leads to false alarms and logical failures in the positioning system. When the target device is blocked by walls or located in an inaccessible, concealed space, although the signal observation values ​​show that the near-field ranging threshold has been met, the user cannot actually reach the entity. If the system determines that the target has been reached based solely on this RF parameter and automatically terminates the guidance, the search task will be unexpectedly interrupted before completion.

[0004] To address this, a method and system for mutual location tracking of wearable devices based on the principles of entry and exit are proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for mutual positioning of wearable devices based on the principle of entry and exit, so as to achieve mutual positioning of wearable devices through the principle of entry and exit.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The base station is set as a fixed anchor point, and a multi-antenna array is used to obtain the angle information of the first wearable device and the second wearable device. The distance information is estimated by channel detection, and time-slotted positioning is performed at the same time.

[0008] Real-time monitoring of the posture change data of the first wearable device is used to construct a dynamic human shadow model to identify the line-of-sight communication sector and human occlusion blind zone of the first wearable device relative to the base station; a dynamic sector confidence gating strategy is adopted to reduce the measurement noise weight when in the human occlusion blind zone; the relative pose vector of the first wearable device pointing to the second wearable device is calculated by combining geometric constraints and extended Kalman filtering; and a guidance command is generated and fed back to the first wearable device.

[0009] When the relative distance is less than a preset contact threshold, near-field verification is initiated. The first wearable device is controlled to emit an coded vibration sequence, and environmental vibration data sensed by the second wearable device is acquired. The time-domain wave packet cross-correlation between the environmental vibration data and the coded vibration sequence is calculated. If the time-domain wave packet cross-correlation satisfies the resonance matching condition, it is determined that the two are in the same physical medium environment, and the search mode is automatically exited.

[0010] Preferably, the process of acquiring the angle information and distance information involves the base station periodically broadcasting a synchronization beacon frame containing a time slot allocation table, scheduling the first wearable device and the second wearable device to send probe reference signals in different orthogonal time slots; the base station uses a multi-antenna array to receive the probe reference signals, analyzes the channel state information of each subcarrier, and constructs a spatial spectrum covariance matrix;

[0011] The covariance matrix is ​​decomposed using a multi-signal classification algorithm to filter out non-line-of-sight reflection path components and extract angle information; the phase slope of the detection reference signal between different subcarriers is analyzed, and the distance information is calculated based on the multi-carrier phase ranging principle.

[0012] Preferably, the human body shadow dynamic model is established by creating a human body coordinate system with the center of the torso of the wearer of the first wearable device as the origin, and defining the human torso as a rigid cylindrical model with a physical radius.

[0013] Based on the posture change data, the spatial coordinates and antenna normal vector of the first wearable device in the human body coordinate system are calculated, and a line-of-sight path vector connecting the spatial coordinates and the base station coordinates is constructed. The geometric perpendicular distance from the center of the torso to the line-of-sight path vector is calculated, and the included angle between the antenna normal vector and the line-of-sight path vector is calculated. When the geometric perpendicular distance is less than the physical radius, causing the line-of-sight path vector to cut the rigid cylindrical model, it is determined to be the human body occlusion blind zone. And when the included angle is an obtuse angle, causing the antenna normal vector to deviate from the line-of-sight path vector, it is determined to be the human body occlusion blind zone.

[0014] The direction that is not identified as the blind spot of the human body is determined as the line-of-sight communication sector.

[0015] Preferably, the dynamic sector confidence gating strategy involves constructing an adaptive observation noise covariance matrix for the extended Kalman filter and setting a baseline noise parameter.

[0016] In response to the current direction being determined to be the human body occlusion blind zone, a noise expansion coefficient is applied to the reference noise parameters to generate a high-damped covariance matrix and suppress Kalman gain;

[0017] In response to the current direction being identified as the line-of-sight communication sector, the reference noise parameters are maintained, and the state update is biased towards the observations of the measurement update step; the geometric constraint is to introduce a maximum motion rate limit during the filtering process to eliminate abnormal pose abrupt values ​​that exceed the limits of human kinematics.

[0018] Preferably, the relative pose vector calculation method is to obtain the position coordinates of the first wearable device and the position coordinates of the second wearable device in the base station coordinate system;

[0019] Calculate the difference vector in the base station coordinate system, where the difference vector represents the straight-line distance and direction between two points; obtain the heading angle of the first wearable device and construct a rotation matrix;

[0020] The difference vector is transformed from the base station coordinate system to the local body coordinate system of the first wearable device using the rotation matrix to obtain the relative pose vector; the relative pose vector includes the yaw angle deviation of the target relative to the current wearer's orientation and the relative distance.

[0021] Preferably, the generated guidance instruction is configured as the first guidance instruction to drive the linear motor inside the first wearable device, which is mapped to tactile coded vibrations in different directions according to the yaw angle deviation, and mapped to changes in vibration frequency according to the relative distance;

[0022] The second guidance command is transmitted to the TV terminal via a wireless local area network. The TV terminal draws a radar scan map on the graphical user interface according to the command. The relative position of the second wearable device is indicated by directional arrows, and the approach degree is displayed intuitively by dynamically shortening distance bars or color gradient halos. Status prompt text pops up in real time.

[0023] Preferably, the time-domain wave packet cross-correlation is configured as a linear frequency modulated signal with a frequency that scans linearly over time, and the linear frequency modulated signal covers the resonant bandwidth of the haptic actuator;

[0024] A sliding time window is used to extract the high-frequency acceleration component from the environmental vibration data. The high-frequency acceleration component is then subjected to a discrete-time sliding dot product operation with the linear frequency modulated signal to generate a cross-correlation function curve.

[0025] Extract the maximum peak energy and main lobe width from the cross-correlation function curve; the resonance matching condition is that the maximum peak energy is higher than the preset energy coupling threshold, and the peak-to-sidelobe ratio of the main peak to the maximum sidelobe of the cross-correlation function curve is higher than the preset waveform fidelity threshold.

[0026] Wearable device mutual location system based on entry and exit principles includes:

[0027] Radio frequency sensing module: The base station is set as a fixed anchor point and a multi-antenna array is used to obtain the angle information of the first wearable device and the second wearable device. The distance information is estimated by channel detection and time-slotted positioning is performed at the same time.

[0028] Posture modeling module: Real-time monitoring of posture change data of the first wearable device, constructing a dynamic human shadow model to identify the line-of-sight communication sector of the first wearable device relative to the base station and the blind spot of human occlusion;

[0029] Pose fusion module: adopts dynamic sector confidence gating strategy to reduce measurement noise weight when in the blind zone of the human body, and combines geometric constraints and extended Kalman filter to calculate the relative pose vector of the first wearable device pointing to the second wearable device, and generates guidance instructions to be fed back to the first wearable device.

[0030] Near-field verification module: When the relative distance is less than a preset contact threshold, near-field verification is entered, the first wearable device is controlled to emit an coded vibration sequence, and the environmental vibration data sensed by the second wearable device is acquired; the time-domain wave packet cross-correlation between the environmental vibration data and the coded vibration sequence is calculated; if the time-domain wave packet cross-correlation satisfies the resonance matching condition, it is determined that the two are in the same physical medium environment, and the search mode is automatically exited.

[0031] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0032] 1. This invention accurately identifies human occlusion blind spots in communication links by real-time monitoring of the posture change data of wearable devices and constructing a rigid cylindrical model of the human body, combined with the calculation of geometric vertical distance and antenna normal vector. On this basis, a dynamic sector confidence gating strategy is used to adaptively adjust the observation noise covariance matrix of the extended Kalman filter, and a noise expansion coefficient is applied to the measurement data under occlusion to suppress the Kalman gain. This effectively solves the problem of non-line-of-sight signal drift and positioning result jump caused by dynamic human occlusion, and significantly improves the robustness and smoothness of relative pose calculation in complex indoor environments.

[0033] 2. This invention introduces a linear frequency modulated coded vibration sequence with pulse compression characteristics during the near-field verification stage, uses a sliding time window to extract high-frequency acceleration components and calculates their time-domain wave packet cross-correlation with the transmitted sequence; further, by verifying whether the maximum peak energy and peak-sidelobe ratio of the cross-correlation function meet the stringent resonance matching conditions, it can accurately distinguish between airborne acoustic wave coupling and rigid physical medium transmission between devices, thereby achieving intelligent and seamless exit upon contact while eliminating false alarms caused by misjudgment of walls, effectively reducing the ineffective power consumption of wearable devices.

[0034] 3. This invention acquires channel state information and constructs a spatial spectrum covariance matrix through a base station multi-antenna array. It applies a multi-signal classification algorithm and a multi-carrier phase ranging principle to filter multipath interference and obtain high-precision angle and distance information. Combined with coordinate system transformation technology, the differential vector in the global coordinate system of the base station is mapped in real time to the relative pose vector in the local body coordinate system of the first wearable device. This allows guidance commands to be directly converted into directional tactile feedback or a visual radar scan map that conforms to the user's tactile senses, greatly reducing the user's search cognitive load and improving interaction efficiency. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of a wearable device mutual location method based on the principle of entry and exit.

[0036] Figure 2 This is a schematic diagram of the anti-occlusion positioning process based on human shadow model and dynamic gating of the present invention;

[0037] Figure 3 This is a schematic diagram of the near-field physical contact verification process based on linear frequency modulation vibration sequence of the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] Example 1:

[0040] Please see Figure 1 This invention provides a method for mutual location tracking of wearable devices based on the principle of entry and exit, and the technical solution is as follows:

[0041] The base station is set as a fixed anchor point, and a multi-antenna array is used to obtain the angle information of the first wearable device and the second wearable device. The distance information is estimated by channel detection, and time-slotted positioning is performed at the same time.

[0042] Real-time monitoring of the posture change data of the first wearable device is used to construct a dynamic human shadow model to identify the line-of-sight communication sector and human occlusion blind zone of the first wearable device relative to the base station; a dynamic sector confidence gating strategy is adopted to reduce the measurement noise weight when in the human occlusion blind zone; the relative pose vector of the first wearable device pointing to the second wearable device is calculated by combining geometric constraints and extended Kalman filtering; and a guidance command is generated and fed back to the first wearable device.

[0043] When the relative distance is less than a preset contact threshold, near-field verification is initiated. The first wearable device is controlled to emit an coded vibration sequence, and environmental vibration data sensed by the second wearable device is acquired. The time-domain wave packet cross-correlation between the environmental vibration data and the coded vibration sequence is calculated. If the time-domain wave packet cross-correlation satisfies the resonance matching condition, it is determined that the two are in the same physical medium environment, and the search mode is automatically exited.

[0044] The process of acquiring the angle and distance information involves the base station periodically broadcasting a synchronization beacon frame containing a time slot allocation table, scheduling the first wearable device and the second wearable device to send probe reference signals in different orthogonal time slots; the base station uses a multi-antenna array to receive the probe reference signals, analyzes the channel state information of each subcarrier, and constructs a spatial spectrum covariance matrix.

[0045] The covariance matrix is ​​decomposed using a multi-signal classification algorithm to filter out non-line-of-sight reflection path components and extract angle information; the phase slope of the detection reference signal between different subcarriers is analyzed, and the distance information is calculated based on the multi-carrier phase ranging principle.

[0046] The base station is configured with a multi-antenna array of known geometric arrangement. In this embodiment, a uniform circular array or a uniform linear array is used as the receiving end to capture spatial signals.

[0047] The time-slotted positioning process begins with the base station broadcasting a synchronization beacon frame containing a time slot allocation table, scheduling the first wearable device and the second wearable device to synchronize their times; the first wearable device and the second wearable device parse the allocation table, and switch from receive mode to transmit mode in non-overlapping orthogonal time slots respectively, sending physical layer detection reference signals covering effective subcarriers;

[0048] The detection reference signal is received using a multi-antenna array. The received signal is then subjected to cyclic prefix removal and fast Fourier transform to obtain frequency domain data. By removing the modulation component of the transmitted signal using a local reference sequence, the complex channel response of each subcarrier on each antenna element is analyzed, and the responses of all elements at the same time are combined into a snapshot data vector. The spatial spectrum covariance matrix is ​​constructed by calculating the product of multiple consecutive snapshot data vectors and their conjugate transposes and taking the statistical average.

[0049] The base station uses a uniform circular array with a physical radius of 0.2m, and its operating center frequency is set to 5.8GHz, corresponding to a wavelength of approximately 51mm. The physical spacing between adjacent antenna array elements is approximately 2.6cm. When executing the multi-signal classification algorithm, the system constructs a covariance matrix by collecting 100 snapshot data samples and performs a fine scan search in 0.5° increments within an omnidirectional angle range of -180° to 180°. To accurately eliminate multipath interference, a signal strength judgment threshold is set, retaining only angle values ​​with an intensity greater than 0.5 times the maximum spectral peak intensity. Furthermore, the delay data of the channel impulse response is combined to select the angle with the smallest delay as the effective angle of the line-of-sight direct path.

[0050] The specific steps for extracting angle information using the multi-signal classification algorithm include: performing eigenvalue decomposition on the spatial spectrum covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors; dividing the eigenvector space into a signal subspace corresponding to large eigenvalues ​​and a noise subspace corresponding to small eigenvalues ​​based on the differences in the magnitude of the eigenvalues; constructing a spatial spectrum function using the orthogonality between the noise subspace and the signal direction vector; scanning the spatial spectrum function within a preset angle search range (in this embodiment, [-90°, 90°]); searching for the extreme peaks of the spatial spectrum function; eliminating non-line-of-sight false peaks caused by multipath reflections based on the first-to-reach path principle; and determining the angles corresponding to the retained line-of-sight path peaks as the target angle information.

[0051] Extract the channel state information on the line-of-sight path determined by the above steps, and arrange the phase values ​​of different subcarriers in descending order of frequency; perform phase unwinding processing for the periodic phase jump phenomenon to restore the continuous linear change characteristics of the phase; use the least squares method to linearly fit the unwound phase value with the subcarrier frequency index, calculate the slope of the phase frequency line, and use the product of the slope and the speed of light to deduce the physical distance.

[0052] The time-slotted scheduling mechanism effectively avoids signal collisions and co-channel interference between multiple devices; combined with the multi-signal classification algorithm and the first-to-reach path principle, it accurately filters out non-line-of-sight multipath interference, significantly improving the angle estimation accuracy in complex environments; at the same time, based on multi-carrier phase slope and least squares fitting, it achieves high-precision distance sensing with strong noise resistance, ensuring the accuracy and stability of positioning results.

[0053] See Figure 2 The human body shadow dynamic model is established by creating a human body coordinate system with the center of the torso of the wearer of the first wearable device as the origin, and defining the human torso as a rigid cylindrical model with a physical radius.

[0054] Based on the posture change data, the spatial coordinates and antenna normal vector of the first wearable device in the human body coordinate system are calculated, and a line-of-sight path vector connecting the spatial coordinates and the base station coordinates is constructed. The geometric perpendicular distance from the center of the torso to the line-of-sight path vector is calculated, and the included angle between the antenna normal vector and the line-of-sight path vector is calculated. When the geometric perpendicular distance is less than the physical radius, causing the line-of-sight path vector to cut the rigid cylindrical model, it is determined to be the human body occlusion blind zone. And when the included angle is an obtuse angle, causing the antenna normal vector to deviate from the line-of-sight path vector, it is determined to be the human body occlusion blind zone.

[0055] The direction that is not identified as the blind spot of the human body is determined as the line-of-sight communication sector.

[0056] Specifically, the establishment of the human body shadow dynamic model is based on human biomechanical characteristics and geometric simplification principles. A human coordinate system is established with the torso center of the wearer of the first wearable device as the origin. The vertical axis of the torso center when the wearer is standing is defined as the vertical axis, the direction pointing directly forward from the torso center of the wearer is defined as the forward axis, and the lateral direction perpendicular to the vertical axis and the forward axis is defined as the lateral axis. In this coordinate system, the human torso is abstractly defined as a rigid cylindrical model with a fixed physical radius. The physical radius is preset according to the wearer's body shape parameters or set using general anthropometric average values. The height range of the rigid cylindrical model covers the area from the wearer's shoulders to the hips, forming a virtual cylindrical spatial obstacle.

[0057] By constructing a simplified rigid cylindrical model of the human body, this invention abstracts the complex human body shape into regular geometric obstacles, significantly reducing computational complexity to meet the low power consumption requirements of wearable devices, providing a precise geometric benchmark for identifying signal occlusion, and effectively avoiding positioning drift caused by dynamic occlusion of the torso.

[0058] The process of real-time monitoring of the attitude change data of the first wearable device and calculating spatial parameters is as follows: The first wearable device has a built-in inertial measurement unit that collects raw sensor data, including accelerometer, gyroscope, and magnetometer data, in real time. Through an attitude calculation algorithm, the raw sensor data is converted into attitude change data in quaternion or Euler angle form. Combined with pre-calibrated wearing position parameters, a rotation matrix transformation is used to map the state of the first wearable device in its own carrier coordinate system to the human body coordinate system. The spatial coordinate position of the first wearable device in the human body coordinate system and the normal vector of the antenna surface of the first wearable device are calculated in real time. The antenna normal vector represents the main lobe radiation direction of the antenna at the current moment.

[0059] The relative position information of the base station with respect to the first wearable device is obtained. A straight line connecting the current spatial coordinates of the first wearable device and the coordinates of the base station is constructed in the human body coordinate system. This straight line is the line-of-sight path vector. Subsequently, on the horizontal projection plane, the vertical distance from the center of the torso (as the origin) to the straight line containing the line-of-sight path vector is calculated. This is the geometric vertical distance. Using the point-to-line distance calculation method, the numerical value of this geometric vertical distance is obtained through analytical geometric operations, representing the degree of proximity between the communication link between the base station and the first wearable device and the center of the human torso.

[0060] The calculated geometric perpendicular distance is compared with the physical radius of the rigid cylindrical model. When the calculated geometric perpendicular distance is less than the physical radius, it means that the line-of-sight path vector passes through the rigid cylindrical model representing the human torso in geometric space, i.e., the communication link is blocked by the wearer's body. In this case, the current state is determined to be a blind spot due to human body obstruction. Under these circumstances, the wireless signal cannot directly propagate to the base station through a straight line and must rely on reflection or diffraction, resulting in poor channel quality.

[0061] The included angle between the antenna normal vector and the line-of-sight path vector is calculated. The cosine of the angle between the two vectors is obtained through vector dot product operation, and then the angle value is derived. When the included angle is an obtuse angle, that is, the angle value is greater than 90 degrees, it indicates that the main radiation direction of the antenna of the first wearable device is away from the direction of the base station, and the back lobe or side lobe of the antenna is facing the base station, resulting in a significant decrease in antenna gain. At this time, even if there is no physical torso obstruction, the current state is still determined to be in the human body obstruction blind zone.

[0062] Based on the above determination results, the line-of-sight communication sector is determined. During real-time monitoring, the direction is determined as the line-of-sight communication sector only when two conditions are met simultaneously: the geometric vertical distance is greater than or equal to the physical radius, so that the line-of-sight path vector does not cut the rigid cylindrical model, and the included angle is an acute angle or a right angle, so that the antenna normal vector roughly points to the direction of the base station.

[0063] By constructing a simplified rigid cylindrical model of the human body shadow, the algorithm complexity is significantly reduced, making it suitable for the low power consumption requirements of wearable devices. By utilizing a dual decision mechanism of geometric vertical distance and antenna normal angle, physical obstruction and antenna back-facing blind spots are accurately identified, effectively eliminating measurement noise caused by non-line-of-sight propagation and low gain. By actively avoiding interference at the source, ensuring that only high-confidence line-of-sight data is used, the accuracy and robustness of mutual search and positioning in dynamic scenes are greatly improved.

[0064] The dynamic sector confidence gating strategy involves constructing an adaptive observation noise covariance matrix for the extended Kalman filter and setting a baseline noise parameter.

[0065] In response to the current direction being determined to be the human body occlusion blind zone, a noise expansion coefficient is applied to the reference noise parameters to generate a high-damped covariance matrix and suppress Kalman gain;

[0066] In response to the current direction being identified as the line-of-sight communication sector, the reference noise parameters are maintained, and the state update is biased towards the observations of the measurement update step; the geometric constraint is to introduce a maximum motion rate limit during the filtering process to eliminate abnormal pose abrupt values ​​that exceed the limits of human kinematics.

[0067] The dynamic sector confidence gating strategy establishes an adaptive observation noise covariance matrix based on an extended Kalman filter architecture. During the initialization phase, a set of reference noise parameters is pre-set for the observation data such as angle and distance information obtained by the base station. The reference noise parameters are determined based on the variance values ​​determined by the signal statistical error characteristics in an unobstructed ideal radio environment. The reference noise parameters are assigned to the diagonal elements of the observation noise covariance matrix as the default operating state of the filter.

[0068] Regarding the specific operating parameters of the extended Kalman filter, the sampling time interval is set to 10ms, and the noise weight is dynamically adjusted according to the communication status. When in the line-of-sight communication sector, the standard deviation of the angle observation noise is set to 3°, and the standard deviation of the distance observation noise is set to 0.32m. Once the direction is determined to be a blind spot due to human occlusion, the noise expansion coefficient is immediately set to 5, that is, the variance of the observation noise is amplified to 5 times the baseline value, thereby reducing the weight of the data in the filter. In addition, in order to eliminate abnormal jumps caused by multipath effects, human kinematic constraints are introduced, and the maximum movement speed is set to 5m / s. If the calculated instantaneous speed exceeds this value, the predicted value is forced to replace the current observation value.

[0069] In response to the determination result of the aforementioned human shadow dynamic model, when it is confirmed that the direction of the first wearable device relative to the base station is in the human occlusion blind zone, the noise expansion mechanism is immediately activated; the specific process is to select a preset noise expansion coefficient greater than one, multiply the noise expansion coefficient by the aforementioned reference noise parameter, thereby significantly increasing the value of the observed noise covariance matrix and generating a high-damped covariance matrix.

[0070] In the iterative calculation of the extended Kalman filter, the high-damped covariance matrix directly leads to a significant reduction in the calculated Kalman gain. Therefore, when performing posterior state estimation, the filter greatly suppresses the weight of the observation residual at the current moment, that is, it ignores the measurement data that is severely affected by occlusion interference at this moment, and retains and relies more on the predicted value derived from the state at the previous moment, thereby smoothing the transition signal fluctuations and preventing drastic changes in the positioning trajectory;

[0071] When the current direction is determined to be the line-of-sight communication sector, indicating that the current communication link quality is good and the measurement data is reliable, the noise inflation mechanism is revoked, and the reference noise parameters are directly used to construct the observation noise covariance matrix. In this state, the observation noise covariance matrix is ​​smaller, resulting in a relatively larger calculated Kalman gain. This causes the extended Kalman filter's state update process to focus on the real-time observations in the trust measurement update step, thereby enabling sensitive tracking of the actual pose changes of the first wearable device and ensuring high positioning accuracy and low latency.

[0072] The geometric constraints are implemented in parallel. A maximum movement speed limit threshold (set to 5 s / m in this embodiment) is pre-defined based on human biomechanical characteristics, representing the wearer's maximum achievable movement speed in the physical world. Within each time step of the filter, the change between the current pose and the previous pose is calculated, and the instantaneous velocity is deduced. If this instantaneous velocity exceeds the maximum movement speed limit, the current calculation result is determined to be an aberration in pose that violates physical laws. Such aberrations are either eliminated or subjected to forced amplitude limiting and smoothing, effectively filtering out long-distance drift points occasionally caused by multipath effects.

[0073] By employing a dynamic sector confidence gating strategy, the observation weights of the extended Kalman filter are adaptively adjusted to effectively suppress signal fluctuations in the blind zone of human occlusion and ensure high-sensitivity tracking in the line-of-sight sector, achieving an intelligent balance between anti-interference and high precision. Combined with geometric constraints based on the limits of human motion, multipath drift anomalies are eliminated from the physical level, significantly improving the smoothness and reliability of the positioning trajectory of wearable devices in complex dynamic environments.

[0074] The relative pose vector calculation method is to obtain the position coordinates of the first wearable device and the position coordinates of the second wearable device in the base station coordinate system;

[0075] Calculate the difference vector in the base station coordinate system, where the difference vector represents the straight-line distance and direction between two points; obtain the heading angle of the first wearable device and construct a rotation matrix;

[0076] The difference vector is transformed from the base station coordinate system to the local body coordinate system of the first wearable device using the rotation matrix to obtain the relative pose vector; the relative pose vector includes the yaw angle deviation of the target relative to the current wearer's orientation and the relative distance.

[0077] From the final state estimate output of the extended Kalman filter, the smoothed position coordinates of the first wearable device in the base station coordinate system and the position coordinates of the second wearable device in the base station coordinate system are extracted. The base station coordinate system is a pre-constructed global absolute coordinate system with the base station as the origin, and the position coordinates include horizontal and vertical coordinate components on the horizontal plane.

[0078] Subtract the corresponding horizontal and vertical coordinate components of the first wearable device from the horizontal and vertical coordinate components of the second wearable device to obtain the Euclidean space vector pointing from the first wearable device to the second wearable device. The difference vector physically represents the linear displacement and absolute orientation of the target device relative to the searching device in a global stationary reference frame.

[0079] The real-time heading angle of the first wearable device is obtained and a rotation matrix is ​​constructed. The heading angle is calculated by fusing the magnetometer and gyroscope built into the first wearable device, and represents the angle between the current forward direction of the first wearable device and a fixed reference axis of the base station coordinate system. A two-dimensional rotation matrix is ​​constructed based on the heading angle. The rotation matrix is ​​composed of the cosine and sine function values ​​of the heading angle arranged according to coordinate transformation rules, and is used to describe the mapping relationship from the global coordinate system to the local coordinate system.

[0080] The difference vector is multiplied on the left by the inverse of the rotation matrix, and the difference vector is projected onto the local body coordinate system of the first wearable device. The local body coordinate system is a relative reference system centered on the first wearable device, with its front as the vertical axis and its right side as the horizontal axis.

[0081] The relative pose vector is obtained by analyzing the converted data. The vector contains two key navigation parameters: the yaw angle deviation of the target relative to the current wearer's orientation and the relative distance. The relative distance is obtained by calculating the modulus of the converted vector in the local body coordinate system. The yaw angle deviation is obtained by calculating the inverse trigonometric function values ​​of the horizontal and vertical components of the converted vector.

[0082] By mapping the absolute difference vector in the base station coordinate system to the local body coordinate system of the first wearable device through a coordinate transformation algorithm, an intuitive transformation from global absolute positioning to first-person relative guidance is achieved. By combining the real-time heading angle to construct a rotation matrix, it is ensured that the relative pose vector can be dynamically corrected according to the wearer's turning action. No matter how the wearer changes direction, it can output the accurate yaw angle deviation and distance relative to the wearer's current viewpoint.

[0083] The generated guidance instruction is configured to drive the linear motor inside the first wearable device, and the yaw angle deviation is mapped to tactile coded vibrations in different directions, and the relative distance is mapped to changes in vibration frequency.

[0084] The second guidance command is transmitted to the TV terminal via a wireless local area network. The TV terminal draws a radar scan map on the graphical user interface according to the command, in which directional arrows are used to indicate the relative position of the second wearable device, a dynamically shortened distance bar is used to intuitively display the degree of approach, and a status prompt text pops up in real time.

[0085] The relative pose vector is decomposed into two components: yaw angle deviation and relative distance, and corresponding control signals are generated through the tactile feedback channel and the visual feedback channel, respectively.

[0086] The configuration and tactile feedback process of the first guidance command drives the linear motor integrated inside the first wearable device to perform tactile interaction. For the yaw angle deviation, a pre-set orientation mapping logic divides the wearer's forward field of vision into a left yaw zone, a right yaw zone, and a forward locking zone. When the yaw angle deviation indicates the target is in the left yaw zone, the linear motor outputs a tactile coded vibration representing a left turn, using a short and rapid pulse sequence; when the yaw angle deviation indicates the target is in the right yaw zone, the linear motor outputs a tactile coded vibration representing a right turn, using a long and gentle pulse sequence; when the target enters the forward locking zone, a continuous steady-state vibration is output to prompt the wearer to maintain the current orientation.

[0087] Based on the change in vibration frequency mapped by the relative distance, a maximum search distance and a minimum contact distance are set as the mapping interval, and an inverse proportional mapping function is used to control the driving frequency of the linear motor. As the relative distance gradually decreases, the vibration frequency of the linear motor increases linearly or exponentially, and the vibration interval gradually shortens. This change in vibration rhythm from sparse to dense simulates the physical feedback mechanism of a heartbeat, allowing the wearer to intuitively perceive the approach of the target solely through skin touch without relying on vision.

[0088] The generation and visual feedback process of the second guidance command involves the first wearable device using its built-in wireless communication module to encapsulate and send the relative pose vector and identification information in real time to the associated TV or large-screen display device via a wireless local area network. The TV runs a dedicated receiving and rendering program. After receiving the data, it draws a radar scan map on the graphical user interface of the screen. The radar scan map of the mobile phone uses the center of the screen as the origin to represent the position of the first wearable device. In the graphical user interface, directional arrows are used to indicate the relative orientation of the second wearable device. Based on the received yaw angle deviation, the rotation angle is calculated in real time, and the directional arrows are controlled to rotate around the origin on the radar disk so that the tip of the arrow always dynamically points to the actual physical orientation of the second wearable device.

[0089] The approximation of proximity is visually displayed using a dynamically shortening distance bar or a color-gradient halo. Based on the relative distance value, a variable-length distance progress bar is drawn on one side of the interface, and the length of the progress bar is reduced proportionally as the distance decreases. At the same time, the interface displays status prompts in real time, showing text information such as "searching," "turning left to find," "approaching," and "about to meet" based on the current distance and orientation, to assist the user in making decisions.

[0090] At the close-wearing end, the coded vibration and frequency changes of the linear motor enable blind operation guidance without taking up visual attention, allowing the wearer to perceive direction and distance without frequently raising their hand to look at the screen; at the far-end display end, large-screen devices such as TVs provide intuitive vantage points and radar maps, making up for the shortcomings of wearable devices having small screens and limited information.

[0091] See Figure 3 The time-domain wave packet cross-correlation is configured as a linear frequency modulated signal whose frequency scans linearly with time, and the linear frequency modulated signal covers the resonant bandwidth of the haptic actuator;

[0092] A sliding time window is used to extract the high-frequency acceleration component from the environmental vibration data. The high-frequency acceleration component is then subjected to a discrete-time sliding dot product operation with the linear frequency modulated signal to generate a cross-correlation function curve.

[0093] Extract the maximum peak energy and main lobe width from the cross-correlation function curve; the resonance matching condition is that the maximum peak energy is higher than the preset energy coupling threshold, and the peak-to-sidelobe ratio of the main peak to the maximum sidelobe of the cross-correlation function curve is higher than the preset waveform fidelity threshold.

[0094] The activation logic of the near-field verification process involves real-time monitoring of the relative distance between the first wearable device and the second wearable device, calculated using an extended Kalman filter. When the relative distance is detected to be less than a preset contact threshold (set to 20cm in this embodiment), it is determined that the two devices have entered the potential physical contact range, the near-field verification mode is immediately triggered, and the update of the far-field radio positioning guidance command is paused.

[0095] The specific implementation process of controlling the first wearable device to transmit the coded vibration sequence is as follows: the microprocessor of the first wearable device generates a specific digital driving waveform, which is configured as a linear frequency modulation signal with a frequency that increases linearly with time; the start frequency and end frequency of the signal are set to cover the effective resonant bandwidth range of the haptic actuator, and the signal is set to linearly scan from 100Hz to 300Hz within a duration of 200 milliseconds, converting the digital waveform into an analog voltage signal, driving the linear motor inside the first wearable device to generate physical vibrations of a fingerprint with a specific frequency, which are then propagated to the external medium as a verification handshake signal;

[0096] The process of acquiring environmental vibration data sensed by the second wearable device involves the second wearable device receiving a sampling trigger command synchronously transmitted via a wireless link. It then activates its built-in high-precision accelerometer or inertial measurement unit to continuously collect environmental vibration data at a sampling rate higher than the Nyquist frequency. The acquisition process employs a sliding time window mechanism, capturing data frames of fixed time lengths in real time. For the captured data frames, a high-pass filter removes gravitational acceleration components and low-frequency interference components generated by human limb movements, retaining only the high-frequency acceleration components that reflect high-frequency vibration characteristics.

[0097] The specific calculation process for the time-domain wave packet cross-correlation involves using the original linear frequency modulated signal emitted by the first wearable device as a reference template and the high-frequency acceleration component acquired by the second wearable device as the observed signal. In the processor's processing unit, a discrete-time sliding dot product operation is performed on the reference template and the observed signal. This involves sliding the reference template point-by-point relative to the observed signal along the time axis, calculating the sum of the products of corresponding sampling points of the two signals at each sliding step, thereby generating a cross-correlation function curve reflecting the change in the similarity of the two signal waveforms with time delay.

[0098] The process of determining and exiting the resonance matching condition involves extracting features from the generated cross-correlation function curve, identifying the main peak with the largest amplitude in the curve, and determining the value of the main peak as the maximum peak energy. At the same time, the secondary peaks adjacent to the main peak are identified, and the ratio of the amplitude of the main peak to the amplitude of the maximum secondary peak, i.e., the peak-side lobe ratio, is calculated to characterize the fidelity and sharpness of the waveform.

[0099] The extracted feature values ​​are compared with a preset threshold. The first comparison determines whether the maximum peak energy is higher than the preset energy coupling threshold, which represents the minimum energy intensity required for the vibration signal to be transmitted through the physical medium, and is used to eliminate weak vibration interference propagated in the air. The second comparison determines whether the peak-sidelobe ratio is higher than the preset waveform fidelity threshold, which is used to confirm whether the received vibration mode strictly conforms to the specific coding characteristics of the linear frequency modulated signal, so as to prevent misjudgment caused by random environmental collision noise.

[0100] For signal processing in the near-field verification phase, the transmitter generates a linear frequency modulated signal with a frequency that linearly increases from 100Hz to 300Hz and a duration of 200ms, covering the resonant range of a general linear motor. The receiver collects environmental vibration data at a sampling rate of 4kHz and filters it to retain only the components in the 100Hz to 400Hz frequency band for cross-correlation calculation. To ensure the accuracy of contact judgment, a dual decision threshold is set: first, the energy coupling threshold is set to -20dB, which is considered airborne acoustic interference; second, the waveform fidelity threshold, i.e., the peak-to-sidelobe ratio, is set to 8dB, which is considered waveform distortion caused by transmission through a wall. Only when both conditions are met simultaneously is it determined to be a real physical contact and the search mode is automatically exited.

[0101] If both of the above conditions are met simultaneously, it indicates that the specific vibration generated by the first wearable device has been clearly and strongly sensed by the second wearable device, determining that the two are in the same rigid physical medium environment or have made direct contact. At this point, the mutual search task is confirmed to be complete, the positioning algorithm is automatically terminated, the search mode is exited, and a successful connection feedback prompt is triggered.

[0102] By employing linear frequency modulated signals as vibration codes for near-field verification and leveraging their strong autocorrelation characteristics in the time-frequency domain, the problem of traditional single-frequency vibrations being easily interfered with by environmental noise is effectively solved. Combined with time-domain wave packet cross-correlation algorithm and peak-sidelobe ratio verification, weak physical contact signals can be identified with extreme accuracy. In noisy dynamic environments, zero-false-alarm verification of touch confirmation is achieved, significantly improving the intelligent experience and certainty of the final step of device interoperability.

[0103] To address the technical challenges of existing wearable devices' mutual search processes, such as location drift caused by human occlusion and the lack of precise arrival confirmation mechanisms, this paper constructs a dynamic human shadow model and a dynamic sector confidence gating strategy. This enables real-time identification and suppression of non-line-of-sight measurement noise in blind spots caused by human occlusion, effectively avoiding signal fluctuations and location abrupt changes caused by wearer limb occlusion, and significantly improving the robustness and smoothness of the dynamic guidance process. Simultaneously, by combining a near-field verification mechanism based on coded vibration sequences and utilizing wave packet cross-correlation to accurately determine the physical contact status, this paper effectively solves the problem that relying solely on wireless signals cannot distinguish between proximity to a dividing wall and actual contact. This achieves an intelligent automatic closed loop from far-field guidance to near-field confirmation, greatly improving the accuracy of exiting the search mode and enhancing the user experience.

[0104] Example 2:

[0105] This embodiment applies a wearable device mutual location system based on entry and exit principles to the complex environment of a large warehousing and logistics center. The industrial-grade wireless positioning anchor points suspended from the steel structure at the top of the warehousing center are configured as the base stations. The smart work badges worn by warehouse supervisors are configured as the first wearable device, and the smart safety helmets worn by sorters working deep inside the shelves are configured as the second wearable device. The system includes the following modules and their specific operation processes:

[0106] The operation of the radio frequency sensing module involves setting the industrial-grade wireless positioning anchor point as a fixed anchor point, using its built-in large-aperture multi-antenna array to cover the work area, periodically broadcasting a synchronization beacon frame containing a time slot allocation table, and having the smart work badge of the unified scheduling supervisor and the smart safety helmet of the sorting worker send broadband detection reference signals in non-overlapping orthogonal time slots. The base station receives the above signals and parses the channel state information. By constructing a spatial spectrum covariance matrix and applying a multi-signal classification algorithm, it filters out non-line-of-sight reflection path components, accurately extracts the angle information of the two devices relative to the base station, and estimates the distance information based on the multi-carrier phase ranging principle, thus completing the basic time-slotted positioning data acquisition.

[0107] The posture modeling module operates by real-time monitoring of posture change data output by the inertial measurement unit built into the supervisor's smart badge, establishing a human coordinate system with the center of the supervisor's torso as the origin, and defining the supervisor's torso as a rigid cylindrical model with a preset physical radius; real-time calculation of the smart badge's spatial coordinates and antenna normal vector in the human coordinate system, constructing a line-of-sight path vector connecting the spatial coordinates and the base station coordinates; calculating the geometric perpendicular distance from the center of the torso to the line-of-sight path vector and the included angle between the antenna normal vector and the line-of-sight path vector; when the calculation results show that the geometric perpendicular distance is less than the physical radius, causing the line-of-sight path vector to cut the rigid cylindrical model, or the included angle is an obtuse angle, causing the antenna normal vector to deviate from the line-of-sight path vector, the system determines that it is currently in a human body occlusion blind zone; otherwise, the direction not determined as a human body occlusion blind zone is identified as a line-of-sight communication sector;

[0108] The operation of the pose fusion module is as follows: A dynamic sector confidence gating strategy is adopted. When the determination result of being in the human body occlusion blind zone is received, a noise expansion coefficient is automatically applied to the adaptive observation noise covariance matrix of the extended Kalman filter to generate a high-damped covariance matrix to suppress the Kalman gain, thereby reducing the weight of the measurement noise at this time. When in the line-of-sight communication sector, the reference noise parameters are maintained. Combined with the geometric constraints that introduce a maximum motion rate limit, the relative pose vector of the smart badge pointing to the smart safety helmet is fused and calculated, a guidance command is generated and fed back to the smart badge, driving the built-in linear motor to generate tactile encoded vibration, and a radar scan map is drawn on the supervisor's handheld terminal via a wireless network.

[0109] The near-field verification module operates as follows: when the system detects that the relative distance is less than a preset contact threshold, it automatically enters the near-field verification process, controlling the smart badge to transmit a coded vibration sequence that is linearly scanned over time; simultaneously, it controls the smart safety helmet to collect and sense environmental vibration data, and calculates the time-domain wave packet cross-correlation between the environmental vibration data and the coded vibration sequence; if the calculated time-domain wave packet cross-correlation meets the resonance matching conditions that include the maximum peak energy and the peak-sidelobe ratio, the system determines that the two are in the same physical medium environment (i.e., confirms that physical contact or extremely close-range transmission in the same medium has occurred), automatically exits the search mode, and completes the personnel docking task.

[0110] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for mutual positioning of wearable devices based on the principle of entry and exit, characterized in that, include: The base station is set as a fixed anchor point, and a multi-antenna array is used to obtain the angle information of the first wearable device and the second wearable device. The distance information is estimated by channel detection, and time-slotted positioning is performed at the same time. Real-time monitoring of posture change data of the first wearable device; construction of a dynamic human shadow model to identify the line-of-sight communication sector of the first wearable device relative to the base station and the blind spot of human occlusion. A dynamic sector confidence gating strategy is adopted to reduce the measurement noise weight when the device is in the blind zone of the human body. The relative pose vector of the first wearable device pointing to the second wearable device is calculated by combining geometric constraints and extended Kalman filtering, and a guidance command is generated and fed back to the first wearable device. When the relative distance is less than a preset contact threshold, near-field verification is initiated. The first wearable device is controlled to emit an coded vibration sequence, and environmental vibration data sensed by the second wearable device is acquired. The time-domain wave packet cross-correlation between the environmental vibration data and the coded vibration sequence is calculated. If the time-domain wave packet cross-correlation satisfies the resonance matching condition, it is determined that the two are in the same physical medium environment, and the search mode is automatically exited.

2. The wearable device mutual positioning method based on the entry and exit principle according to claim 1, characterized in that, The process of acquiring the angle information and distance information involves the base station periodically broadcasting a synchronization beacon frame containing a time slot allocation table, and scheduling the first wearable device and the second wearable device to send detection reference signals in different orthogonal time slots. The base station uses a multi-antenna array to receive the detection reference signal, analyzes the channel state information of each subcarrier, and constructs a spatial spectrum covariance matrix; The covariance matrix is ​​decomposed using a multiple signal classification algorithm to filter out non-line-of-sight reflection path components and extract angle information. The phase slope of the detection reference signal between different subcarriers is analyzed, and the distance information is calculated based on the multi-carrier phase ranging principle.

3. The wearable device mutual positioning method based on the entry and exit principle according to claim 1, characterized in that, The human body shadow dynamic model is established by creating a human body coordinate system with the center of the torso of the wearer of the first wearable device as the origin, and defining the human torso as a rigid cylindrical model with a physical radius. Based on the posture change data, the spatial coordinates and antenna normal vector of the first wearable device in the human coordinate system are calculated, and a line-of-sight path vector connecting the spatial coordinates and the base station coordinates is constructed. The geometric perpendicular distance from the center of the torso to the line-of-sight path vector is calculated, and the included angle between the antenna normal vector and the line-of-sight path vector is calculated. When the geometric perpendicular distance is less than the physical radius, causing the line-of-sight path vector to cut the rigid cylindrical model, it is determined to be the human body occlusion blind zone. And when the included angle is an obtuse angle, causing the antenna normal vector to deviate from the line-of-sight path vector, it is determined to be the human body occlusion blind zone. The direction that is not identified as the blind spot of the human body is determined as the line-of-sight communication sector.

4. The wearable device mutual positioning method based on the entry and exit principle according to claim 1, characterized in that, The dynamic sector confidence gating strategy involves constructing an adaptive observation noise covariance matrix for the extended Kalman filter and setting a baseline noise parameter. In response to the current direction being determined to be the human body occlusion blind zone, a noise expansion coefficient is applied to the reference noise parameters to generate a high-damped covariance matrix and suppress Kalman gain; In response to the current direction being identified as the line-of-sight communication sector, the reference noise parameters are maintained, and the state update is biased towards the observations of the measurement update step; geometric constraints are used to introduce a maximum motion rate limit during the filtering process to eliminate abnormal pose abrupt values ​​that exceed the limits of human kinematics.

5. The wearable device mutual positioning method based on the entry and exit principle according to claim 1, characterized in that, The relative pose vector calculation method is to obtain the position coordinates of the first wearable device and the position coordinates of the second wearable device in the base station coordinate system; Calculate the difference vector in the base station coordinate system, where the difference vector represents the straight-line distance and direction between two points; obtain the heading angle of the first wearable device and construct a rotation matrix; The difference vector is transformed from the base station coordinate system to the local body coordinate system of the first wearable device using the rotation matrix to obtain the relative pose vector; the relative pose vector includes the yaw angle deviation of the target relative to the current wearer's orientation and the relative distance.

6. The wearable device mutual positioning method based on the entry and exit principle according to claim 1, characterized in that, The generated boot instructions include a first boot instruction and a second boot instruction; The first guidance command is configured to drive the linear motor inside the first wearable device, which maps the yaw angle deviation to tactile coded vibrations in different directions and maps them to changes in vibration frequency based on the relative distance. The second guidance command is transmitted to the TV via a wireless local area network. The TV draws a radar scan map on the graphical user interface according to the command. The directional arrows indicate the relative position of the second wearable device, the approach level is displayed intuitively by dynamically shortening distance bars or color gradient halos, and status prompt text pops up in real time.

7. The wearable device mutual positioning method based on the entry and exit principle according to claim 1, characterized in that, The time-domain wave packet cross-correlation is configured as a linear frequency-modulated signal with a frequency that scans linearly with time, and the linear frequency-modulated signal covers the resonant bandwidth of the haptic actuator; A sliding time window is used to extract the high-frequency acceleration component from the environmental vibration data. The high-frequency acceleration component is then subjected to a discrete-time sliding dot product operation with the linear frequency modulated signal to generate a cross-correlation function curve. Extract the maximum peak energy and main lobe width from the cross-correlation function curve; the resonance matching condition is that the maximum peak energy is higher than the preset energy coupling threshold, and the peak-to-sidelobe ratio of the main peak to the maximum sidelobe of the cross-correlation function curve is higher than the preset waveform fidelity threshold.

8. A wearable device mutual positioning system based on the principle of entry and exit, characterized in that, include: Radio frequency sensing module: The base station is set as a fixed anchor point and a multi-antenna array is used to obtain the angle information of the first wearable device and the second wearable device. The distance information is estimated by channel detection and time-slotted positioning is performed at the same time. Posture modeling module: Real-time monitoring of posture change data of the first wearable device, constructing a dynamic human shadow model to identify the line-of-sight communication sector of the first wearable device relative to the base station and the blind spot of human occlusion; Pose fusion module: adopts dynamic sector confidence gating strategy to reduce measurement noise weight when in the blind zone of the human body, and combines geometric constraints and extended Kalman filter to calculate the relative pose vector of the first wearable device pointing to the second wearable device, and generates guidance instructions to be fed back to the first wearable device. Near-field verification module: When the relative distance is less than a preset contact threshold, near-field verification is entered, the first wearable device is controlled to emit an coded vibration sequence, and the environmental vibration data sensed by the second wearable device is acquired; the time-domain wave packet cross-correlation between the environmental vibration data and the coded vibration sequence is calculated; if the time-domain wave packet cross-correlation satisfies the resonance matching condition, it is determined that the two are in the same physical medium environment, and the search mode is automatically exited.

Images