GEO fence triggering method and triggering system based on multi-mode positioning data fusion

Abstract

The invention relates to the technical field of geo-fence triggering, in particular to a multi-mode positioning data fused GEO fence triggering method and triggering system. The method comprises the following steps that according to the distance between equipment and the fence boundary, a layered positioning strategy is adopted to demarcate a safe area, a transition area and a critical area; when the device is located in the safe area, outputting a coarse position coordinate of the device and a corresponding coarse positioning confidence coefficient by adopting a low-power-consumption coarse positioning algorithm; when the equipment is located in the transition area, fusing the GPS data and the IMU data by adopting an adaptive fusion algorithm to obtain an intermediate fusion position, an intermediate fusion speed and an intermediate positioning confidence coefficient; fusing the multi-modal data when the equipment is located in the critical zone, and predicting the probability that the equipment invades the fence by combining a trajectory risk assessment algorithm based on the equipment acceleration and the distance between the equipment and the fence boundary. Through hierarchical positioning strategy and multi-modal data fusion, the problems of insufficient precision and high power consumption of a single positioning mode in a complex scene are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of geofence triggering, and more specifically, to a GEO fence triggering method and a triggering system for multi-modal positioning data fusion. Background Art

[0002] The virtual boundary of a GEO fence is based on a globally common coordinate reference system (such as WGS-84, EPSG:4326), and a circular or polygonal area is defined by a center point + radius or vertex latitudes and longitudes. The device location acquisition is also based on this coordinate system, and the boundary distance is calculated through the great circle distance (haversine) formula to ensure the consistency between the fence boundary and the device positioning.

[0003] Existing GEO fence triggering methods mostly rely on a single positioning mode (such as GPS or cellular network), and it is difficult to achieve a balance among power consumption, accuracy, and real-time performance. Especially in complex scenarios (such as high-rise building occlusion and multipath interference in urban areas), the positioning error is significant, resulting in missed triggers or false triggers. How to select an appropriate positioning mode in different distance intervals and improve the fusion accuracy and triggering frequency when approaching the boundary is an urgent problem to be solved. Therefore, a GEO fence triggering method and a triggering system for multi-modal positioning data fusion are provided. Summary of the Invention

[0004] The purpose of the present invention is to provide a GEO fence triggering method and a triggering system for multi-modal positioning data fusion, so as to solve the problem in the above background art that existing GEO fence triggering methods mostly rely on a single positioning mode (such as GPS or cellular network), and it is difficult to achieve a balance among power consumption, accuracy, and real-time performance. Especially in complex scenarios (such as high-rise building occlusion and multipath interference in urban areas), the positioning error is significant, resulting in missed triggers or false triggers.

[0005] To achieve the above purpose, the present invention aims to provide a GEO fence triggering method for multi-modal positioning data fusion, including the following steps: S1. According to the distance between the device and the fence boundary, a hierarchical positioning strategy is adopted to demarcate a safe area, a transition area, and a critical area respectively, and the area to which the device belongs is demarcated based on the current position of the device; S2. When the device is located in the safe area, a low-power coarse positioning algorithm is used to output the coarse position coordinates of the device and the corresponding coarse positioning confidence ; S3. When the device is located in the transition area, an adaptive fusion algorithm is used to fuse GPS data and IMU data to obtain an intermediate fusion position , an intermediate fusion speed and an intermediate positioning confidence ; S4. When the device is located in the critical area, a multi-modal data fusion algorithm is used to fuse the multi-modal data, and the risk fusion position is output and the risk fusion confidence ; Based on the device acceleration and the distance between the device and the fence boundary, the probability of the device invading the fence is predicted by combining the trajectory risk assessment algorithm; S5. Based on the multi-source signals generated in S2 - S4, the final execution instruction is generated through the preset multi-condition decision logic.

[0006] As a further improvement of this technical solution, the specific method involved in the hierarchical positioning strategy is: Calculate the current position coordinates of the device based on the low-power rough positioning algorithm, and calculate the shortest boundary distance between the current position of the device and the fence boundary in combination with the fence boundary configuration; Set the safety transition threshold and the critical transition threshold respectively based on the shortest boundary distance; When the shortest boundary distance is less than the safety transition threshold, it is designated as the safe area; When the shortest boundary distance is greater than the safety transition threshold and less than the safety transition threshold , it is designated as the transition area; When the shortest boundary distance is less than the critical transition threshold, it is designated as the critical area; Among them, when the device distance exactly falls near the safety transition threshold or the critical transition threshold, a switching suppressor is introduced to maintain the current positioning mode for seconds.

[0007] As a further improvement of this technical solution, the low-power rough positioning algorithm is specifically: Scan the list of visible base stations around and record the corresponding received signal strength indication values ; Perform short-term smoothing processing on the original RSSI and introduce an interference factor , and obtain the smoothed value ; After performing interference correction on the , introduce the path loss model and correct it through the interference amplification coefficient to obtain the distance between the base station and the device ; According to the corrected distance , perform coordinate calculation through the trilateration algorithm, and output the rough position coordinates of the device ; Meanwhile, based on the signal variance, interference variance and geometric condition number , calculate the rough positioning reliability .

[0008] As a further improvement of this technical solution, in S3, the intermediate position information and the fusion confidence are obtained by the adaptive fusion algorithm. The specific steps involved are as follows: Collect the device's longitude and latitude through the GPS module , speed and the positioning accuracy variance ; Obtain the device's acceleration from the IMU module , speed , angular velocity and its own noise variance ; Based on the acceleration and the angular velocity , calculate the relative displacement through the Newton-Euler equation ; Convert the IMU acceleration to the global coordinates , and based on the global coordinates predict the position and speed at the current moment through the Newton-Euler equation ; Calculate the Kalman gain by the predicted state variance matrix and the GPS observation noise ; Fuse the IMU predicted position and speed with the GPS observed position and speed by weighting according to the Kalman gain and output the intermediate fusion position and the intermediate fusion speed ; Introduce the device's longitude and latitude , and calculate the GPS fusion weight based on the positioning accuracy variance ; Based on the noise variance and the positioning accuracy variance , and introduce the geometric distribution factor to calculate the intermediate positioning reliability ; According to the real-time distance between the device and the fence boundary and the intermediate positioning reliability , dynamically adjust the detection frequency​ Perform secondary adjustment by introducing a frequency smoother and restricting ; Based on the fusion speed and acceleration , predict the displacement within the future ; If the predicted position enters the fence, trigger an alarm in advance.

[0009] As a further improvement of this technical solution, the specific steps involved in the multi-modal data fusion algorithm are as follows: Use the intermediate fusion position as the basic prediction value, introduce the rough position coordinates , and calculate the position deviation ; Set the deviation threshold . If the position deviation is greater than the deviation threshold , then correct the intermediate fusion position to obtain the corrected position coordinates ; Use as the third type of observation, use the observation matrix and the observation noise covariance , and substitute them into the Kalman update; Synchronously calculate the Kalman gain , and update the state and covariance matrix of the Kalman filter; Then the first component of the state of the updated Kalman filter is the final output risk fusion position , and the risk fusion confidence is calculated inversely from the diagonal element of the covariance .

[0010] As a further improvement of this technical solution, the specific steps involved in the trajectory risk assessment algorithm for predicting the probability of a device invading the fence are as follows: Based on the current risk fusion position and the intermediate fusion speed , construct the cloud state vector ; Based on the acceleration and the angular velocity , calculate the predicted displacement within the future time through the Newton-Euler equation; Introduce the acceleration noise variance and the angular velocity noise variance , and perform first-order linear error propagation to obtain the predicted displacement variance ; Based on the future displacement within a period of time and the predicted displacement variance , the probability that the device invades the fence is calculated as ; Set a preset threshold , if the probability that the device invades the fence is , and it exceeds the suppression holding time , then an intrusion alarm is output.

[0011] As a further improvement of this technical solution, the multi-condition decision logic is specifically as follows: Based on the multi-source signals generated in the above steps S2 - S4, generate the final decision logic: When the device is in the safe area, only pay attention to the rough positioning confidence , set the high confidence threshold for the safe area , if , then maintain the current low-power positioning mode without switching; if , then enter the transition area positioning mode and enter the medium-frequency fusion mode; When the device is in the transition area, pay attention to the intermediate positioning confidence , set the high confidence threshold for the transition area and the low confidence threshold for the transition area , if , then maintain the medium-frequency fusion mode; If and , then trigger multi-source verification and re-evaluate the intermediate positioning confidence ; If , then forcefully fallback to the safe area mode and reduce to the low-power positioning mode; When the device is in the critical area, pay attention to the risk fusion confidence and the intrusion probability , preset the intrusion probability threshold , the high confidence alarm threshold for the critical area and the low confidence suppression threshold for the critical area ; If ≥ and ≥ , then output the high-priority positioning mode; If but , then output the warning mode and turn on the auxiliary positioning; If when the device is in the critical area, the above two conditions are not met, then introduce weighted scoring , if the weighted score If it is greater than the comprehensive critical threshold, enter the warning mode; otherwise, maintain the conventional monitoring in the critical area.

[0012] On the other hand, the present invention provides a GEO fence triggering system for multi-modal positioning data fusion, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the multi-modal positioning data fusion GEO fence triggering method described in any one of the above.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. In the multi-modal positioning data fusion GEO fence triggering method and triggering system, through the hierarchical positioning strategy and multi-modal data fusion, the problems of insufficient accuracy and high power consumption of a single positioning mode in complex scenarios are solved. By using regional division and dynamic mode switching, the optimal positioning scheme is selected in different distance intervals, and resource allocation is optimized by introducing a switching suppressor and dynamic frequency adjustment, reducing ineffective energy consumption.

[0014] 2. In the multi-modal positioning data fusion GEO fence triggering method and triggering system, by dynamically allocating the weights of GPS and IMU through the Kalman gain, introducing rough positioning as an observation value, and triggering position correction through a deviation threshold, the problem of requiring high-precision positioning in complex scenarios is solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is the overall method flow chart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0017] Embodiment 1: Please refer to Figure 1 As shown, this embodiment provides a GEO fence triggering method for multi-modal positioning data fusion, including the following steps: S1. According to the distance between the device and the fence boundary, use a hierarchical positioning strategy to delimit a safe area, a transition area, and a critical area respectively, and delimit the area to which the device belongs based on the current position of the device; In this embodiment, the specific method involved in the hierarchical positioning strategy is as follows: Based on a low-power rough positioning algorithm, calculate the current position coordinates of the device, and combine the fence boundary configuration to calculate the shortest boundary distance between the current position of the device and the fence boundary; Set the safety transition threshold and the critical transition threshold respectively based on the shortest boundary distance and the critical transition threshold ; When the shortest boundary distance is less than the safety transition threshold, it is designated as a safe area; When the shortest boundary distance is greater than the safety transition threshold and less than the safety transition threshold it is designated as a transition area, indicating that the current position of the device is between the safe area boundary and the transition area boundary; When the shortest boundary distance is less than the critical transition threshold, it is designated as a critical area, indicating that the current position of the device is between the fence boundary and the critical area boundary; In this embodiment, the safety transition threshold represents the distance of the device from the fence boundary; The critical transition threshold represents the distance of the device from the fence boundary, and ; Set an overlapping buffer at the safety transition threshold and the critical transition threshold; Furthermore, when the distance of the device exactly falls near the safety transition threshold or the critical transition threshold, the system will frequently switch back and forth between the safe area - transition area and the transition area - critical area, causing jitter in the positioning mode, not only wasting energy consumption, but also causing fluctuations in confidence. Therefore, a switching suppressor is introduced to maintain the current positioning mode for at least seconds after crossing the threshold to prevent frequent mode switching.

[0018] In this embodiment, a switching suppressor is introduced to prevent the device from frequently switching the positioning mode due to position fluctuations near the fence boundary; If the safety transition threshold of the fence boundary is 10 meters and the critical transition threshold is 5 meters, and the device moves back and forth near the safety transition threshold (9 - 11 meters), its position fluctuates between the safe area and the transition area due to signal interference; When the switching suppressor is not introduced, if a mode switch is triggered every 30 seconds due to position fluctuations, it will cause frequent activation of the GPS and IMU sensors, resulting in increased power consumption, and the rough positioning and fusion positioning results are alternately output, the intrusion risk assessment is unstable. At the same time, it causes a brief decrease in the confidence of the transition area, and may wrongly trigger a fallback to the safe area mode, missing the actual risk; Set the suppression time of the switching suppressor to 60 seconds, then when the device reduces from a distance (the distance of the device from the fence boundary) of 11 meters to 9.5 meters (crossing the safety transition threshold ), the system switches to the transition zone mode, enables GPS+IMU integrated positioning, and starts the suppression timer; Within the next 60 seconds, even if the device briefly returns to 10.5 meters (safe zone), the transition zone mode is still maintained, and the integrated positioning continues; After 60 seconds, the position is re-detected. If the device is still in the safe zone, it switches back to the low-power mode; if the device enters the critical zone (such as the distance drops to 4 meters) during the suppression period, the critical zone mode is immediately triggered and the suppression timer is reset; the switching suppressor realizes a smooth transition of the mode switch in the multi-modal positioning system by introducing a time lag mechanism.

[0019] S2. When the device is in the safe zone, a low-power rough positioning algorithm is used to output the rough position coordinates of the device and the corresponding rough positioning confidence ; In this embodiment, the low-power rough positioning algorithm is specifically: Scan the list of visible base stations around , and record the corresponding Received Signal Strength Indicator (RSSI) values ; Perform short-term smoothing processing on the original RSSI (such as moving average), eliminate the jump values, and introduce an interference factor due to the RSSI jitter caused by interference sources (such as multipath, co-frequency noise) in the environment , to obtain the smoothed value ; ; In the formula, represents the original RSSI value, the original received signal strength indicator of the th base station scanned by the device, and the unit is dBm; represents the smoothed RSSI value; represents the th base station's interference factor; represents the short-term smoothing processing function; After performing interference correction on , introduce the path loss model and correct it through the interference amplification coefficient to obtain the distance between the base station and the device ; ; In the formula, represents the distance from the device to the th base station, and the unit is meter; represents the signal strength calibration value at the reference distance (usually 1 meter), and the unit is dBm; represents the path loss exponent, reflecting the signal attenuation rate; is the interference amplification coefficient, which is used to reflect the amplification error of the distance estimation caused by the degradation of the signal quality in the environment; According to the corrected distance , the coordinate solution is performed through the trilateration algorithm, and the rough position coordinates of the device are output ; Among them, the trilateration algorithm: For each signal source, a corresponding estimated distance is generated , and at least 3 sources with the strongest signals are selected , …, ; Establish a system of equations: Use differential elimination to linearize it and solve to obtain the rough position coordinates of the device ; In the formula, represents the corrected distance from the device to the first base station; represents the corrected distance from the device to the second base station; represents the corrected distance from the device to the third base station; represents the coordinates of the first base station; represents the coordinates of the second base station; represents the coordinates of the third base station; represents the rough position coordinates of the device; At the same time, based on the signal variance, interference variance and geometric condition number , calculate the rough positioning reliability .

[0020] Specifically, calculate the RSSI variance ; Calculate the geometric condition number , where is the source direction vector; Comprehensively obtain the positioning reliability ; ; ; In the formula, is the interference factor variance index; represents the geometric condition number, which is used to reflect the influence of the base station layout on the positioning accuracy; represents the smoothed RSSI variance; represents the variance of the interference factor, reflecting the volatility of the environmental interference; represents a very small constant to prevent the denominator from being zero; represents the The unitized result of the direction vector of a base station.

[0021] S3. When the device is located in the transition zone, the GPS data and IMU data are fused using an adaptive fusion algorithm to obtain the intermediate fusion position , the intermediate fusion speed and the intermediate positioning confidence ; In this embodiment, the specific steps for obtaining the intermediate position information and fusion confidence by the adaptive fusion algorithm are as follows: The longitude and latitude of the device are collected through the GPS module , the speed and the positioning accuracy variance ; The acceleration of the device is obtained from the IMU module , the speed , the angular velocity and its own noise variance ; ; ; Based on the acceleration and the angular velocity , the relative displacement is calculated through the Newton-Euler equation ; ; In the formula, represents the sampling time interval; represents the IMU speed at the previous moment; represents the IMU acceleration at the current moment; the longitude and latitude of the device represents the longitude and latitude coordinates of the device measured by GPS; the speed represents the speed of the device measured by GPS; the acceleration represents the three-axis acceleration of the device measured by the inertial measurement unit (IMU); the angular velocity represents the three-axis angular velocity of the device measured by the IMU; The IMU acceleration is converted to the global coordinate , and based on the global coordinate the position and speed at the current moment are predicted through the Newton-Euler equation; Among them: ; In the formula, is the component of gravity in the body coordinate system, which is separated from the original acceleration through low-pass filtering; represents the net acceleration in the global coordinate system; represents the rotation matrix from the IMU body coordinate system to the global coordinate system, reflecting the device's attitude (pitch angle, roll angle, yaw angle), and is usually updated in real time through gyroscope integration or attitude calculation (such as quaternion); ; ; In the formula, represents the fused position at the previous moment (global coordinate system); represents the fused velocity at the previous moment (global coordinate system); represents the time interval; represents the predicted position at the current moment; represents the predicted velocity at the current moment; Through the predicted state variance matrix and the GPS observation noise calculate the Kalman gain ; In the formula, represents the predicted state covariance matrix, indicating the uncertainty of the predicted position and velocity; represents the observation matrix, mapping the state vector to the observation space (such as GPS only observes the position); represents the observation noise covariance matrix (GPS positioning accuracy variance); represents the Kalman gain, determining the weight distribution between the predicted value and the observed value; represents mapping the information in the observation space back to the state space in reverse; When the IMU prediction error is large and the GPS observation accuracy is high ( small): increases, and it depends more on GPS to correct the IMU prediction error; When the IMU prediction error is small and the GPS observation noise is large ( large): decreases, and it trusts more in the short-term prediction of the IMU; is the real-time assessment of the sensor reliability by the system: When in an open area (strong GPS signal), is smaller, is close to 1, and the fusion result is dominated by GPS; When in a tunnel (GPS signal lost), increases, is close to 0, and the fusion result completely depends on IMU calculation.

[0022] The predicted position of the IMU and velocity are weighted and fused with the GPS observed position and velocity according to the Kalman gain to output the intermediate fused position and the intermediate fused velocity ; Suppress the integration drift of the IMU and the jump noise of the GPS through weighted averaging: ; ; Introduce the longitude and latitude of the device , and calculate the GPS fusion weight based on the positioning accuracy variance ; ; If the GPS signal is abnormal ( ), trigger anomaly detection and mask the current GPS data; Based on the noise variance and the positioning accuracy variance , and introduce a geometric distribution factor (such as the satellite DOP value) to calculate the intermediate positioning confidence , which is used to quantify the global reliability of the fused positioning result and guide system-level decisions (such as mode switching, multi-source verification); ; In the formula, represents the DOP normalization coefficient; represents the accuracy attenuation factor; Hierarchical response: When , high confidence, directly output the result; When , trigger multi-source verification (map matching / visual-aided positioning), correct the fused result and recalculate the confidence; When , forcefully fallback to the safe zone mode and reduce the detection frequency to save power consumption; According to the real-time distance between the device and the fence boundary and the intermediate positioning confidence , dynamically adjust the detection frequency , and introduce the intermediate positioning confidence to perform a secondary adjustment on the detection frequency , by introducing a frequency smoother and restricting to avoid sampling jitter, represents the frequency change rate limit; Distance-driven frequency reference: ; Introduce Perform secondary adjustment on the frequency to enhance scene adaptability: ; Introduce a frequency smoother to perform first-order low-pass filtering on the detected frequency to limit the maximum change rate (limit ); In the formula, represents the basic detection frequency (default frequency in the safe zone); represents the maximum detection frequency (highest frequency in the critical zone); represents the safe transition threshold; represents the critical transition threshold; represents the real-time distance from the current position of the device to the fence boundary; Furthermore, based on the fusion speed and acceleration , predict the displacement within the future time: Among them, ; is the component of gravity in the body coordinate system, separated from the original acceleration through low-pass filtering; represents the net acceleration in the global coordinate system; represents the rotation matrix from the IMU body coordinate system to the global coordinate system, reflecting the device's attitude (pitch angle, roll angle, yaw angle), usually updated in real time through gyroscope integration or attitude solution (such as quaternion); Then ; If the predicted position enters the fence , an alarm is triggered in advance.

[0023] S4. When the device is located in the critical zone, use a multi-modal data fusion algorithm to fuse multi-modal data and output the risk fusion position and the risk fusion confidence ; In this embodiment, the specific steps involved in the multi-modal data fusion algorithm are as follows: Take the intermediate fusion position as the basic prediction value, introduce the rough position coordinates , calculate the position deviation , and the position deviation represents the difference between the intermediate fusion position and the rough positioning position, used to detect IMU integration drift or GPS sudden jumps; To suppress the long-term integration drift of the IMU and correct the GPS sudden jump, a deviation threshold is set , if the position deviation is greater than the deviation threshold , then correct the intermediate fusion position to obtain the corrected position coordinates ; ; Take as the third type of observation, use the observation matrix and the observation noise covariance , and substitute them into the Kalman update. The observation matrix is ; Among them, the correction coefficient is dynamically adjusted by to obtain: In the formula, represents the correction coefficient (dynamically adjusted depending on the confidence level); represents the deviation between the intermediate fusion position and the rough positioning; represents the rough positioning confidence; represents a very small constant; represents the intermediate fusion confidence; Synchronously calculate the Kalman gain , and update the state and covariance matrix of the Kalman filter; The noise covariance ; Calculate the Kalman gain ; Update the state and covariance matrix of the Kalman filter, and optionally perform exponential decay on the covariance ; Among them, ; ; In the formula, represents the observation matrix; represents the predicted state covariance matrix, reflecting the uncertainty of the predicted position and speed; represents the observation noise covariance; represents the final fusion state; is the predicted value, representing the device position and speed deduced by integrating the IMU acceleration or speed; represents the corrected position coordinates; Dynamically adjust the process noise: ; The first component of the state of the updated Kalman filter is the final output risk fusion position , and the risk fusion confidence is calculated inversely from the diagonal elements of the covariance ; ; ; In the formula, represents the process noise covariance matrix; represents the adaptive process noise covariance; represents the position deviation; represents the deviation threshold; represents the variance of the position component in the state covariance matrix after Kalman filter update.

[0024] Furthermore, based on the device acceleration and the distance between the device and the fence boundary, the probability of the device invading the fence is predicted by combining the trajectory risk assessment algorithm; The specific steps involved in the trajectory risk assessment algorithm predicting the probability of the device invading the fence are as follows: Based on the current risk fusion position and the intermediate fusion speed , a cloud state vector is constructed; Among them, , in the formula, represents the abscissa of the risk fusion position ; represents the ordinate of the risk fusion position ; represents the horizontal component of the intermediate fusion speed ; represents the vertical component of the intermediate fusion speed ; represents the lateral acceleration component of the device in the global coordinate system (converted from IMU acceleration); represents the longitudinal acceleration component of the device in the global coordinate system; the state vector represents the integration of position, speed, and acceleration information and is used to predict the future movement trajectory of the device; Based on the acceleration and the angular velocity , the displacement predicted for the future time is calculated through the Newton-Euler equation; ; To evaluate the uncertainty of , the acceleration noise variance and the angular velocity noise variance are introduced, and first-order linear error propagation is performed to obtain the predicted displacement variance ; represents the fusion speed at the previous moment; represents the net acceleration in the global coordinate system (the motion acceleration after removing gravity); represents the prediction time window; ; wherein, ; ; In the formula, represents the acceleration noise variance; represents the angular velocity noise variance; represents the total variance of the predicted displacement; represents the contribution of the acceleration noise; represents the contribution of the angular velocity noise; Based on the displacement within the future time and the predicted displacement variance , the probability that the device intrudes into the fence is calculated as ; The condition for the device to intrude into the fence is: The device moves from the current position (the distance from the current position to the fence is ) in the direction of the fence, and its displacement satisfies: ; That is, when the device moves at least distance will it intrude into the fence, and ; In the formula, represents the critical transition threshold; represents the real-time distance from the current position of the device to the fence boundary; represents the predicted displacement standard deviation; represents the base of the natural logarithm; represents the integration variable, which is a dummy variable in the integration, representing the possible displacement values of the device within the future time; Specifically, in the above formula for calculating the probability that the device intrudes into the fence, by quantifying the risk probability of the future behavior of the device, complex multi-sensor data (such as position, speed, acceleration, noise) are fused into an intuitive risk indicator, providing a dynamic decision-making basis for the fence monitoring system, and by predicting the future displacement of the device in real time, combining environmental noise and sensor errors, judging the possibility of intruding into the fence; predicting the future displacement of the device in real time, combining environmental noise and sensor errors, judging the possibility of intruding into the fence.

[0025] Furthermore, a preset threshold is set. , if the probability of the device invading the fence is , and it exceeds the suppression holding time , then an intrusion alarm is output; Otherwise, maintain the normal state and last at least seconds in the critical area mode to avoid frequent switching of the positioning mode.

[0026] S5. Based on the multi-source signals generated in S2 - S4, generate a final execution instruction through a preset multi-condition decision logic; In this embodiment, the multi-condition decision logic is specifically: Based on the multi-source signals generated in the above steps S2 - S4, generate a final decision logic: When the device is in the safe area, only focus on the rough positioning reliability , set a high confidence threshold for the safe area , if , then maintain the current low-power positioning mode without switching; if , then enter the transition area positioning mode and enter the medium-frequency fusion mode; When the device is in the transition area, focus on the intermediate positioning reliability , set a high confidence threshold for the transition area and a low confidence threshold for the transition area , if , then maintain the medium-frequency fusion mode (maintain GPS + IMU fusion, detection frequency ); If and , then trigger multi-source verification and re-evaluate the intermediate positioning reliability ; If , then forcefully fallback to the safe area mode and reduce to the low-power positioning mode; When the device is in the critical area, focus on the risk fusion confidence and the intrusion probability , preset an intrusion probability threshold , a high confidence alarm threshold for the critical area and a low suppression threshold for the critical area ; If ≥ and ≥ , then output a high-priority positioning mode (send an emergency alarm signal to the upper-layer security system and the user); If but , the warning mode is output (when the positioning confidence is insufficient and the risk is still low, first raise the warning level, increase the detection frequency, and turn on the assisted positioning), and the assisted positioning is turned on (such as vision, map matching); If the device does not meet the above two conditions when it is in the critical area, a weighted score is introduced , if the weighted score is greater than the comprehensive critical threshold , it enters the warning mode, otherwise it maintains the routine monitoring in the critical area (high-frequency fusion but no alarm).

[0027] Specifically, ; In the formula, represents the confidence weight. If the environment (such as tunnels, buildings) seriously affects the GPS / IMU fusion accuracy, it can be appropriately reduced. Conversely, it can be increased in open scenarios; represents the distance factor weight. If a high requirement for quick response to crossing the boundary is required, it can be increased. If more attention is paid to the fusion confidence and probability prediction, it can be reduced; represents the intrusion probability weight. In dynamic scenarios (such as vehicles, high-speed movement), probability prediction is more critical and can be increased. If it is in a static or slow-moving scenario, it can be appropriately reduced; , according to different scenarios or the sensor status during operation, the weights are adjusted online, so that the system can more flexibly adapt to the environment; represents the normalization of the intermediate positioning confidence; represents the normalization of the distance factor. The closer to the fence ( the smaller), the larger the value; represents the distance from the current position of the device to the fence boundary; represents the critical transition threshold; represents the intrusion probability; represents the maximum achievable value of the risk fusion confidence; represents the comprehensive risk score; High confidence threshold for the safe area is used to determine whether the rough positioning result in the safe area is reliable enough. If not, it enters the transition area; High confidence threshold for the transition area is used to determine whether the intermediate fusion result can be directly output; Low confidence threshold for the transition area is used to determine whether multi-source verification or fallback to the safe area needs to be triggered; High confidence warning threshold for the critical area is used to ensure an alarm is issued under high confidence conditions; Low confidence suppression threshold for the critical area is used to trigger an early warning and assisted positioning first when the confidence is extremely low to avoid false alarms.

[0028] In this embodiment, only the low-power coarse positioning algorithm is enabled in the low-power positioning mode, and GPS and IMU are in the sleep state. Among them, the sampling frequency takes the minimum reference value. ; In the medium-frequency fusion mode, GPS and IMU are both turned on, and the adaptive fusion algorithm is used to fuse GPS and IMU data, and the intermediate fusion position , intermediate fusion speed and intermediate positioning confidence are output. Among them, the sampling frequency is increased to the medium frequency , and the sampling frequency is driven by distance and dynamically adjusted in combination with confidence, and the change rate is limited by ; In the high-priority positioning mode, full-scale multi-modal fusion (GPS, IMU, RSSI / AP) is adopted to output the risk fusion position and risk fusion confidence , and the probability of the device invading the fence is calculated based on the trajectory risk assessment algorithm as .

[0029] Embodiment 2: This embodiment provides a GEO fence trigger system for multi-modal positioning data fusion, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the multi-modal positioning data fusion GEO fence trigger method described in any one of the above.

[0030] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A GEO fence triggering method for multi-modal positioning data fusion, characterized in that, It includes the following steps: S1. According to the distance between the device and the fence boundary, adopt a hierarchical positioning strategy to delimit the safety zone, transition zone, and critical zone respectively, and delimit the area to which the device belongs based on the current position of the device; S2. When the device is located in the safe area, a low-power rough positioning algorithm is used to output the rough position coordinates of the device and the corresponding rough positioning confidence ; S3. When the device is located in the transition zone, an adaptive fusion algorithm is used to fuse GPS data and IMU data to obtain an intermediate fusion position , an intermediate fusion speed and an intermediate positioning confidence ; S4. When the device is located in the critical area, a multi-modal data fusion algorithm is used to fuse the multi-modal data, and the risk fusion position and the risk fusion confidence ; And based on the device acceleration and the distance between the device and the fence boundary, combine the trajectory risk assessment algorithm to predict the probability of the device invading the fence; S5. Based on the multi-source signals generated in S2 - S4, generate the final execution instruction through the preset multi-condition decision logic.

2. The GEO fence triggering method for multi-modal positioning data fusion according to claim 1, wherein: The specific method involved in the hierarchical positioning strategy is: Based on the low-power rough positioning algorithm, calculate the current position coordinates of the device, and combine the fence boundary configuration to calculate the shortest boundary distance between the current position of the device and the fence boundary; Set the safety transition threshold and the critical transition threshold respectively based on the shortest boundary distance and the critical transition threshold ; When the shortest boundary distance is less than the safe transition threshold, it is designated as a safe area; When at the safety transition threshold Shortest boundary distance When at the safety transition threshold, it is designated as the transition zone; When the shortest boundary distance is less than the critical transition threshold, it is defined as the critical area; Among them, when the device distance exactly falls near the safety transition threshold or the critical transition threshold, a switching suppressor is introduced to maintain the current positioning mode after crossing the threshold. seconds.

3. The GEO fence triggering method for multi-modal positioning data fusion according to claim 2, wherein: The low-power rough positioning algorithm is specifically: Scan the list of visible base stations around and record the corresponding received signal strength indication values; Perform short-term smoothing on the original RSSI and introduce an interference factor , and obtain the smoothed value; The smoothed value after interference correction is introduced into the path loss model and the distance between the base station and the device is obtained by correction through the interference amplification factor ;​ According to the corrected distance , the coordinate solution is carried out by the trilateral positioning algorithm, and the rough position coordinates of the device are output ; Calculate the rough positioning reliability based on the signal variance, interference variance, and geometric condition number .

4. The GEO fence triggering method for multi-modal positioning data fusion according to claim 1, characterized in that: In S3, the intermediate position information and fusion confidence are obtained by the adaptive fusion algorithm, and the specific steps involved are: Collect the device's longitude and latitude through the GPS module , speed and the variance of the positioning accuracy ; Obtain the acceleration of the device from the IMU module , speed , angular velocity and its own noise variance ; Based on acceleration and angular velocity , the relative displacement is calculated by the Newton-Euler equation ; Convert the IMU acceleration to global coordinates , and based on the global coordinates predict the position and velocity at the current moment By predicting the state variance matrix and the GPS observation noise calculate the Kalman gain ; The predicted position of the IMU and velocity are weighted and fused with the GPS observed position and velocity according to the Kalman gain to output the intermediate fused position and the intermediate fused velocity ; Introduce the longitude and latitude of the device , based on the variance of positioning accuracy , calculate the GPS fusion weight ; Based on the noise variance and the positioning accuracy variance , and introducing a geometric distribution factor, calculate the intermediate positioning confidence ; According to the real-time distance between the device and the fence boundary and the intermediate positioning confidence , dynamically adjust the detection frequency , and introduce the intermediate positioning confidence to perform a secondary adjustment on the detection frequency by introducing a frequency smoother and restricting ; Based on the combined speed and acceleration , predict the displacement within the future ; If the predicted position enters the fence, an alarm is triggered in advance.

5. The GEO fence triggering method for multi-modal positioning data fusion according to claim 1, wherein: The specific steps involved in the multi-modal data fusion algorithm are: Take the intermediate fusion position as the basic predicted value, and introduce the rough position coordinates , and calculate the position deviation ; Set the deviation threshold , if the position deviation is greater than the deviation threshold , then correct the intermediate fusion position to obtain the corrected position coordinates ; Take as the third type of observation, use the observation matrix and the observation noise covariance , and substitute them into the Kalman update; Synchronous calculation of the Kalman gain , and update the state and covariance matrix of the Kalman filter; Then the first component of the state of the updated Kalman filter is the final output of the risk-fused position , which is obtained by back-calculating the risk-fused confidence from the diagonal elements of the covariance . .

6. The GEO fence triggering method for multimodal positioning data fusion according to claim 5, characterized in that: The specific steps involved in the trajectory risk assessment algorithm predicting the probability of the device invading the fence are: Based on the current risk fusion location and the intermediate fusion speed , construct the cloud state vector ; Based on acceleration and angular velocity , the displacement in the predicted future time is calculated through the Newton-Euler equation ; Introduce the acceleration noise variance and the angular velocity noise variance , and perform first-order linear error propagation to obtain the predicted displacement variance ; Based on the future Displacement within a certain period of time And the predicted displacement variance The probability that the device invades the fence is calculated to be ; Set a preset threshold , if the probability of the device invading the fence is , and exceeds the suppression hold time , then an intrusion alarm is output.

7. The GEO fence triggering method for multi-modal positioning data fusion according to claim 1, wherein: The multi-condition decision logic is specifically: Based on the multi-source signals generated in the above steps S2 - S4, generate the final decision logic: When the device is in the safe area, only focus on the rough positioning confidence , set the high confidence threshold for the safe area , if , then maintain the current low-power positioning mode without switching; if , then enter the positioning mode in the transition area and enter the medium-frequency fusion mode; When the device is in the transition zone, pay attention to the intermediate fixed position confidence , set the high confidence threshold for the transition zone and the low confidence threshold for the transition zone , if , then maintain the medium frequency fusion mode; If and , then trigger re-evaluation of the intermediate positioning confidence after multi-source verification ; If , then force a fallback to the safe zone mode and reduce to the low-power positioning mode; When the device is in the critical area, pay attention to the risk fusion confidence and the intrusion probability , the preset intrusion probability threshold , the high-confidence alarm threshold for the critical area and the low-confidence suppression threshold for the critical area ; If ≥ and ≥ , output the high-priority positioning mode; If However , an early warning mode is output and auxiliary positioning is enabled; If the above two conditions are not met when the device is in the critical area, a weighted score is introduced , if the weighted score is greater than the comprehensive critical threshold, the early warning mode is entered; otherwise, the routine monitoring of the critical area is maintained.

8. A GEO fence trigger system for multi-modal positioning data fusion, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes the computer program to implement the GEO fence triggering method for multi-modal positioning data fusion as described in any one of claims 1 - 7.

Images