Sensor jump point detection method, device, equipment and medium

Abstract

The invention provides a sensor jump point detection method and device, equipment and a medium, and relates to the technical field of high-precision positioning, and the method comprises the steps: obtaining sensor data collected by a to-be-detected sensor, and obtaining the state information of the to-be-detected sensor when the to-be-detected sensor collects the sensor data; determining a dynamically adjustable threshold using the state information; detecting whether the to-be-detected sensor has a jump point or not based on the sensor data and the dynamic adjustable threshold value through a pre-constructed sensor jump point detection model set, and obtaining an initial jump point detection result output by each sensor jump point detection model in the sensor jump point detection model set; and determining a target jump point detection result corresponding to the to-be-detected sensor according to the initial jump point detection result output by each sensor jump point detection model. According to the method, the sensor jump point detection capability under various speed change scenes can be remarkably improved, and the robustness and the sensitivity of jump point detection are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of high-precision positioning technology, and in particular to a sensor jump point detection method, device, equipment and medium. Background Art

[0002] Sensor jump point phenomenon refers to the sudden change or error of sensor observation value caused by some unexpected factors. Taking RTK (Real-time kinematic, real-time dynamic carrier phase difference technology) as an example, although RTK technology is very mature, in some specific scenarios, such as scenarios with large altitude changes and fast speed changes, the RTK jump point detection accuracy and real-time performance are limited, and RTK jump points cannot be effectively identified, thus affecting positioning accuracy and system performance. Summary of the invention

[0003] In view of this, an object of the present invention is to provide a sensor jump point detection method, device, equipment and medium, which can significantly improve the sensor jump point detection capability in various speed change scenarios and enhance the robustness and sensitivity of the jump point detection.

[0004] In a first aspect, the present invention provides a sensor jump point detection method, comprising: Acquire sensor data collected by the sensor to be detected, and acquire status information of the sensor to be detected when collecting the sensor data; Determining a dynamically adjustable threshold using the state information; Through the pre-built sensor jump point detection model set, based on sensor data and dynamically adjustable thresholds, whether the sensor to be detected has a jump point is detected, and the initial jump point detection result output by each sensor jump point detection model in the sensor jump point detection model set is obtained; According to the initial jump point detection result output by each sensor jump point detection model, the target jump point detection result corresponding to the sensor to be detected is determined.

[0005] In one embodiment, the state information includes sensor speed information and sensor speed change rate; and determining the dynamically adjustable threshold using the state information includes: Determine the real-time observation standard deviation corresponding to the sensor to be detected according to the observation accuracy range and sensor speed information corresponding to the sensor to be detected; The dynamically adjustable threshold is determined based on the real-time observed standard deviation; or, the dynamically adjustable threshold is determined based on the real-time observed standard deviation and the sensor speed change rate.

[0006] In one embodiment, the state information further includes a sensor update frequency; determining a dynamically adjustable threshold based on a real-time observed standard deviation and a sensor speed change rate includes: The sensor speed change rate is corrected using the sensor update frequency to obtain the true sensor speed change rate; A dynamically adjustable threshold is determined based on the real-time observed standard deviation and the true sensor speed change rate.

[0007] In one embodiment, a sensor jump detection model set is pre-built to detect whether a sensor to be detected has a jump point based on sensor data and a dynamically adjustable threshold, including: For any sensor jump point detection model in the pre-built sensor jump point detection model set, use the sensor jump point detection model to perform the following operations: Insert the sensor data to the end of the sensor data sequence, control the counter to increase, and when the value of the counter is greater than or equal to the size of the sliding window, set the value of the counter to the size of the sliding window; Control the sliding window to slide on the sensor data sequence; determining a reference benchmark based on historical sensor data within the sliding window after the slide; Based on the deviation between the sensor data and the reference benchmark and the dynamically adjustable threshold, the sensor to be detected is tested for whether a jump point occurs, and an initial jump point detection result output by the sensor jump point detection model is obtained.

[0008] In one implementation, the sensor jump point detection model is a sliding window difference detection model; based on the deviation between the sensor data and the reference benchmark and the dynamically adjustable threshold, whether the sensor to be detected has a jump point is detected, and an initial jump point detection result output by the sensor jump point detection model is obtained, including: If the deviation between the sensor data and the reference datum is greater than the dynamically adjustable threshold, then the initial jump point detection result is determined to be a probability value of a jump point occurring in the sensor to be detected being greater than a preset probability threshold.

[0009] In one implementation, the sensor jump point detection model is an accumulation and detection model; based on the deviation between the sensor data and the reference benchmark and the dynamically adjustable threshold, whether the sensor to be detected has a jump point is detected, and an initial jump point detection result output by the sensor jump point detection model is obtained, including: Based on a dynamically adjustable threshold, the deviation between the sensor data and the reference benchmark is accumulated in the positive direction and the negative direction respectively to obtain the positive direction cumulative sum and the negative direction cumulative sum; If any of the cumulative sum in the positive direction and the cumulative sum in the negative direction is greater than the dynamically adjustable threshold, it is determined that the initial jump point detection result is that the probability value of the jump point of the sensor to be detected is greater than the preset probability threshold.

[0010] In one implementation, determining a target jump point detection result corresponding to the sensor to be detected according to an initial jump point detection result output by each sensor jump point detection model includes: When the initial jump point detection results output by each sensor jump point detection model are all that the probability value of the sensor to be detected having a jump point is greater than a preset probability threshold, it is determined that the target jump point detection result is that the sensor to be detected has a jump point.

[0011] In a second aspect, the present invention further provides a sensor jump point detection device, comprising: An acquisition module, for acquiring sensor data collected by the sensor to be detected, and acquiring status information of the sensor to be detected when collecting the sensor data; A threshold adjustment module, used to determine a dynamically adjustable threshold using state information; A jump point detection module is used to detect whether a jump point occurs in the sensor to be detected based on the sensor data and a dynamically adjustable threshold through a pre-built sensor jump point detection model set, and obtain an initial jump point detection result output by each sensor jump point detection model in the sensor jump point detection model set; The result determination module is used to determine the target jump point detection result corresponding to the sensor to be detected according to the initial jump point detection result output by each sensor jump point detection model.

[0012] In a third aspect, the present invention further provides an electronic device, comprising a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement any one of the methods provided in the first aspect.

[0013] In a fourth aspect, the present invention further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement any one of the methods provided in the first aspect.

[0014] The present invention provides a sensor jump point detection method, device, equipment and medium, which first obtains sensor data collected by the sensor to be detected, and obtains state information when the sensor to be detected collects sensor data, so as to determine a dynamically adjustable threshold value by using the state information; then, through a pre-constructed sensor jump point detection model set, based on the sensor data and the dynamically adjustable threshold value, detects whether a jump point occurs in the sensor to be detected, and obtains the initial jump point detection result output by each sensor jump point detection model in the sensor jump point detection model set; finally, according to the initial jump point detection result output by each sensor jump point detection model, determines the target jump point detection result corresponding to the sensor to be detected. The above method uses the state information when the sensor to be detected collects sensor data to adaptively determine the dynamic adjustable threshold value, so as to improve the sensor jump point detection capability in various speed change scenarios, and comprehensively determines the target jump point detection result according to the initial jump point detection result output by each sensor jump point detection model in the sensor jump point detection model set for the sensor to be detected, which helps to enhance the robustness and sensitivity of jump point detection.

[0015] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0016] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A schematic diagram of a flow chart of a sensor jump point detection method provided by an embodiment of the present invention; Figure 2 A flowchart of an RTK jump point detection method provided by an embodiment of the present invention; Figure 3 A comprehensive jump point detection logic diagram provided by an embodiment of the present invention; Figure 4 A logical diagram for initializing or updating a sliding window provided by an embodiment of the present invention; Figure 5A sliding window differential jump point detection logic diagram provided by an embodiment of the present invention; Figure 6 A CUSUM jump point detection logic diagram provided by an embodiment of the present invention; Figure 7 A schematic diagram of the structure of a sensor jump point detection device provided by an embodiment of the present invention; Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described in combination with the embodiments below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] At present, the existing technology mainly adopts the jump detection method with fixed threshold, which has the following main problems: (1) Fixed threshold limitation: It is difficult to adapt to different speed change scenarios. For example, from hovering to full-speed lifting, the position change rate is greatly different, resulting in a decrease in jump detection accuracy; (2) Lack of multi-method fusion: A single method is not robust enough to noise and short-term jumps, which can easily cause false alarms or missed alarms.

[0021] Based on this, the present invention implements a sensor jump point detection method, device, equipment and medium, which can significantly improve the sensor jump point detection capability in various speed change scenarios and enhance the robustness and sensitivity of jump point detection.

[0022] To facilitate understanding of this embodiment, a sensor jump point detection method disclosed in an embodiment of the present invention is first described in detail. Figure 1 The flowchart of a sensor jump point detection method shown in FIG. 1 mainly includes the following steps S102 to S108: Step S102, obtaining sensor data collected by the sensor to be detected, and obtaining state information of the sensor to be detected when collecting the sensor data. The state information includes at least sensor speed information, and may also include sensor speed change rate, sensor update frequency, etc.

[0023] Step S104: Determine a dynamically adjustable threshold using the state information.

[0024] The dynamically adjustable threshold refers to a critical parameter that changes in real time according to current sensor speed information (such as hovering, full speed increase, etc.) of the sensor to be detected. The critical parameter is called a jump point detection threshold.

[0025] In one example, the dynamically adjustable threshold value can be determined based only on the sensor speed information; in another example, the dynamically adjustable threshold value can be determined in combination with the sensor speed information and the sensor speed change rate; in another example, the sensor speed change rate can be corrected using the sensor update frequency, and then the dynamically adjustable threshold value can be determined in combination with the sensor speed information and the corrected sensor speed change rate.

[0026] Step S106, using the pre-built sensor jump point detection model set, based on the sensor data and the dynamically adjustable threshold, detect whether the sensor to be detected has a jump point, and obtain the initial jump point detection result output by each sensor jump point detection model in the sensor jump point detection model set. The sensor jump point detection model set includes at least a sliding window difference detection model and a cumulative sum (CUSUM) detection model, and may also include other detection models, such as a machine learning-based detection model, a deep neural network-based detection model, etc.; the initial jump point detection result is used to characterize whether the probability value of the sensor to be detected having a jump point is greater than a preset probability threshold, that is, whether a jump point is suspected to have occurred.

[0027] In one embodiment, after acquiring the latest sensor data, it can be inserted into the end of the existing sensor data sequence. Each sensor jump point detection model in the sensor jump point detection model set can detect whether a jump point occurs in the sensor to be detected based on the sensor data, the sensor data sequence and the dynamically adjustable threshold, and output the corresponding initial jump point detection result.

[0028] Step S108, determining a target jump point detection result corresponding to the sensor to be detected according to the initial jump point detection result output by each sensor jump point detection model, wherein the target jump point detection result is used to indicate whether the sensor to be detected has a jump point.

[0029] In one embodiment, when the initial jump point detection results output by each sensor jump point detection model indicate that the sensor to be detected is suspected of having a jump point, it is determined that the sensor to be detected has a jump point. In another embodiment, a fixed or adjustable confidence level may be configured for each sensor jump point detection model, and the initial jump point detection results output by each sensor jump point detection model are comprehensively considered based on the confidence level to determine whether a jump point has occurred in the detection sensor.

[0030] The sensor jump point detection method provided in an embodiment of the present invention utilizes the state information of the sensor to be detected when collecting sensor data to adaptively determine a dynamically adjustable threshold value, so as to improve the sensor jump point detection capability in various speed change scenarios, and comprehensively determines the target jump point detection result based on the initial jump point detection result output by the sensor to be detected by each sensor jump point detection model in the sensor jump point detection model set, which helps to enhance the robustness and sensitivity of jump point detection.

[0031] For ease of understanding, an embodiment of the present invention provides a specific implementation of a sensor jump point detection method.

[0032] For the aforementioned step S104, the embodiment of the present invention provides a specific implementation method of using state information to determine a dynamically adjustable threshold, including the following steps 1a to 1b: Step 1a, according to the observation accuracy range and sensor speed information of the sensor to be detected, determine the real-time observation standard deviation corresponding to the sensor to be detected. The observation accuracy range includes the accuracy of the sensor when it is stationary. and the accuracy of the sensor at high speeds .

[0033] Sensor speed information can be used as a reflection of the signal drift rate and used to dynamically adjust the threshold (That is, the aforementioned dynamically adjustable threshold). In the actual use of the sensor, different working conditions may be involved; taking RTK as an example, the judgment threshold of the jump under static working conditions and the judgment threshold of RTK under highly mobile working conditions (such as 15m / s high-speed flight of a drone) should obviously not be the same number, and the latter should be much larger than the former. If the speed is low but the position data has changed significantly, then this jump is more likely to be abnormal; if the speed is high, the position change may be the natural result of the movement itself and should not be misjudged as a jump.

[0034] Similarly, in sensor jump point detection models (such as CUSUM detection models), RTK speed information can be used to adjust the cumulative sum change rate. For example, when the RTK speed is high, the cumulative sum change rate is allowed to be faster to adapt to the violent motion environment; when the speed is low, the change is required to be smoother to prevent false detection.

[0035] In the present invention, the threshold The method is based on the accuracy characteristics of RTK itself. When stationary, its accuracy is at the centimeter level. Assume that the accuracy in the RTK hardware manual is ; When moving at high speed, its accuracy is ; then according to the speed from 0 to , within the range of observation accuracy arrive Establish a mapping relationship to obtain the real-time observation standard deviation : ; in, is the real-time observation standard deviation, is the observation accuracy when the sensor is stationary, is the observation accuracy of the sensor when it moves at high speed, is the maximum speed of the sensor movement, is the real-time speed of the sensor movement.

[0036] Step 1b, determining a dynamically adjustable threshold based on a real-time observed standard deviation; or determining a dynamically adjustable threshold based on a real-time observed standard deviation and a sensor speed change rate.

[0037] Method 1: Based only on real-time observation standard deviation Determine the dynamically adjustable threshold , real-time observation standard deviation is the real-time calculation threshold Optional, dynamically adjustable thresholds Standard deviation from real-time observation The relationship between is based on the hypothesis verification theory. Hypothesis verification theory is a method used in statistics to determine whether sample data supports a hypothesis (called the null hypothesis H). Its core idea is to use probability to evaluate the possibility of random errors, ensure that the conclusion is statistically significant, and avoid making wrong inferences due to sample bias. It compares the deviation between the sample data and the expected value to evaluate whether the deviation is large enough to reject the null hypothesis. It usually includes the following steps: 1. Propose a hypothesis: Set the null hypothesis H0 and the alternative hypothesis H1. H0 usually indicates no significant difference or a default state, and H1 indicates the conclusion that the researcher hopes to prove. In the embodiment of the present invention, the null hypothesis H0 is that the sensor is normal, and the alternative hypothesis H1 is that the sensor has a jump point.

[0038] 2. Select the test method and significance level: Determine the type of statistical test (such as t-test, z-test) and the significance level α, usually 0.05 or 0.01.

[0039] 3. Calculate the test statistic: Use the sample data to calculate the test statistic (such as z-value, t-value) and determine its probability under the assumed distribution.

[0040] 4. Determine the rejection region: Based on the significance level, find the critical value of the rejection region (such as the z value of the normal distribution).

[0041] 5. Make a decision: If the test statistic falls into the rejection region, reject H0 and accept H1; if the test statistic is not in the rejection region, H0 cannot be rejected.

[0042] In specific implementation, since the sensor value may jump in both positive and negative directions, the embodiment of the present invention adopts a two-tailed test. The embodiment of the present invention exemplarily provides a significant level as shown in Table 1, A mapping table of Z-score and rejection region range.

[0043] Table 1 Significant level, Mapping table of value (Z-score) and rejection region range

[0044] Based on the above, since RTK is a high-precision sensor, the real-time observation standard deviation The significance level is 0.1%, then the threshold can be adjusted dynamically Standard deviation from real-time observation The relationship between them is: .

[0045] in, Represents the real-time observation standard deviation Corresponding For example, at a significance level of 0.1%, determine the real-time observed standard deviation Corresponding The value (Z-score) is 3.29, which means that The embodiment of the present invention can achieve the following effects through the above method: the sensitivity of detection is set according to the sensor speed information (for example, a larger change is allowed at a higher speed); if the sensor speed is close to zero but a significant jump is detected, it may be a signal abnormality, and a jump point is determined; if the sensor speed is large, the signal change may be caused by motion and should not be misjudged, thereby improving the reliability of jump point detection.

[0046] Method 2: Determine the dynamically adjustable threshold based on the real-time observed standard deviation and the sensor speed change rate. In the specific implementation, first determine the initial dynamically adjustable threshold according to the aforementioned method 1, and then adjust the initial dynamically adjustable threshold using the sensor speed change rate to obtain the final dynamically adjustable threshold. For example, determine the product of the sensor speed change rate and the fault tolerance parameter (set to 2.0), and use the sum of the product and the initial dynamically adjustable threshold as the final dynamically adjustable threshold. Therefore, the dynamically adjustable threshold The expression is as follows: ; in, is the sensor speed change rate.

[0047] Furthermore, the sensor speed change rate can be corrected using the sensor update frequency to obtain the true sensor speed change rate, and then the dynamically adjustable threshold is determined based on the real-time observation standard deviation and the true sensor speed change rate. For example, the sensor speed change rate can be corrected according to the following formula: : ;in, is the sensor update frequency. On this basis, the final dynamically adjustable threshold is obtained according to the above method 2.

[0048] For the aforementioned step S106, the embodiment of the present invention provides a specific implementation method of detecting whether a sensor to be detected has a jump point based on sensor data and a dynamically adjustable threshold through a pre-constructed sensor jump point detection model set, and obtaining an initial jump point detection result output by each sensor jump point detection model. For any sensor jump point detection model in the pre-constructed sensor jump point detection model set, the sensor jump point detection model is used to perform the following steps 2a to 2d: Step 2a, inserting the sensor data to the end of the sensor data sequence, controlling the counter to increase, and when the value of the counter is greater than or equal to the size of the sliding window, setting the value of the counter to the size of the sliding window; Step 2b, controlling the sliding window to slide on the sensor data sequence; Step 2c, determining a reference benchmark based on historical sensor data located in the sliding window after sliding; Step 2d: Based on the deviation between the sensor data and the reference benchmark and the dynamically adjustable threshold, detect whether the sensor to be detected has a jump point, and obtain an initial jump point detection result output by the sensor jump point detection model.

[0049] In one example, when the sensor jump point detection model is a sliding window differential detection model, if the deviation between the sensor data and the reference benchmark is greater than a dynamically adjustable threshold, the initial jump point detection result is determined to be a probability value of a jump point occurring in the sensor to be detected that is greater than a preset probability threshold.

[0050] The sliding window differential detection model is used to detect the deviation between the current position and the sliding window mean, and quickly determine whether a significant jump occurs. It detects whether the new sensor data has a jump by comparing the deviation between the average value (or median) of the data in the sliding window and the new sensor data. The principle of the sliding window differential detection model is: the sliding window saves the sensor data sequence of the most recent period; calculates the average value (or median) of the sensor data sequence in the sliding window as a reference benchmark; calculates the deviation between the new sensor data and the reference benchmark; if the deviation exceeds the dynamically adjustable threshold, it is determined that the sensor to be detected is suspected of having a jump. The sliding window differential detection model has the advantages of being simple and easy to implement, small amount of calculation, and sensitive to the detection of sudden changes with a single large change. In addition, the sliding window differential detection model is more sensitive to signal noise, so it is easily affected by single-point outliers, and it is impossible to distinguish whether the sudden change is a single anomaly or a trend change. Therefore, the sliding window differential detection model is suitable for detecting single emergencies or short-term outliers. Among them, the preset probability threshold can be 0. When the probability value of the sensor to be detected having a jump is greater than the preset probability threshold, it can be determined that the sensor to be detected is suspected of having a jump.

[0051] Based on this, the embodiment of the present invention provides a specific implementation method of using a sliding window difference detection model to detect whether a sensor has a jump point, including: 1. Use a sliding window of length WINDOW_SIZE to calculate the mean of the data in the window , the mean The reference benchmark is: ; 2. For new sensor data With the mean The difference between To perform the test: ; 3. Determine whether the difference exceeds the dynamically adjustable threshold: ; If the difference exceeds the dynamically adjustable threshold, it is determined that the probability value of the sensor to be detected having a jump point is greater than the preset probability threshold; otherwise, it is determined that the sensor to be detected has not had a jump point.

[0052] In one example, when the sensor jump point detection model is a cumulative sum detection model, based on a dynamically adjustable threshold, the deviation between the sensor data and the reference benchmark is accumulated in the positive direction and the negative direction, respectively, to obtain the positive direction cumulative sum and the negative direction cumulative sum; if any of the positive direction cumulative sum and the negative direction cumulative sum is greater than the dynamically adjustable threshold, then the initial jump point detection result is determined to be that the probability value of a jump point occurring in the sensor to be detected is greater than a preset probability threshold.

[0053] The cumulative sum detection model is used to further detect the accumulated deviation and filter out the influence of noise, thereby improving the reliability and robustness of jump detection. It detects the change trend and jump of new sensor data by accumulating the deviation in the positive and negative directions. The principle of the cumulative sum detection model is as follows: calculate the deviation between the new sensor data and the reference mean in the sliding window; accumulate the deviation in the positive direction and the negative direction respectively (called positive cumulative sum and negative cumulative sum); if any cumulative sum exceeds the set threshold, it is determined that the signal has jumped; once the jump is detected, reset the cumulative sum so that the subsequent signal can be detected again. The cumulative sum detection model has the advantages of being more sensitive to trend changes (such as the signal gradually deviating from the reference benchmark), more robust to noise, and able to filter out the influence of single-point outliers. However, compared with the sliding window difference model, the cumulative sum detection model has a slightly larger amount of calculation, and needs to reasonably set the deviation parameters and a constant, and the adjustment is more complicated. Therefore, the cumulative sum detection model can be applied to: detecting trend changes or long-term drift of the signal, and scenes with large noise or stable trend changes.

[0054] Based on this, the embodiment of the present invention provides a specific implementation method of using the cumulative sum detection model to detect whether a sensor has a jump point, including: 1. Update the cumulative sum in the positive direction and the cumulative sum in the negative direction: cusumPos=max(0,cusumPos+(newData−meanValue)− ); cusumNeg=max(0,cusumNeg+(meanValue−(newData+ ); Among them, cusumPos and cusumNeg are positive cumulative sum and negative cumulative sum respectively, newData is the new sensor data, meanValue is the reference benchmark, A dynamically adjustable threshold.

[0055] 2. Jump compensation: If the jump detection function returns true and the previous and next frames are in a fixed solution state, the jump compensation value is calculated: compensation=(window[2]−window[1]); Among them, window[2] is the sensor data of the previous frame, and window[1] is the new sensor data (that is, the sensor data of the current frame); Limit the compensation value range: compensation=LIMIT(compensation,−1000.0,1000.0); The new sensor data is thus compensated using the compensation value.

[0056] For the aforementioned step S108, an embodiment of the present invention provides an implementation method for determining a target jump point detection result corresponding to a sensor to be detected based on an initial jump point detection result output by each sensor jump point detection model, that is: when the initial jump point detection result output by each sensor jump point detection model is that the probability value of a jump point occurring in the sensor to be detected is greater than a preset probability threshold, the target jump point detection result is determined to be a jump point occurring in the sensor to be detected, which helps to avoid misjudgment by a single sensor jump point detection model.

[0057] The sensor jump point detection method provided in the embodiment of the present invention can be used to detect the inevitable jump point situations of important sensors on drones, such as low RTK satellite number or RTK interference. The method has the following characteristics: providing RTK jump change detection capability adapting to multiple scenarios, from low-speed hovering to high-speed ascent and descent (such as 10m / s); innovatively combining the sliding window difference method and the CUSUM method to enhance the robustness and sensitivity of jump change detection; designing an adaptive mechanism to adjust the critical parameters of jump change detection in real time to adapt to different speed change scenarios.

[0058] For example, the embodiment of the present invention takes RTK as an example to provide an application example of a sensor jump point detection method, see Figure 2 The flowchart of a RTK jump point detection method shown in the figure includes the following (i) to (ix): 1. Input the RTK data rtk_data_height and RTK update frequency rtk_data_vel_down.

[0059] (ii) Determine whether initialization is complete; if not, execute (iii) and skip subsequent checks, and the process ends; if yes, execute (iv).

[0060] (3) Call the sensor change point detection initialization program detect_change_point_init to skip subsequent detection.

[0061] 4. Calculating the real sensor change rate : .

[0062] (V) Call the sensor jump point detection model detect_change_poin to determine whether a jump is detected; if yes, execute (VII); if not, execute (VI).

[0063] The embodiment of the present invention provides a specific implementation method of calling the sensor jump point detection model detect_change_poin to determine whether a jump is detected. Figure 3 A comprehensive jump point detection logic diagram shown includes the following steps: initializing or updating the sliding window; calculating the sliding window difference diff; judging whether the sliding window detects a jump; if so, determining that the sliding window detects a jump; if not, determining that the sliding window does not detect a jump; calculating the CUSUM cumulative sum; judging whether the CUSUM method detects a jump; if so, determining that the CUSUM method detects a jump; if not, determining that the CUSUM method does not detect a jump; if both methods detect a jump, returning true; if either method does not detect a jump, returning false.

[0064] Furthermore, the embodiment of the present invention also provides a specific implementation method for initializing or updating the sliding window, see Figure 4 A logic diagram for initializing or updating a sliding window is shown, comprising the following steps: inputting new sensor data newData; determining whether the sliding window is full; if so, returning true; if not, inserting the new sensor data newData into the sliding window and increasing the initialization counter; determining whether the value of the counter is less than the window size; if so, returning false; if not, setting the value of the counter to the window size and returning true.

[0065] Furthermore, the embodiment of the present invention also provides Figure 5 The sliding window differential jump point detection logic diagram shown in the figure includes the following steps: inputting new sensor data newData and sensor speed change rate ; Determine the dynamically adjustable threshold ; Calculate the mean in the sliding window ; Calculate the sensor data newData and the mean in the sliding window The difference between the two values ​​is diff; determine whether the difference diff is greater than the dynamically adjustable threshold ; If yes, the sliding window detects a jump and returns true; if no, the sliding window does not detect a jump and returns false.

[0066] Furthermore, the embodiment of the present invention also provides Figure 6 The CUSUM jump point detection logic diagram shown in the figure includes the following steps: inputting new sensor data newData and sensor speed change rate ; Determine dynamically adjustable thresholds ; Calculate the mean in the sliding window ; Update the cumulative sum cusumPos, cusumNeg; judge cusumPos> orcusumNeg< ; If yes, reset cusumPos and cusumNeg, confirm that the CUSUM method detects a jump, and return true; if no, confirm that the CUSUM method does not detect a jump, and return false.

[0067] (vi) Transition detection is not triggered.

[0068] 7. Calculate compensation.

[0069] (VIII) Constraint compensation value range.

[0070] 9. Update the fixed solution rtk_data_fix_last and the process ends.

[0071] In summary, the sensor jump point detection method provided by the embodiment of the present invention has at least the following characteristics: multi-scenario adaptability: the adaptive mechanism adjusts the jump detection parameters in real time to meet the needs of different speed change scenarios; high detection accuracy: the sliding window difference method quickly responds to single-frame jumps, and the CUSUM method captures trend changes, reducing the missed detection rate and false detection rate under double protection; real-time: the algorithm optimizes the processing speed and adapts to high-frequency (10Hz) RTK data streams to meet real-time requirements; enhanced robustness: the fault-tolerant mechanism is combined with noise processing to improve the robustness of the algorithm in complex environments.

[0072] Based on the above embodiments, the present invention provides a sensor jump point detection device. Figure 7 The schematic diagram of the structure of a sensor jump point detection device shown in FIG. 1 mainly includes the following parts: An acquisition module 702 is used to acquire sensor data collected by the sensor to be detected, and to acquire state information of the sensor to be detected when collecting the sensor data; A threshold adjustment module 704, configured to determine a dynamically adjustable threshold using the state information, where the dynamically adjustable threshold is a jump point detection threshold; A jump point detection module 706 is used to detect whether a jump point occurs in the sensor to be detected based on the sensor data and the dynamically adjustable threshold through a pre-built sensor jump point detection model set, and obtain an initial jump point detection result output by each sensor jump point detection model in the sensor jump point detection model set; The result determination module 708 is used to determine the target jump point detection result corresponding to the sensor to be detected according to the initial jump point detection result output by each sensor jump point detection model.

[0073] The sensor jump point detection device provided in an embodiment of the present invention utilizes the state information of the sensor to be detected when collecting sensor data to adaptively determine a dynamically adjustable threshold value to improve the sensor jump point detection capability in various speed change scenarios, and comprehensively determines the target jump point detection result based on the initial jump point detection result output by the sensor to be detected by each sensor jump point detection model in the sensor jump point detection model set, which helps to enhance the robustness and sensitivity of jump point detection.

[0074] In one implementation, the state information includes sensor speed information and sensor speed change rate; the threshold adjustment module 704 includes: A standard deviation determination unit, used to determine a real-time observation standard deviation corresponding to the sensor to be detected according to an observation accuracy range and sensor speed information corresponding to the sensor to be detected; The threshold determination unit is used to: determine the dynamically adjustable threshold based on the real-time observation standard deviation; or determine the dynamically adjustable threshold based on the real-time observation standard deviation and the sensor speed change rate.

[0075] In one implementation, the state information further includes a sensor update frequency; and the threshold determination unit is specifically configured to: The sensor speed change rate is corrected using the sensor update frequency to obtain the true sensor speed change rate; A dynamically adjustable threshold is determined based on the real-time observed standard deviation and the true sensor speed change rate.

[0076] In one implementation, the jump point detection module 706 is specifically used to: For any sensor jump point detection model in the pre-built sensor jump point detection model set, use the sensor jump point detection model to perform the following operations: Insert the sensor data to the end of the sensor data sequence, control the counter to increase, and when the value of the counter is greater than or equal to the size of the sliding window, set the value of the counter to the size of the sliding window; Control the sliding window to slide on the sensor data sequence; determining a reference benchmark based on historical sensor data within the sliding window after the slide; Based on the deviation between the sensor data and the reference benchmark and the dynamically adjustable threshold, the sensor to be detected is tested for whether a jump point occurs, and an initial jump point detection result output by the sensor jump point detection model is obtained.

[0077] In one implementation, the sensor jump point detection model is a sliding window difference detection model, and the jump point detection module 706 is specifically used to: If the deviation between the sensor data and the reference datum is greater than the dynamically adjustable threshold, then the initial jump point detection result is determined to be a probability value of a jump point occurring in the sensor to be detected being greater than a preset probability threshold.

[0078] In one implementation, the sensor jump point detection model is an accumulation and detection model, and the jump point detection module 706 is specifically used to: Based on a dynamically adjustable threshold, the deviation between the sensor data and the reference benchmark is accumulated in the positive direction and the negative direction respectively to obtain the positive direction cumulative sum and the negative direction cumulative sum; If any of the cumulative sum in the positive direction and the cumulative sum in the negative direction is greater than the dynamically adjustable threshold, it is determined that the initial jump point detection result is that the probability value of the jump point of the sensor to be detected is greater than the preset probability threshold.

[0079] In one implementation, the result determination module 708 is specifically configured to: When the initial jump point detection results output by each sensor jump point detection model are all that the probability value of the sensor to be detected having a jump point is greater than a preset probability threshold, it is determined that the target jump point detection result is that the sensor to be detected has a jump point.

[0080] The device provided in the embodiment of the present invention has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the device embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0081] An embodiment of the present invention provides an electronic device. Specifically, the electronic device includes a processor and a storage device. The storage device stores a computer program, and when the computer program is executed by the processor, it executes the method described in any one of the above-mentioned implementation methods.

[0082] Figure 8 A structural diagram of an electronic device provided in an embodiment of the present invention, the electronic device 100 includes: a processor 80, a memory 81, a bus 82 and a communication interface 83, wherein the processor 80, the communication interface 83 and the memory 81 are connected via the bus 82; the processor 80 is used to execute an executable module stored in the memory 81, such as a computer program.

[0083] The memory 81 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 83 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0084] The bus 82 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0085] Among them, the memory 81 is used to store programs, and the processor 80 executes the program after receiving the execution instruction. The method executed by the device for flow process definition disclosed in any embodiment of the above-mentioned embodiments of the present invention can be applied to the processor 80 or implemented by the processor 80.

[0086] The processor 80 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 80. The above processor 80 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present invention can be directly embodied as a hardware decoding processor to execute, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 81, and the processor 80 reads the information in the memory 81 and completes the steps of the above method in combination with its hardware.

[0087] The computer program product of the readable storage medium provided in the embodiment of the present invention includes a computer-readable storage medium storing program code, and the instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be referred to the previous method embodiments, which will not be repeated here.

[0088] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0089] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A sensor jump point detection method, characterized in that: include: Acquire sensor data collected by the sensor to be detected, and acquire state information of the sensor to be detected when collecting the sensor data, wherein the state information includes sensor speed information and sensor speed change rate; determining a dynamically adjustable threshold using the state information; Using a pre-built sensor jump point detection model set, based on the sensor data and the dynamically adjustable threshold, detecting whether the sensor to be detected has a jump point, and obtaining an initial jump point detection result output by each sensor jump point detection model in the sensor jump point detection model set; Determine a target jump point detection result corresponding to the sensor to be detected according to the initial jump point detection result output by each sensor jump point detection model; Determining a dynamically adjustable threshold using the state information includes: determining a real-time observation standard deviation corresponding to the sensor to be detected based on an observation accuracy range corresponding to the sensor to be detected and the sensor speed information; The dynamically adjustable threshold is determined based on the real-time observed standard deviation, or the dynamically adjustable threshold is determined based on the real-time observed standard deviation and the sensor speed change rate.

2. The sensor jump point detection method according to claim 1, characterized in that: The state information also includes a sensor update frequency; determining a dynamically adjustable threshold based on the real-time observation standard deviation and the sensor speed change rate includes: Correcting the sensor speed change rate using the sensor update frequency to obtain a true sensor speed change rate; A dynamically adjustable threshold is determined based on the real-time observed standard deviation and the true sensor speed change rate.

3. The sensor jump point detection method according to claim 1, characterized in that: By using a pre-built sensor jump point detection model set, based on the sensor data and the dynamically adjustable threshold, detecting whether the sensor to be detected has a jump point, including: For any sensor jump point detection model in the pre-built sensor jump point detection model set, use the sensor jump point detection model to perform the following operations: Inserting the sensor data to the end of the sensor data sequence, controlling a counter to increase, and when the value of the counter is greater than or equal to the size of the sliding window, setting the value of the counter to the size of the sliding window; Controlling the sliding window to slide on the sensor data sequence; Determining a reference benchmark based on historical sensor data within the sliding window after sliding; According to the deviation between the sensor data and the reference benchmark and the dynamically adjustable threshold, whether a jump point occurs in the sensor to be detected is detected to obtain an initial jump point detection result output by the sensor jump point detection model.

4. The sensor jump point detection method according to claim 3, characterized in that: The sensor jump point detection model is a sliding window difference detection model; according to the deviation between the sensor data and the reference benchmark and the dynamically adjustable threshold, whether the sensor to be detected has a jump point is detected, and an initial jump point detection result output by the sensor jump point detection model is obtained, including: If the deviation between the sensor data and the reference datum is greater than the dynamically adjustable threshold, then the initial jump point detection result is determined to be that the probability value of the jump point occurring in the sensor to be detected is greater than a preset probability threshold.

5. The sensor jump point detection method according to claim 3, characterized in that: The sensor jump point detection model is an accumulation and detection model; according to the deviation between the sensor data and the reference benchmark and the dynamically adjustable threshold, whether the sensor to be detected has a jump point is detected, and an initial jump point detection result output by the sensor jump point detection model is obtained, including: Based on the dynamically adjustable threshold, the deviation between the sensor data and the reference datum is accumulated in a positive direction and a negative direction respectively to obtain a positive direction cumulative sum and a negative direction cumulative sum; If any of the cumulative sum in the positive direction and the cumulative sum in the negative direction is greater than the dynamically adjustable threshold, it is determined that the initial jump point detection result is that the probability value of the jump point of the sensor to be detected is greater than the preset probability threshold.

6. The sensor jump point detection method according to claim 1, characterized in that: Determining a target jump point detection result corresponding to the sensor to be detected according to the initial jump point detection result output by each sensor jump point detection model includes: When the initial jump point detection results output by each of the sensor jump point detection models are all those in which the probability value of the jump point of the sensor to be detected is greater than a preset probability threshold, it is determined that the target jump point detection result is that the jump point of the sensor to be detected occurs.

7. A sensor jump point detection device, characterized in that: include: an acquisition module, for acquiring sensor data collected by the sensor to be detected, and acquiring state information of the sensor to be detected when collecting the sensor data, wherein the state information includes sensor speed information and sensor speed change rate; A threshold adjustment module, configured to determine a dynamically adjustable threshold using the state information; A jump point detection module is used to detect whether a jump point occurs in the sensor to be detected based on the sensor data and the dynamically adjustable threshold through a pre-built sensor jump point detection model set, and obtain an initial jump point detection result output by each sensor jump point detection model in the sensor jump point detection model set; A result determination module, configured to determine a target jump point detection result corresponding to the sensor to be detected according to the initial jump point detection result output by each sensor jump point detection model; The threshold adjustment module is specifically used to: determine the real-time observation standard deviation corresponding to the sensor to be detected according to the observation accuracy range corresponding to the sensor to be detected and the sensor speed information; determine the dynamically adjustable threshold based on the real-time observation standard deviation, or determine the dynamically adjustable threshold based on the real-time observation standard deviation and the sensor speed change rate.

8. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the method according to any one of claims 1 to 6.

Images