Power line inspection robot control method based on dynamic event triggering

Abstract

The invention provides a power line inspection robot control method based on dynamic event triggering, and relates to the technical field of power line inspection, comprising the following steps: receiving a first early warning signal comprising a first prediction start moment of a dynamic event, and determining a mapping factor and an observation window period corresponding to the dynamic event; when entering an observation window period, acquiring real-time monitoring data of each mapping factor in real time to obtain a mapping factor set, and judging the event occurrence probability of the dynamic event at the first prediction starting moment according to the mapping factor set; if the event occurrence probability is smaller than a first preset threshold value, controlling the inspection robot to maintain a default inspection mode; and if the event occurrence probability is greater than or equal to a first preset threshold value, controlling the inspection robot to switch from a default inspection mode to an event response mode corresponding to the dynamic event at a first prediction starting moment. According to the invention, unnecessary mode switching can be reduced, so that energy consumption is reduced, and the stability of the inspection system is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of power line inspection, and in particular to a control method of a power line inspection robot based on dynamic event triggering. Background Art

[0002] In the current power system, the safe and stable operation of power lines is of vital importance. In order to ensure the normal working state of power lines, inspection robots are widely used in power line inspection tasks. However, existing power line inspection robots often have some problems when facing dynamic events.

[0003] Current inspection robots mainly rely on weather forecasts or alarms about dynamic meteorological events issued by weather stations to respond to upcoming dynamic events, such as switching to inspection mode when dynamic events occur. However, the accuracy and reliability of warning signals such as weather forecasts are still unclear. For example, the meteorological system may issue warnings of possible strong winds, heavy rains, etc. due to changes in the local microclimate, but in reality these dynamic events may not actually occur.

[0004] In this case, the inspection robot will frequently switch between different control modes according to the warning signal. Each mode switch requires the robot to mobilize different sensors, perform different actions, and adjust operating parameters. In actual applications, due to the uncertainty and frequency of warning signals, the robot performs multiple unnecessary mode switches in a short period of time, resulting in excessive energy consumption. This not only increases the inspection cost, but also affects the robot's endurance and work efficiency. Summary of the invention

[0005] In view of the above defects or deficiencies in the prior art, the present invention aims to provide a control method of a power line inspection robot based on dynamic event triggering, comprising the following steps: receiving a first warning signal of a dynamic event, determining at least one mapping factor corresponding to the dynamic event, and determining an observation window period corresponding to the dynamic event; the mapping factor is a variable reflected on or around the power line before the dynamic event occurs; the first warning signal includes a first predicted start time of the dynamic event; the observation window period is before the first predicted start time; When entering the observation window period, real-time monitoring data of each mapping factor is acquired in real time at intervals of a first preset time length to obtain a mapping factor set; the mapping factor set is composed of real-time monitoring sequences corresponding to a plurality of mapping factors; Determining, according to the mapping factor set, the probability of occurrence of the dynamic event at the first prediction start time; If the probability of the event occurring is less than the first preset threshold, the inspection robot is controlled to maintain the default inspection mode; if the probability of the event occurring is greater than or equal to the first preset threshold, the inspection robot is controlled to switch from the default inspection mode to the event response mode corresponding to the dynamic event at the start time of the first prediction.

[0006] According to the technical solution provided by the present invention, the determining of at least one mapping factor corresponding to the dynamic event specifically includes the following steps: Retrieve and traverse a historical database, wherein the historical database includes multiple characteristic factors of the power line and a historical monitoring sequence composed of multiple historical monitoring data corresponding to each characteristic factor in time sequence; wherein each of the historical monitoring sequences is marked with a historical start time of at least one historical event, and the historical monitoring data in the historical monitoring sequence that is within a second preset time length before the historical start time is marked as the first reference monitoring data; The characteristic factor of each of the first reference monitoring data that is in continuous change within the second preset time period is used as the mapping factor corresponding to the dynamic event of the same type as the historical event.

[0007] According to the technical solution provided by the present invention, the plurality of first reference monitoring data corresponding to each of the mapping factors constitute a first reference monitoring sequence; The determining, according to the mapping factor set, the probability of occurrence of the dynamic event at the first prediction start time specifically comprises the following steps: respectively calculating a first distance of each of the mapping factors, the first distance being a Euclidean distance between the real-time monitoring sequence and the first reference monitoring sequence; The first distances of the mapping factors are weighted to obtain an average value to obtain the event occurrence probability.

[0008] According to the technical solution provided by the present invention, judging the probability of occurrence of the dynamic event at the first prediction start time according to the mapping factor set specifically includes the following steps: According to the mapping factor set, multiple reference sets are found in the historical database, the reference sets including second reference monitoring sequences corresponding to multiple mapping factors respectively; in the reference set and the mapping factor set, the Euclidean distance between the real-time monitoring sequence and the second reference monitoring sequence of the same mapping factor is less than or equal to a second preset threshold; A reference set in the history database that is marked with the history start time within a third preset time period after the reference set and in which the history event corresponding to the history start time is of the same type as the dynamic event is used as a credible set; The ratio of the number of the credible set to the number of the reference set is taken as the probability of the event occurring.

[0009] According to the technical solution provided by the present invention, the determination of the observation window period corresponding to the dynamic event specifically includes the following steps: Based on the time leading interval and duration range of the trusted set relative to the corresponding historical start time, the observation window period corresponding to the dynamic event is obtained.

[0010] According to the technical solution provided by the present invention, after retrieving and traversing the historical database, the following steps are also included: Determine whether the historical database contains the historical event of the same type as the dynamic event; The method specifically includes the following steps: taking the characteristic factor of each of the first reference monitoring data that is in continuous change within the second preset time period as the mapping factor corresponding to the dynamic event of the same type as the historical event: If the historical event of the same type as the dynamic event is matched in the historical database, the characteristic factor of each first reference monitoring data that is in continuous change within the second preset time period is used as the mapping factor corresponding to the dynamic event of the same type as the historical event.

[0011] According to the technical solution provided by the present invention, after determining whether the historical database contains a historical event of the same type as the dynamic event, the following steps are also included: If the historical event of the same type as the dynamic event is not matched in the historical database, then based on principal component analysis, the plurality of characteristic factors in the historical database are converted into a plurality of principal components, each of which is composed of a plurality of corresponding characteristic factors and a weight of each characteristic factor, and the number of the principal components is less than the number of characteristic factors; Relevant components of the dynamic event are obtained, and the principal component matching the relevant components is used as the mapping factor corresponding to the dynamic event.

[0012] According to the technical solution provided by the present invention, the historical database is also marked with the second predicted starting time of each historical event; The method further comprises the following steps: When the dynamic event occurs, the actual start time of the dynamic event is recorded, and the dynamic event is used as a historical event to update the historical database; wherein the actual start time of the dynamic event is used as the historical start time, and the first predicted start time is used as the second predicted start time.

[0013] According to the technical solution provided by the present invention, after receiving the first warning signal of the dynamic event, the following steps are also included: Retrieving and traversing the history database, and determining whether the history database contains the history event of the same type as the dynamic event; If there is at least one historical event of the same type as the dynamic event, the historical event of the same type as the dynamic event is taken as a matching event; Determining the observation window period corresponding to the dynamic event specifically includes the following steps: If the deviation between the historical start time of all the matching events and the second predicted start time is less than or equal to a third preset threshold, the observation window period corresponding to the dynamic event is determined according to the first predicted start time.

[0014] According to the technical solution provided by the present invention, after taking the historical event of the same type as the dynamic event as a matching event, the following steps are also included: If there are at least N matching events whose deviations between the historical start time and the second predicted start time are greater than the third preset threshold, the N matching events are regarded as abnormal matching events, and the matching events other than the N abnormal matching events are regarded as normal matching events; Obtaining a first adjustment amount according to the second predicted start time of the normal matching event, and obtaining a second adjustment amount according to the second predicted start time of the abnormal matching event; The first adjustment amount matching weight is calculated based on the ratio of the normal matching events to all the matching events, and the second adjustment amount matching weight is calculated based on the ratio of the abnormal matching events to all the matching events, so as to obtain a total adjustment amount; Before determining the observation window period corresponding to the dynamic event, the following steps are also included: The first predicted start time is updated with the total adjustment amount.

[0015] Compared with the prior art, the beneficial effect of the present invention is that the present invention uses the mapping factor of the dynamic event to judge the probability that the dynamic event may actually occur during the observation window period, and makes a more accurate judgment and reasonable response to the warning signal, instead of obtaining the warning signal and switching the mode at the predicted moment of the warning signal, thereby reducing unnecessary mode switching and reducing energy consumption. At the same time, it avoids the time waste and system instability caused by frequent mode switching, so that the inspection robot can perform the inspection task more attentively and improve work efficiency. In addition, a more stable working mode also helps to reduce the wear and failure risk of equipment. A reasonable control mode switching strategy can improve the inspection robot's ability to respond to dynamic events, ensure stable operation under normal conditions, and enhance the reliability of the entire power line inspection system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flowchart of the steps of a control method of a power line inspection robot based on dynamic event triggering provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It should also be noted that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0018] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0019] Example 1 As mentioned in the background technology, in order to solve the problems in the prior art, the present invention proposes a control method of a power line inspection robot based on dynamic event triggering, please refer to Figure 1 As shown, the following steps are included: S1. Receive a first warning signal of a dynamic event, determine at least one mapping factor corresponding to the dynamic event, and determine an observation window period corresponding to the dynamic event; the mapping factor is a variable reflected on or around the power line before the dynamic event occurs; the first warning signal includes a first predicted start time of the dynamic event; the observation window period is before the first predicted start time; Specifically, dynamic events can be meteorological events (strong winds, heavy rains, lightning, snowstorms, etc.), human-related events (construction activities, man-made damage, etc.) or biological events (bird activities, large animal activities). For different dynamic events, variables that may be reflected on or around power lines before they occur are used as mapping factors. Exemplarily, for the dynamic event of strong winds, the mapping factors may include wind speed, wind direction, power line sway amplitude, sway frequency, etc.; for man-made construction events, the mapping factors may include construction noise intensity, etc. For example, the first warning signal of a dynamic event such as a meteorological-related event is issued by a meteorological station, and the first warning signal also includes the dynamic event type, such as strong wind weather, and the first prediction start time is the start time of the strong wind weather.

[0020] Specifically, the observation window period is before the first predicted start time of the dynamic event, so that there is enough time for monitoring and judgment. For example, for a strong wind event, the observation window period can be set to 1 hour before the predicted strong wind arrives.

[0021] Specifically, the entire inspection system establishes connections with external data sources such as meteorological departments and construction monitoring systems to receive the first warning signal of dynamic events in a timely manner. The first warning signal should include the first predicted start time of the dynamic event, as well as other possible relevant information, such as event type, expected intensity, etc.

[0022] In a preferred embodiment, the determining of at least one mapping factor corresponding to the dynamic event specifically comprises the following steps: Retrieve and traverse a historical database, wherein the historical database includes multiple characteristic factors of the power line and a historical monitoring sequence composed of multiple historical monitoring data corresponding to each characteristic factor in time sequence; wherein each of the historical monitoring sequences is marked with a historical start time of at least one historical event, and the historical monitoring data in the historical monitoring sequence that is within a second preset time length before the historical start time is marked as the first reference monitoring data; Specifically, the historical database is obtained from the data recorded during the historical operation of the power line, and the characteristic factors include but are not limited to: meteorological factors such as wind speed, wind direction, rainfall, air humidity, temperature, etc.; factors related to the power line's own state such as power line tension, current, voltage, etc.; factors related to the surrounding environment such as surrounding magnetic field strength, etc. Among them, each historical monitoring sequence also corresponds to the historical event type marked with the historical event.

[0023] The characteristic factor of each of the first reference monitoring data that is in continuous change within the second preset time period is used as the mapping factor corresponding to the dynamic event of the same type as the historical event.

[0024] For example, if within 1 hour before the historical start time of the strong wind event, the monitoring data of the characteristic factor of wind speed continues to increase and the monitoring data of the characteristic factor of wind direction also changes continuously, then the wind speed and wind direction are determined as mapping factors of the strong wind dynamic event.

[0025] In a preferred embodiment, after retrieving and traversing the historical database, the following steps are also included: Determine whether the historical database contains the historical event of the same type as the dynamic event; The method specifically includes the following steps: taking the characteristic factor of each of the first reference monitoring data that is in continuous change within the second preset time period as the mapping factor corresponding to the dynamic event of the same type as the historical event: If the historical event of the same type as the dynamic event is matched in the historical database, the characteristic factor of each first reference monitoring data that is in continuous change within the second preset time period is used as the mapping factor corresponding to the dynamic event of the same type as the historical event.

[0026] Specifically, the historical event type of the historical event corresponding to the mark in each historical monitoring sequence is used to determine whether there are historical events of the same type as the dynamic event in the historical database. Since the types of these historical events are the same as the current dynamic events, their characteristic change patterns before the event have reference value. For example, for dynamic events such as strong winds, if there are records of strong wind events in the historical database, characteristic factors that continue to change within the second preset time period before the strong wind arrives (such as 1 hour before the strong wind arrives) can be found based on these records, such as wind speed, wind direction, etc. This method of determining mapping factors based on historical events of the same type can more accurately capture key variables related to the current dynamic event and avoid blindly selecting characteristic factors that may be irrelevant.

[0027] S2, when entering the observation window period, acquiring real-time monitoring data of each mapping factor at intervals of a first preset time length in real time to obtain a mapping factor set; the mapping factor set is composed of real-time monitoring sequences corresponding to a plurality of mapping factors; Specifically, when approaching the observation window period, the inspection robot begins to prepare for real-time monitoring of the mapping factors. During the observation window period, at intervals of a first preset duration (e.g., 5 minutes), the inspection robot obtains the real-time monitoring data of each mapping factor through sensors, and forms a mapping factor set with these data. The mapping factor set consists of real-time monitoring sequences corresponding to multiple mapping factors, each of which records the monitoring values ​​of the mapping factors at different time points; wherein the first preset duration is the same as the interval between two adjacent characteristic factors in the historical monitoring sequence.

[0028] S3. determining, according to the mapping factor set, the probability of occurrence of the dynamic event at the first prediction start time; In a preferred embodiment, a plurality of the first reference monitoring data corresponding to each of the mapping factors constitute a first reference monitoring sequence; The determining, according to the mapping factor set, the probability of occurrence of the dynamic event at the first prediction start time specifically comprises the following steps: respectively calculating a first distance of each of the mapping factors, the first distance being a Euclidean distance between the real-time monitoring sequence and the first reference monitoring sequence; For example, for strong wind events, the monitoring data of wind speed, wind direction and other mapping factors within 1 hour before the start of the historical strong wind event respectively constitute their own first reference monitoring sequences; during the observation window period, the inspection robot obtains real-time monitoring data of wind speed and wind direction every 5 minutes to form their own real-time monitoring sequences. For the wind speed mapping factor, the real-time monitoring sequence is [V1, V2, V3,..., Vn], and the first reference monitoring sequence is [W1, W2, W3,...,Wn]. Then the first distance of the wind speed mapping factor is .

[0029] The first distances of the mapping factors are weighted to obtain an average value to obtain the event occurrence probability.

[0030] Specifically, a weight value is set for each mapping factor according to the degree of influence of each mapping factor on the dynamic event. A mapping factor with a greater degree of influence can be assigned a higher weight, and a mapping factor with a smaller degree of influence can be assigned a lower weight. The weight setting can be determined by expert experience, historical data analysis, or machine learning algorithms. For example, for a strong wind event, the weight of the wind speed mapping factor is relatively high, while the weight of the wind direction mapping factor is relatively low.

[0031] Specifically, the probability of an event occurring can be calculated using the following formula: , where P is the probability of an event occurring, is the weight of the i-th mapping factor, is the first distance of the i-th mapping factor, and m is the number of mapping factors.

[0032] In a preferred embodiment, judging the probability of occurrence of the dynamic event at the first prediction start time according to the mapping factor set can be obtained by another method, which specifically includes the following steps: According to the mapping factor set, multiple reference sets are found in the historical database, the reference sets including second reference monitoring sequences corresponding to multiple mapping factors respectively; in the reference set and the mapping factor set, the Euclidean distance between the real-time monitoring sequence and the second reference monitoring sequence of the same mapping factor is less than or equal to a second preset threshold; Specifically, if the distance between the historical monitoring sequence in the historical database and the real-time monitoring sequence of a mapping factor of the current dynamic event meets the conditions, it is included in the corresponding reference set.

[0033] The reference set in the historical database that is marked with the historical start time within a third preset time period after each reference set and the historical event corresponding to the historical start time is of the same type as the dynamic event is used as a credible set; Exemplarily, if the current dynamic event is a strong wind event, within a period of time (a third preset time length) after a time point corresponding to a reference set, if the historical start time of the strong wind event appears in the historical database, then this reference set becomes a trusted set, and the historical database is traversed to find all trusted sets.

[0034] The ratio of the number of the credible set to the number of the reference set is taken as the probability of the event occurring.

[0035] In a preferred embodiment, the determining of the observation window period corresponding to the dynamic event specifically comprises the following steps: Based on the time leading interval and duration range of the trusted set relative to the corresponding historical start time, the observation window period corresponding to the dynamic event is obtained.

[0036] Specifically, the time lead interval refers to the time period from the time point corresponding to the trusted set to the start time of the historical event. For example, if the starting time point in the time series corresponding to the trusted set is T1, and the start time of the historical event corresponding to the trusted set is T2, then the time lead interval is [T1, T2). For the duration range, if the durations of the time series corresponding to several trusted sets are t1, t2, t3...tn, respectively, then the lower limit of the duration range can be the minimum value min (t1, t2, t3...tn) among these durations, and the upper limit can be the maximum value max (t1, t2, t3...tn).

[0037] S4. If the probability of the event occurring is less than a first preset threshold, the inspection robot is controlled to maintain a default inspection mode; if the probability of the event occurring is greater than or equal to the first preset threshold, the inspection robot is controlled to switch from the default inspection mode to the event response mode corresponding to the dynamic event at the start time of the first prediction.

[0038] Exemplarily, if the calculated probability of occurrence of a strong wind event is lower than the first preset threshold, it means that the strong wind is unlikely to occur at the first prediction start time, and there may be a false alarm in the external signal source. The inspection robot continues to work according to the preset default inspection mode. The default inspection mode usually performs routine inspections on power lines according to a certain route and frequency. For example, move along the power line at a stable speed, monitor the basic parameters such as the appearance, temperature, and current of the power line, and record the data and upload it to the control center. When the probability of an event is greater than or equal to the first preset threshold, it indicates that the current dynamic event is likely to occur, and the warning signal predicted by the external signal source is credible. For example, the calculated probability of occurrence of a rainstorm event reaches or exceeds the first preset threshold, which means that the rainstorm is likely to fall at the first prediction start time. After the control system of the inspection robot receives the judgment result of the probability of occurrence of the event, it immediately begins to prepare to switch the inspection mode. Each inspection robot has a different event response mode for dynamic events, and the event response mode will not be described in detail here. The event response mode includes notifying each sensor and actuator of the robot to prepare for corresponding adjustments.

[0039] Exemplarily, the event response mode for heavy rain events includes notifying waterproof sensors to strengthen monitoring, reducing movement speed, strengthening monitoring of power line insulation performance, checking whether the pole tower foundation is affected by rain erosion, etc., so as to better deal with safety risks under severe weather conditions.

[0040] In a preferred embodiment, after determining whether the historical database contains a historical event of the same type as the dynamic event, the following steps are further included: If the historical event of the same type as the dynamic event is not matched in the historical database, then based on principal component analysis, the plurality of characteristic factors in the historical database are converted into a plurality of principal components, each of which is composed of a plurality of corresponding characteristic factors and a weight of each characteristic factor, and the number of the principal components is less than the number of characteristic factors; Relevant components of the dynamic event are obtained, and the principal component matching the relevant components is used as the mapping factor corresponding to the dynamic event.

[0041] Specifically, the covariance matrix between the characteristic factors is calculated, and then the covariance matrix is ​​subjected to eigenvalue decomposition to obtain eigenvalues ​​and eigenvectors. The eigenvalues ​​are arranged in descending order, and the eigenvectors corresponding to the first few larger eigenvalues ​​are selected to form the basis vectors of the principal components. Each principal component is composed of a combination of its corresponding multiple characteristic factors and the weight of each characteristic factor, and the number of principal components is less than the number of original characteristic factors. For example, if the dynamic event is an unknown meteorological phenomenon that affects the power lines, the relevant components can be determined by some key information in the meteorological forecast, or the relevant components related to the dynamic event type in the first warning signal can be obtained, and then at least one principal component is matched according to the relevant components corresponding to the dynamic event type as the mapping factor of the dynamic event type.

[0042] Exemplarily, data of all characteristic factors are extracted from the historical database, including wind speed, wind direction, rainfall, air humidity, power line tension, current, voltage, etc. After preprocessing, the covariance matrix is ​​calculated and the eigenvalue decomposition is performed. Assuming that three principal components are obtained, the first principal component is mainly composed of wind speed, wind direction and rainfall, and the weights are w11, w12, w13 respectively; the second principal component is mainly composed of power line tension, current and voltage, and the weights are w21, w22, w23 respectively; the third principal component is mainly composed of air humidity and some other characteristic factors, and the weights are w31, w32, etc., respectively. By analyzing the weather forecast and the current power consumption, it is determined that the relevant components of the dynamic event type include specific meteorological conditions (such as strong wind, dry air) and the increase in current caused by the peak power consumption. Comparing the characteristics of the relevant components and the main components of the dynamic event, it is found that the strong wind related characteristic factor in the first principal component and the air humidity related characteristic factor in the third principal component are related to the meteorological part of the dynamic event, while the current related characteristic factor in the second principal component is related to the peak power consumption part. Therefore, these three principal components are used as mapping factors of the current dynamic events.

[0043] This embodiment provides a flexible solution for determining mapping factors. It does not rely on a specific type of historical experience, but is based on the intrinsic structure of the data itself (through principal component analysis) and the analysis of dynamic event-related components. It mines the potential relationship between the feature factors through a combination of multiple feature factors and their weights, thereby obtaining the mapping factors.

[0044] In a preferred embodiment, the historical database is also marked with the second predicted start time of each historical event; The method further comprises the following steps: When the dynamic event occurs, the actual start time of the dynamic event is recorded, and the dynamic event is used as a historical event to update the historical database; wherein the actual start time of the dynamic event is used as the historical start time, and the first predicted start time is used as the second predicted start time.

[0045] Specifically, suppose that in the past, when the historical event of strong wind actually occurred, the system recorded the time as 10 am, then this time is the actual start time of the historical event of strong wind, and at the same time, the start time predicted when the warning signal of strong wind was received before, such as 9:50 am, is used as the second predicted start time. At this time, the dynamic event has completed and can be updated to the historical database as a new historical event.

[0046] In a preferred embodiment, after receiving the first warning signal of the dynamic event, the method further includes the following steps: Retrieving and traversing the history database, and determining whether the history database contains the history event of the same type as the dynamic event; If there is at least one historical event of the same type as the dynamic event, the historical event of the same type as the dynamic event is taken as a matching event; Determining the observation window period corresponding to the dynamic event specifically includes the following steps: If the deviation between the historical start time of all the matching events and the second predicted start time is less than or equal to a third preset threshold, the observation window period corresponding to the dynamic event is determined according to the first predicted start time.

[0047] Specifically, if the deviation between the historical start time of all matching events and the second predicted start time is less than or equal to the third preset threshold, it means that the predicted start time of the historical event is close to the actual start time, which means that the predicted start time is relatively accurate. The observation window period can be determined based on the first predicted start time of the current dynamic event.

[0048] Exemplarily, the third preset threshold is set to 10 minutes. If in past rainstorm events, the time difference between the historical start time and the second predicted start time is within 10 minutes, then the observation window period can be determined based on the first predicted start time (for example, 11 a.m.) of the dynamic event of the current rainstorm warning, for example, 10:30 a.m. to 11 a.m. is the observation window period.

[0049] In a preferred embodiment, after taking the historical event of the same type as the dynamic event as a matching event, the method further includes the following steps: If there are at least N matching events whose deviations between the historical start time and the second predicted start time are greater than the third preset threshold, the N matching events are regarded as abnormal matching events, and the matching events other than the N abnormal matching events are regarded as normal matching events; Exemplarily, set N = 3. If there are 5 historical rainstorm events as matching events in the historical database, and the deviation between the historical start time of 3 of the events and the second predicted start time is greater than 10 minutes, then these 3 events are abnormal matching events, and the other 2 are normal matching events.

[0050] Obtaining a first adjustment amount according to the second predicted start time of the normal matching event, and obtaining a second adjustment amount according to the second predicted start time of the abnormal matching event; Specifically, the average time difference between the second predicted start time and the historical start time of the normal matching event can be calculated as the first adjustment amount. Similarly, the second adjustment amount is obtained.

[0051] For example, the average time difference of normal matching events is 5 minutes ahead, so the first adjustment amount is -5 minutes (assuming that ahead is negative and later is positive); the average time difference of abnormal matching events is 8 minutes later, so the second adjustment amount is +8 minutes.

[0052] The first adjustment amount matching weight is calculated based on the ratio of the normal matching events to all the matching events, and the second adjustment amount matching weight is calculated based on the ratio of the abnormal matching events to all the matching events, so as to obtain a total adjustment amount; For example, there are 5 matching events in the history database, 2 normal matching events and 3 abnormal matching events, so the ratio of normal matching events is 2 / 5 and the ratio of abnormal matching events is 3 / 5. That is, the total adjustment amount = (2 / 5)×(-5) + (3 / 5)×8 = (-2) + (24 / 5) = -2 + 4.8 = 2.8 minutes.

[0053] Before determining the observation window period corresponding to the dynamic event, the following steps are also included: The first predicted start time is updated with the total adjustment amount.

[0054] Exemplarily, the first predicted start time included in the first warning signal is 11 a.m., and according to the above, the total adjustment amount is 2.8 minutes in advance, so the updated first predicted start time is 10:57.2 a.m., and the observation window period is determined based on the updated first predicted start time (10:57.2 a.m.).

[0055] By adjusting the first prediction start time, this embodiment can more reasonably determine the range of the observation window period, so that the inspection robot can start monitoring the mapping factors at a more appropriate time, without wasting resources too early or missing key information too late, thereby better responding to dynamic events and improving the overall efficiency and performance of the system.

[0056] Furthermore, the total adjustment amount can be obtained by using abnormal matching events in the following manner: after taking the N matching events as abnormal matching events and taking the matching events other than the N abnormal matching events as normal matching events, the following steps are further included: Analyze the inducing factors that cause a large deviation between the predicted start time and the actual start time in the abnormal matching events, wherein the inducing factors are factors that exist in all the abnormal matching events and do not exist in the normal matching events; Determine whether the dynamic event has the same inducing factor; if it has the same inducing factor, use the average value of the deviation between the historical start time of the abnormal matching event and the second predicted start time as the total adjustment amount.

[0057] Exemplarily, by comparing the surrounding environment and situation corresponding to the time of occurrence of abnormal matching events, it is found that they all have a common situation, that is, before these abnormal matching events occur, there is large-scale road construction nearby, which does not happen in normal matching events. Therefore, it is determined that there is large-scale road construction nearby as an inducing factor. If there is also large-scale road construction at present, it means that the first predicted start time contained in the first warning signal is likely to be inaccurate due to large-scale road construction in the surrounding area. At this time, the average value of the deviation value between the historical start time of all previous abnormal matching events and the second predicted start time is used as the total adjustment amount to avoid the problem of poor prediction accuracy caused by inducing factors.

[0058] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. The above is only a preferred implementation method of the present invention. It should be pointed out that due to the limitations of textual expression, there are objectively infinite specific structures. For ordinary technicians in this technical field, without departing from the principles of the present invention, several improvements, modifications or changes can be made, and the above technical features can be combined in an appropriate manner; these improvements, modifications, changes or combinations, or the direct application of the inventive concept and technical solution to other occasions without improvement, should be regarded as the protection scope of the present invention.

Claims

1. A control method for a power line inspection robot based on dynamic event triggering, characterized in that: The following steps are involved: receiving a first warning signal of a dynamic event, determining at least one mapping factor corresponding to the dynamic event, and determining an observation window period corresponding to the dynamic event; The mapping factor is a variable that is reflected on or around the power line before the dynamic event occurs; the first warning signal includes a first predicted start time of the dynamic event; the observation window period is before the first predicted start time; When entering the observation window period, real-time monitoring data of each mapping factor is acquired in real time at intervals of a first preset time length to obtain a mapping factor set; The mapping factor set is composed of real-time monitoring sequences corresponding to a plurality of the mapping factors; Determining, according to the mapping factor set, the probability of occurrence of the dynamic event at the first prediction start time; If the probability of the event occurring is less than the first preset threshold, the inspection robot is controlled to maintain the default inspection mode; if the probability of the event occurring is greater than or equal to the first preset threshold, the inspection robot is controlled to switch from the default inspection mode to the event response mode corresponding to the dynamic event at the start time of the first prediction.

2. The control method of the power line inspection robot based on dynamic event triggering according to claim 1 is characterized in that: The determining of at least one mapping factor corresponding to the dynamic event specifically comprises the following steps: Retrieve and traverse a historical database, wherein the historical database includes multiple characteristic factors of the power line and a historical monitoring sequence composed of multiple historical monitoring data corresponding to each characteristic factor in time sequence; wherein each of the historical monitoring sequences is marked with a historical start time of at least one historical event, and the historical monitoring data in the historical monitoring sequence that is within a second preset time length before the historical start time is marked as the first reference monitoring data; The characteristic factor of each of the first reference monitoring data that is in continuous change within the second preset time period is used as the mapping factor corresponding to the dynamic event of the same type as the historical event.

3. The control method of the power line inspection robot based on dynamic event triggering according to claim 2 is characterized in that: The plurality of first reference monitoring data corresponding to each of the mapping factors constitute a first reference monitoring sequence; The determining, according to the mapping factor set, the probability of occurrence of the dynamic event at the first prediction start time specifically comprises the following steps: respectively calculating a first distance of each of the mapping factors, the first distance being a Euclidean distance between the real-time monitoring sequence and the first reference monitoring sequence; The first distances of the mapping factors are weighted to obtain an average value to obtain the event occurrence probability.

4. The control method of the power line inspection robot based on dynamic event triggering according to claim 3 is characterized in that: The determining, according to the mapping factor set, the probability of occurrence of the dynamic event at the first prediction start time specifically comprises the following steps: According to the mapping factor set, multiple reference sets are found in the historical database, the reference sets including second reference monitoring sequences corresponding to multiple mapping factors respectively; in the reference set and the mapping factor set, the Euclidean distance between the real-time monitoring sequence and the second reference monitoring sequence of the same mapping factor is less than or equal to a second preset threshold; A reference set in the history database that is marked with the history start time within a third preset time period after the reference set and in which the history event corresponding to the history start time is of the same type as the dynamic event is used as a credible set; The ratio of the number of the credible set to the number of the reference set is taken as the probability of the event occurring.

5. The control method of the power line inspection robot based on dynamic event triggering according to claim 4 is characterized in that: Determining the observation window period corresponding to the dynamic event specifically includes the following steps: Based on the time leading interval and duration range of the trusted set relative to the corresponding historical start time, the observation window period corresponding to the dynamic event is obtained.

6. The control method of the power line inspection robot based on dynamic event triggering according to claim 2 is characterized in that: After retrieving and traversing the historical database, the following steps are also included: Determine whether the historical database contains the historical event of the same type as the dynamic event; The method specifically includes the following steps: taking the characteristic factor of each of the first reference monitoring data that is in continuous change within the second preset time period as the mapping factor corresponding to the dynamic event of the same type as the historical event: If the historical event of the same type as the dynamic event is matched in the historical database, the characteristic factor of each first reference monitoring data that is in continuous change within the second preset time period is used as the mapping factor corresponding to the dynamic event of the same type as the historical event.

7. The control method of the power line inspection robot based on dynamic event triggering according to claim 6 is characterized in that: After determining whether the historical database contains a historical event of the same type as the dynamic event, the following steps are also included: If the historical event of the same type as the dynamic event is not matched in the historical database, then based on principal component analysis, the plurality of characteristic factors in the historical database are converted into a plurality of principal components, each of which is composed of a plurality of corresponding characteristic factors and a weight of each characteristic factor, and the number of the principal components is less than the number of characteristic factors; Relevant components of the dynamic event are obtained, and the principal component matching the relevant components is used as the mapping factor corresponding to the dynamic event.

8. The control method of the power line inspection robot based on dynamic event triggering according to claim 2, characterized in that: The historical database also contains a second predicted start time for each historical event; The method further comprises the following steps: When the dynamic event occurs, the actual start time of the dynamic event is recorded, and the dynamic event is used as a historical event to update the historical database; wherein the actual start time of the dynamic event is used as the historical start time, and the first predicted start time is used as the second predicted start time.

9. The control method of the power line inspection robot based on dynamic event triggering according to claim 8, characterized in that: After receiving the first warning signal of the dynamic event, the method further comprises the following steps: Retrieving and traversing the history database, and determining whether the history database contains the history event of the same type as the dynamic event; If there is at least one historical event of the same type as the dynamic event, the historical event of the same type as the dynamic event is taken as a matching event; Determining the observation window period corresponding to the dynamic event specifically includes the following steps: If the deviation between the historical start time of all the matching events and the second predicted start time is less than or equal to a third preset threshold, the observation window period corresponding to the dynamic event is determined according to the first predicted start time.

10. The control method of the power line inspection robot based on dynamic event triggering according to claim 9, characterized in that: After taking the historical event of the same type as the dynamic event as a matching event, the following steps are also included: If there are at least N matching events whose deviations between the historical start time and the second predicted start time are greater than the third preset threshold, the N matching events are regarded as abnormal matching events, and the matching events other than the N abnormal matching events are regarded as normal matching events; Obtaining a first adjustment amount according to the second predicted start time of the normal matching event, and obtaining a second adjustment amount according to the second predicted start time of the abnormal matching event; The first adjustment amount matching weight is calculated based on the ratio of the normal matching events to all the matching events, and the second adjustment amount matching weight is calculated based on the ratio of the abnormal matching events to all the matching events, so as to obtain a total adjustment amount; Before determining the observation window period corresponding to the dynamic event, the following steps are also included: The first predicted start time is updated with the total adjustment amount.

Images