Intelligent positioning type performance monitoring device for tactical training and application method

Abstract

The present application relates to the technical field of training data analysis, in particular to an intelligent positioning type performance monitoring device for tactical training and an application method, the device comprising an inertial navigation drift identification module, a dynamic position correction module, a behavior path monitoring module, an action deviation judgment module and a tactical performance evaluation module.In the present application, by constructing the joint discrimination condition of the speed change rate sequence and the direction angle change threshold, the inertial navigation drift interval can be identified and the stable interval identification extraction can be realized.Combining the linear fitting of the Euclidean distance difference value between the GNSS positioning result and the inertial navigation estimated value, the path dynamic compensation parameter generation and position estimation correction are completed.Combining the preset tactical action chain and the training instruction execution time sequence, the position deviation and index distance analysis are carried out, the deviation segment and the jump behavior in the action sequence are accurately identified, the accuracy of the path coherence mutation identification and the quantitative analysis ability of the action execution abnormality are improved, and the multi-dimensional dynamic judgment of the tactical stage performance state is realized.

Description

Technical Field

[0001] This invention relates to the field of training data analysis technology, and in particular to an intelligent positioning-type performance monitoring device and its application method for tactical training. Background Technology

[0002] The field of training data analysis technology involves the collection, organization, statistical analysis, and interpretation of data generated or introduced during the training process to support the evaluation and optimization of training effectiveness. This field includes structured modeling of training data, construction of data indicator systems, design of data analysis algorithms, recognition of training process behaviors, establishment of training effectiveness evaluation models, and generation of auxiliary decision-making information. Training data analysis can serve various fields such as military, sports, industry, and medicine. In tactical training scenarios, its focus is on real-time acquisition of personnel training behaviors, analysis of positional status changes, behavioral path modeling, tactical action recognition, and rule-based or model-based evaluation of training effectiveness.

[0003] Traditional intelligent positioning-based performance monitoring devices for tactical training refer to devices that monitor and analyze the position dynamics and mission behavior information of trainees or equipment during tactical training to evaluate training effectiveness. Traditional devices typically rely on GNSS satellite positioning devices to acquire the position information of the training targets, calculate attitude change parameters through inertial navigation units, and use timestamps to synchronously construct dynamic behavior sequences. They then establish a training area model based on static geographic information data. Subsequently, performance analysis is performed on training completion, maneuver path rationality, and target hit rates using rule-based methods or empirical formulas. Furthermore, the data collected during training is often stored as local files, and analysis relies heavily on manual statistical and comparative processing, lacking real-time processing and feedback capabilities. In multi-target collaborative analysis, they suffer from high computational complexity and large synchronization errors. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide an intelligent positioning-based performance monitoring device and its application method for tactical training. The technical solution is as follows:

[0005] On the one hand, an intelligent positioning-based performance monitoring device for tactical training is provided, the device comprising:

[0006] The inertial navigation drift identification module collects linear acceleration and angular velocity data for each time slice in the inertial navigation device, calculates the difference between the rate of change of velocity and the critical threshold of the rate of change of velocity, and determines whether the difference is lower than the critical threshold of the rate of change of velocity. It also determines whether the angle of change of direction within the time period is lower than the heading change threshold, filters navigation time periods that meet the dual threshold conditions, and generates a set of inertial navigation stable drift interval identifiers.

[0007] The dynamic position correction module, based on the inertial navigation stable drift interval identifier set, obtains the estimated position value of the inertial navigation and the position value provided by the GNSS positioning device within the corresponding time period of the identifier set, calculates the Euclidean distance difference sequence of the two types of positions at the same time point, and counts the linear fitting slope of the difference sequence, extracts the interval that can be constructed for linear compensation, and generates the inertial navigation position correction parameter sequence.

[0008] The behavior path monitoring module calls the inertial navigation position correction parameter sequence, extracts the average travel distance, angular velocity and control response delay time, constructs a behavior continuity vector group of the three indicators under the corresponding time window, calculates the standard deviation change between the vector groups, marks the window segment where the path change standard deviation change exceeds the path coherence threshold, and generates abrupt behavior path coherence segments.

[0009] The action offset judgment module, based on the behavior path continuity abrupt segment and combined with the pre-set standard action chain sequence table in the tactical mission, performs position comparison and index distance calculation, filters out segments that meet the condition that the offset is greater than the action tolerance limit and the jump amplitude exceeds the instruction cycle judgment boundary, and generates tactical action execution abnormal segments.

[0010] The tactical effectiveness evaluation module calls the duration, number of segments and tactical phase identifier of the abnormal segments in the tactical action execution abnormal segment, calculates the ratio of the abnormal segment duration to the standard duration, obtains the execution delay ratio, compares the dynamic time warping distance value of the abnormal action sequence and the standard action sequence as the sequence perturbation index, determines whether the phase has entered the abnormal state area, and generates a tactical phase effectiveness state mapping table.

[0011] As a further embodiment of the present invention, the inertial navigation stable drift interval identifier set includes an interval start point set, an interval end point set, and an interval duration set; the inertial navigation position correction parameter sequence includes a position difference fitting slope group, a compensation offset group, and a synchronization time point index group; the behavior path coherence abrupt change segment includes a change window number group, a window span group, and a window coherence discrete group; the tactical action execution abnormal segment includes an abnormal action number group, an abnormal segment duration group, and a sequence offset amplitude group; and the tactical phase effectiveness state mapping table includes a phase execution delay ratio group, a phase sequence disturbance index group, and a phase state identifier group.

[0012] As a further aspect of the present invention, the inertial navigation drift recognition module includes:

[0013] The inertial data acquisition submodule collects linear acceleration and angular velocity data for each time slice in the inertial navigation device. Based on the acquisition time sequence, it forms multiple sets of acceleration vector sequences and angular velocity vector sequences. It calculates the change of each set of vectors in the time axis direction and constructs a linear velocity change sequence. At the corresponding position of each time slice, it performs velocity value difference processing to obtain the linear velocity change rate values ​​of two adjacent time slices and constructs a velocity change rate data stream under a continuous time axis, generating a velocity change rate sequence value.

[0014] The speed change determination submodule, based on the speed change rate sequence value, calls each value of the speed change rate in the continuous time slice, calculates the difference between it and the set speed change rate critical threshold, performs a judgment on the magnitude of the difference and zero value, filters the set of time slices that meet the condition that the difference is less than zero value, obtains the direction angle data corresponding to the time slice, calculates the adjacent time difference value of the direction angle and judges whether the angle change value is less than the heading change threshold, removes the time slices that do not meet the angle change determination condition, and obtains a set of low speed change direction stable time slices;

[0015] The stable interval filtering submodule calls the set of stable time slices in the low-speed direction, divides the time slice set into continuous time periods according to the principle of time slice continuity, records the start index, end index and duration of each time interval, sorts each time interval according to the start and end index, constructs a continuous navigation time period interval set structure and identifies its sequence number, and establishes an inertial navigation stable drift interval identifier set.

[0016] As a further aspect of the present invention, the setting method of the critical threshold for the rate of change of velocity is specifically as follows: obtaining the rate of change of velocity values ​​within multiple consecutive time slices, and calculating the sum of the arithmetic mean and standard deviation of all the rate of change of velocity values ​​as the critical threshold for the rate of change of velocity.

[0017] The heading change threshold is set by obtaining the difference sequence between the heading angle in each time slice and the heading angle in the adjacent time slice, and calculating the angle value of the 70th decimal place in the heading angle difference sequence as the heading change threshold.

[0018] As a further aspect of the present invention, the dynamic position correction module includes:

[0019] The coordinate difference calculation submodule obtains the estimated position coordinates of the inertial navigation system and the actual position coordinates provided by the GNSS positioning module within each drift interval based on the inertial navigation stable drift interval identifier set. It extracts two position coordinates at the same time point according to the timestamp alignment rule, calculates the Euclidean distance value in three-dimensional space, and arranges them in time slice order to generate continuous position error data and establishes the Euclidean distance difference sequence value.

[0020] The position drift trend extraction submodule calls the Euclidean distance difference sequence value, performs least squares linear fitting on the continuous error value in each time interval, obtains the slope value of the corresponding fitted line, extracts the slope set corresponding to each interval, compares each slope value with the set position drift change rate threshold, filters the time intervals that meet the condition that the slope is greater than the threshold, and obtains the drift trend interval set value.

[0021] The compensation interval construction submodule extracts the estimated inertial navigation coordinates and error growth slope values ​​corresponding to the start time index and end time index based on the drift trend interval set value, establishes a linear compensation structure based on the current coordinates and slope values, constructs a compensation correction parameter list according to the segment number, and generates an inertial navigation position correction parameter sequence.

[0022] As a further aspect of the present invention, the behavior path monitoring module includes:

[0023] The behavior data extraction submodule calls the inertial navigation position correction parameter sequence. In the corrected inertial navigation trajectory, it divides the unit time window according to the time slice order, extracts the displacement value, angular velocity sequence value and command response timestamp in each window, calculates the moving distance value, average angular velocity and control response delay time value in each window, and combines and arranges them according to the window number to generate a three-dimensional behavior index sequence set.

[0024] The continuous vector construction submodule constructs a continuous vector group across windows based on the three-dimensional behavioral indicator sequence set in a time sliding window manner. It extracts the corresponding values ​​of the three behavioral indicators between adjacent windows to form a difference vector, performs standard deviation calculation on the difference vector, obtains the standard deviation change value of each vector group, and arranges the results in time order to generate a standard deviation change sequence.

[0025] The continuity mutation determination submodule extracts each standard deviation change value in sequence according to the time window based on the standard deviation change value sequence, compares it with the set path continuity threshold value, marks the window number where the standard deviation change value is greater than the path continuity threshold value, constructs an abnormal path segment number set according to the number combination, adds index time range information, and generates behavioral path continuity mutation segment.

[0026] As a further aspect of the present invention, the action offset determination module includes:

[0027] The task instruction sequence extraction submodule extracts the tactical instruction code and execution timestamp corresponding to each segment based on the behavioral path coherence abrupt segment, constructs the action execution sequence according to the chronological order and records the index position of each instruction in the path, extracts the standard action chain sequence table from the standard tactical task structure and aligns the index numbers to generate a standard comparison action sequence set.

[0028] The offset feature calculation submodule matches each execution instruction with the corresponding number in the standard action chain according to the index position of the standard action sequence set. It calculates the index difference of the same instruction item in the two sequences and records it as the sequence offset. It extracts the number span of adjacent actions in the standard sequence as the jump amplitude. It counts the offset value and jump amplitude value of each segment to obtain the action offset feature sequence.

[0029] The abnormal action segment identification submodule calls the action offset feature sequence, compares the sequence number offset value with the action tolerance limit threshold, determines whether the jump amplitude value exceeds the instruction cycle judgment boundary threshold, filters the action sequence segments that meet both conditions, extracts and combines the segment number and time range into an abnormal identification structure, and generates tactical action execution abnormal segments.

[0030] As a further aspect of the present invention, the tactical effectiveness evaluation module includes:

[0031] The task phase index calculation submodule calls the duration value, the tactical phase identifier value, and the standard duration value of the corresponding phase in the standard task flow for each segment in the abnormal tactical action execution segment. It classifies all abnormal segments by phase, calculates the ratio between the abnormal duration and the standard duration in each phase, and establishes an index mapping by phase number to generate a phase execution delay ratio sequence.

[0032] The sequence perturbation distance extraction submodule extracts the abnormal action sequence and standard action sequence under each stage based on the determined set of tactical stage numbers in the stage execution delay ratio sequence, performs dynamic time warping according to the time alignment rules, calculates the cumulative distance value under the action alignment path, and records the perturbation measure corresponding to each stage as an index value to obtain the tactical sequence perturbation distance sequence.

[0033] The performance status mapping construction submodule extracts two corresponding indicator values ​​for each tactical stage based on the tactical sequence perturbation distance sequence and the stage execution delay ratio sequence, compares them with preset state judgment thresholds, identifies whether each indicator exceeds the limit, marks the exceeding state with Boolean values, binds the judgment result with the corresponding tactical stage number, constructs a structured result table, and generates a tactical stage performance status mapping table.

[0034] As a further embodiment of the present invention, the setting method of the state judgment threshold is specifically as follows: based on the distribution of the perturbation distance values ​​of each stage in the tactical sequence perturbation distance sequence, the mean and standard deviation of all perturbation distance values ​​are extracted, and the sum of the mean and standard deviation is superimposed as the sequence perturbation distance threshold.

[0035] Based on the statistical results of the ratio values ​​of each stage in the stage execution delay ratio sequence, the 90th digit of the delay ratio value distribution is extracted as the stage execution delay ratio threshold.

[0036] The sequence perturbation distance threshold and the stage execution delay ratio threshold are respectively used as two components of the preset state judgment threshold;

[0037] The index value corresponding to each tactical phase in the tactical sequence perturbation distance sequence and the phase execution delay ratio sequence must be compared with the two components mentioned above.

[0038] When either indicator value in the two comparisons is greater than the corresponding threshold, a Boolean value is generated to mark the result as logically true.

[0039] On the other hand, a method for intelligent positioning performance monitoring for tactical training, which is based on the aforementioned intelligent positioning performance monitoring device for tactical training, includes the following steps:

[0040] S1: Collect linear acceleration and angular velocity data for each time slice in the inertial navigation device, calculate the difference between the rate of change of velocity and the critical threshold of the rate of change of velocity, and determine whether the difference is lower than the critical threshold of the rate of change of velocity. Determine whether the angle of change of direction within the time period is lower than the heading change threshold, filter the navigation time periods that meet the dual threshold conditions, and generate a set of inertial navigation stable drift interval identifiers.

[0041] S2: Based on the inertial navigation stable drift interval identifier set, obtain the inertial navigation estimated position value and the position value provided by the GNSS positioning device within the corresponding time period of the identifier set, calculate the Euclidean distance difference sequence of the two types of positions at the same time point, and count the linear fitting slope of the difference sequence, extract the interval that can be constructed for linear compensation, and generate the inertial navigation position correction parameter sequence.

[0042] S3: Call the inertial navigation position correction parameter sequence, extract the average travel distance, angular velocity and control response delay time, construct the behavior continuity vector group of the three indicators under the corresponding time window, calculate the standard deviation change between the vector groups, mark the window segment where the path change standard deviation change exceeds the path coherence threshold, and generate the behavior path coherence abrupt change segment.

[0043] S4: Based on the behavioral path coherence abrupt segment, combined with the standard action chain sequence table preset in the tactical task, perform position comparison and index distance calculation, filter out segments that meet the condition that the offset is greater than the action tolerance limit and the jump amplitude exceeds the instruction cycle judgment boundary, and generate tactical action execution abnormal segments.

[0044] S5: Invoke the duration, number of segments and tactical phase identifier of the abnormal segment in the execution abnormal segment of the tactical action, calculate the ratio of the abnormal segment duration to the standard duration, obtain the execution delay ratio, compare the dynamic time warping distance value of the abnormal action sequence and the standard action sequence as the sequence perturbation index, determine whether the phase has entered the abnormal state area, and generate a tactical phase effectiveness state mapping table.

[0045] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0046] This invention identifies inertial navigation drift intervals and extracts stable interval markers by constructing a joint discrimination condition of velocity change rate sequence and orientation angle change threshold. It combines linear fitting of the Euclidean distance difference between GNSS positioning results and inertial navigation estimation values ​​to generate path dynamic compensation parameters and correct position estimates. It extracts behavioral continuity index sequences within a window and marks path abrupt change segments based on standard deviation changes. It analyzes position offsets and index distances by combining preset tactical action chains and training command execution sequences, accurately identifying offset segments and jump behaviors in action sequences. It maps tactical effectiveness status through the ratio of abnormal segment duration and dynamic time warping distance values, improving the accuracy of path continuity abrupt change identification and the quantitative analysis capability of action execution anomalies, thus achieving multi-dimensional dynamic judgment of tactical phase effectiveness status. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a schematic diagram of an intelligent positioning-type performance monitoring device for tactical training provided in an embodiment of the present invention;

[0049] Figure 2 This is a schematic diagram of the device frame of the present invention;

[0050] Figure 3 This is a flowchart of the inertial navigation drift recognition module in this invention;

[0051] Figure 4 This is a flowchart of the dynamic position correction module in this invention;

[0052] Figure 5 This is a flowchart of the behavior path monitoring module in this invention;

[0053] Figure 6 This is a flowchart of the action offset judgment module in this invention;

[0054] Figure 7 This is a flowchart of the tactical effectiveness evaluation module in this invention;

[0055] Figure 8 This is a flowchart of an intelligent positioning-based performance monitoring application method for tactical training provided in an embodiment of the present invention. Detailed Implementation

[0056] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0057] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0058] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0059] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0060] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0061] like Figure 1-2 As shown, this embodiment of the invention provides an intelligent positioning-type performance monitoring device for tactical training. The device includes an inertial navigation drift recognition module, a dynamic position correction module, a behavior path monitoring module, a motion deviation judgment module, and a tactical performance evaluation module.

[0062] The inertial navigation drift identification module collects linear acceleration and angular velocity data for each time slice in the inertial navigation device, constructs a velocity change rate sequence for continuous time slices, calculates the difference between each item in the sequence and the critical threshold of velocity change rate, and determines whether the difference in continuous time slices is always lower than the critical threshold of velocity change rate. It also determines whether the angle of change of direction within a time period is lower than the heading change threshold, filters navigation time periods that meet the dual threshold conditions, and generates a set of inertial navigation stable drift interval identifiers.

[0063] The dynamic position correction module is based on the inertial navigation stable drift interval identifier set. It obtains the estimated position value of the inertial navigation and the position value provided by the GNSS positioning device within the corresponding time period of the identifier set. It calculates the Euclidean distance difference sequence of the two types of positions at the same time point and counts the linear fitting slope of the difference sequence. Based on the comparison relationship between the linear fitting slope and the position drift change rate threshold, it extracts the interval that can be linearly compensated and generates the inertial navigation position correction parameter sequence.

[0064] The Euclidean distance difference sequence refers to the spatial distance between the inertial navigation estimated position coordinates and the actual GNSS coordinates at each time slice;

[0065] The behavior path monitoring module calls the inertial navigation position correction parameter sequence. In the corrected navigation path data, it extracts the average distance traveled, the average angular velocity, and the control response delay time within a unit time window. It constructs a behavior continuity vector group for the three indicators under the corresponding time window, calculates the standard deviation change between the vector groups in a continuous sliding window, marks the window segment where the standard deviation change of the path change exceeds the path coherence threshold, and generates abrupt behavior path coherence segments.

[0066] Control response delay time is defined as the difference between the time the command is issued and the time the inertial navigation path responds; the standard deviation change represents the difference in the dispersion of continuous vector values ​​in adjacent windows, and is used to detect sudden changes in behavioral stability.

[0067] The action offset judgment module is based on the behavior path continuity abrupt segment, obtains the tactical instruction encoding sequence and execution timestamp sequence in the corresponding segment, combines the standard action chain sequence table that has been preset in the tactical task, compares the execution position of the two sequences and calculates the index distance, counts the sequence number offset and action jump amplitude between each instruction and the standard chain, filters the segments that meet the condition that the offset is greater than the action tolerance limit and the jump amplitude exceeds the instruction cycle judgment boundary, and generates tactical action execution abnormal segments.

[0068] Sequence offset refers to the difference in the number of the executed action relative to the standard sequence in the abnormal segment; the action jump range is defined as the discontinuous span between action numbers in the execution sequence. For example, a jump from the standard action sequence [1, 2, 3] to [1, 3] is a jump range of 1.

[0069] The tactical effectiveness evaluation module calls the duration, number of segments and tactical phase identifier of the abnormal segments in the tactical action execution abnormal segment. Combined with the standard duration of the corresponding phase in the standard task flow and the action execution sequence, it calculates the ratio of the abnormal segment duration to the standard duration, obtains the execution delay ratio, compares the dynamic time warping distance value between the abnormal action sequence and the standard action sequence as the sequence perturbation index, summarizes the two types of indicators of the tactical phase and sets the state threshold, determines whether the phase has entered the abnormal state area, and generates a tactical phase effectiveness state mapping table.

[0070] Dynamic Time Warping Distance (DTW) is a method for dynamically matching two time series sequences, suitable for comparing sequences with different time axis lengths or time offsets. If A is a standard action sequence and B is an abnormal action sequence, then DTW(A, B) represents the minimum matching cost between the two sequences under the optimal path.

[0071] The inertial navigation stable drift interval identifier set includes the interval start point set, interval end point set, and interval duration set. The inertial navigation position correction parameter sequence includes the position difference fitting slope group, compensation offset group, and synchronization time point index group. The behavior path coherence abrupt change segment includes the abrupt change window number group, window span group, and window coherence discrete group. The tactical action execution abnormal segment includes the abnormal action number group, abnormal segment duration group, and sequence offset amplitude group. The tactical phase effectiveness state mapping table includes the phase execution delay ratio group, phase sequence disturbance index group, and phase state identifier group.

[0072] Specifically, such as Figure 2 , 3 As shown, the inertial navigation drift recognition module includes:

[0073] The inertial data acquisition submodule collects linear acceleration and angular velocity data for each time slice in the inertial navigation device. Based on the acquisition time sequence, it forms multiple sets of acceleration vector sequences and angular velocity vector sequences. It calculates the change of each set of vectors in the time axis direction and constructs a linear velocity change sequence. At the corresponding position of each time slice, it performs velocity value difference processing to obtain the linear velocity change rate values ​​of two adjacent time slices and constructs a velocity change rate data stream under a continuous time axis, generating a velocity change rate sequence value.

[0074] Following a set high-frequency sampling frequency of 1 kHz, the processor reads the X-axis, Y-axis, and Z-axis acceleration values ​​output by the three-axis accelerometers in the inertial measurement unit, as well as the roll, pitch, and yaw angular velocity values ​​output by the three-axis gyroscopes, through the underlying hardware interface. The processor timestamps the six dimensions of data read at each moment with a data acquisition precision down to the microsecond level and stores them in a circular buffer strictly according to the order of the timestamps. For each time slice, the processor calls the vector magnitude calculation logic to sum the squares of the three-axis acceleration and angular velocity values ​​and then take the square root to obtain the composite linear acceleration magnitude and composite angular velocity magnitude corresponding to that time slice. Subsequently, the processor iterates through consecutive time slices in the buffer, subtracting the composite linear acceleration magnitude of the previous time slice from the current time slice's composite linear acceleration magnitude to obtain the change in linear acceleration along the time axis; similarly, the change in angular velocity is calculated. Based on this, the processor performs a first-order difference operation on the linear acceleration changes of adjacent time slices, that is, subtracts the change of the previous time slice from the change of the next time slice, and divides the difference by the sampling time interval to obtain the linear velocity change rate value at that time slice, and finally forms a velocity change rate data stream containing the complete time series.

[0075] The speed change determination submodule, based on the speed change rate sequence value, calls each value of the speed change rate in the continuous time slice and calculates the difference between it and the set speed change rate critical threshold. It performs a comparison between the difference and zero, filters the set of time slices that meet the condition that the difference is less than zero, obtains the azimuth angle data corresponding to the time slice, calculates the adjacent time difference value of the azimuth angle and determines whether the angle change value is less than the heading change threshold, and removes the time slices that do not meet the angle change determination condition to obtain a set of low-speed directional stable time slices.

[0076] The specific method for setting the critical threshold for the rate of change of velocity is as follows: obtain the rate of change of velocity values ​​within multiple consecutive time slices, calculate the sum of the arithmetic mean and standard deviation of all the rate of change of velocity values, and use it as the critical threshold for the rate of change of velocity.

[0077] The specific method for setting the heading change threshold is as follows: obtain the difference sequence between the heading angle in each time slice and the heading angle in the adjacent time slice, calculate the angle value of the 70th tenth of the heading angle difference sequence, and use it as the heading change threshold.

[0078] The process of generating the critical threshold for the rate of change of velocity is as follows: All rate of change values ​​stored in the buffer over the past five seconds are selected, their totals are counted, and the arithmetic mean is obtained by summing all values ​​and dividing by the total. Simultaneously, the dispersion of these values ​​relative to the mean is calculated to obtain the standard deviation. The arithmetic mean and standard deviation are then added directly; the result is the critical threshold for the rate of change of velocity. After setting the threshold, the processor reads the rate of change of velocity sequence values ​​within the current monitoring window one by one, calculating the algebraic difference between each rate of change value and the critical threshold. If the difference is less than zero, it indicates that the motion state within that time slice tends towards stabilization or deceleration, and the processor marks that time slice as an initially selected stabilization point. For the set of initially selected stabilization points, the processor extracts the heading angle data for the corresponding time point and calculates the absolute difference between the heading angle at the current time point and the heading angle at the previous time point. To determine the smoothness of angle changes, the processor sets a heading change threshold: It selects the heading angle differences of all adjacent time slices within a historical one-minute period to form a sequence, sorts this sequence from smallest to largest, and locates the value at the 70th percentile of the sequence length (the 70th decimal place), using this value as the heading change threshold. Finally, the processor detects time slices in the initially selected smooth points where the heading angle difference is less than this heading change threshold, retains them and stores them in the set of smooth time slices in the low-speed-change direction, and discards data that does not meet the criteria.

[0079] The stable interval filtering submodule calls the set of stable time slices in the low-speed direction, divides the time slice set into continuous time periods according to the principle of time slice continuity, records the start index, end index and duration of each time interval, sorts each time interval according to the start and end index, constructs a continuous navigation time period interval set structure and identifies its sequence number, and establishes an inertial navigation stable drift interval identifier set.

[0080] The processor reads all time index numbers from the set of stable time slices in the low-speed-variable direction. It performs a continuous scan, determining if adjacent index numbers are consecutive (i.e., the latter number equals the former number plus one). If the numbers are consecutive, they are grouped into the same time period; if there is a gap, the current time period is truncated, and the new number after the gap is used as the starting point of the next time period. For each consecutive time period, the processor records its start and end time slice indices, and calculates the duration by subtracting the start index from the end index and adding one. Subsequently, the processor sorts all the consecutive time periods in ascending order according to the start index. The processor constructs a structure container containing three fields: start index, end index, and duration. The sorted time period data is filled into this container sequentially, and a unique sequential number is assigned to each time period, ultimately forming a set of inertial navigation stable drift interval identifiers. This set explicitly records all time segments in which the inertial navigation device is in a relatively stationary or uniform straight-line state.

[0081] Specifically, such as Figure 2 , 4 As shown, the dynamic position correction module includes:

[0082] The coordinate difference calculation submodule is based on the inertial navigation stable drift interval identifier set. It obtains the estimated position coordinates of the inertial navigation and the actual position coordinates provided by the GNSS positioning module in each drift interval. According to the timestamp alignment rule, it extracts two position coordinates at the same time point, calculates the Euclidean distance value in three-dimensional space, and arranges them in time slice order to generate continuous position error data and establishes the Euclidean distance difference sequence value.

[0083] Based on the inertial navigation system (INS) stable drift interval identifier set, the processor iterates through each time interval identified as stable. Within each interval, the processor retrieves the estimated position coordinates (including longitude, latitude, and projected Cartesian coordinates X, Y, and Z) output from the INS solution and the actual position coordinates output from the GNSS receiver. The processor ensures, according to timestamp alignment rules, that the INS coordinates and satellite positioning coordinates involved in the calculation belong to the same sampling time. For each pair of matching time points, the processor performs a three-dimensional Euclidean distance calculation: it calculates the coordinate differences between the two points along the X, Y, and Z axes, squares each of these differences, adds the three squares, and then takes the square root. The result is the Euclidean distance of the position error at that time. The processor stores the calculated Euclidean distance values ​​for each time point sequentially into an array according to the time slice order, constructing continuous position error data, ultimately forming a sequence of Euclidean distance differences reflecting the change of error over time.

[0084] The position drift trend extraction submodule calls the Euclidean distance difference sequence value, performs least squares linear fitting on the continuous error value within each time interval, obtains the slope value of the corresponding fitted line, extracts the slope set corresponding to each interval, compares each slope value with the set position drift change rate threshold, filters the time intervals that meet the condition that the slope is greater than the threshold, and obtains the drift trend interval set value.

[0085] The specific method for setting the position drift rate threshold is to calculate the sum of the mean and standard deviation of the continuous error change corresponding to the Euclidean distance difference sequence value in each time interval, and use the sum as the position drift rate threshold.

[0086] Linear regression analysis using the least squares method on the error data within the interval aims to find a straight line that optimally fits the error growth trend. During the calculation, the processor uses time as the independent variable and the Euclidean distance difference as the dependent variable. It solves for the slope parameter of the fitted line by minimizing the sum of squared residuals between the actual error value and the predicted value of the fitted line. This slope value quantifies the rate of positional drift. To determine whether the drift is abnormal, the processor sets a threshold for the rate of change of positional drift: it calculates the arithmetic mean of the differences between adjacent values ​​in the Euclidean distance difference sequence within the current time interval, along with the standard deviation of these differences, and adds the mean and standard deviation as the threshold. The processor compares the fitted slope value with this threshold. If the slope value is greater than the threshold, it indicates that the positional error within that interval shows a rapid divergence trend, indicating a significant drift phenomenon. The processor filters out time intervals that meet this condition, records their interval information, and obtains the set value of the drift trend intervals.

[0087] The compensation interval construction submodule extracts the estimated inertial navigation coordinates and error growth slope values ​​corresponding to the start time index and end time index based on the drift trend interval set values, establishes a linear compensation structure based on the current coordinates and slope values, constructs a compensation correction parameter list according to the segment number, and generates an inertial navigation position correction parameter sequence.

[0088] For each drift interval, the processor extracts the estimated inertial navigation coordinates corresponding to its starting time as a reference point, and also extracts the error growth slope value calculated in the previous step. The processor establishes a linear compensation model, whose correction logic is: for any time within the interval, the correction amount is equal to the time difference between that time and the starting time multiplied by the error growth slope. The processor subtracts this correction amount from the current coordinate value to offset the accumulated error caused by the drift. The processor packages the calculated correction amount and the corresponding start and end time indices, numbers them according to the time sequence of the intervals, and constructs a compensation correction parameter list. This list defines in detail which time periods require how much linear correction amount to apply, ultimately generating an inertial navigation position correction parameter sequence used to correct the navigation trajectory.

[0089] Specifically, such as Figure 2 , 5 As shown, the behavior path monitoring module includes:

[0090] The behavior data extraction submodule calls the inertial navigation position correction parameter sequence. In the corrected inertial navigation trajectory, it divides the unit time window according to the time slice order, extracts the displacement value, angular velocity sequence value and command response timestamp in each window, calculates the movement distance value, average angular velocity and control response delay time value in each window, and combines and arranges them according to the window number to generate a three-dimensional behavior index sequence set.

[0091] The processor divides the corrected trajectory data into continuous unit time windows of fixed duration (e.g., one second). Within each time window, the processor first calculates the straight-line distance between the start and end points of the window, as the travel distance value for that window. Second, the processor accumulates the angular velocity magnitudes of all sampled points within the window and divides them by the number of sampled points to obtain the average angular velocity. Third, the processor retrieves the control command records that occurred within that window and calculates the time difference between the command issuance time and the moment the inertial sensor detects the corresponding action response, as the control response delay value. The processor encapsulates these three metrics (travel distance, average angular velocity, and response delay) into a feature vector and combines these vectors according to the time window numbering order to generate a complete set of three-dimensional behavioral metric sequences for subsequent coherence analysis.

[0092] The continuous vector construction submodule constructs a continuous vector group across windows based on the three-dimensional behavioral indicator sequence set using a time sliding window approach. It extracts the corresponding values ​​of the three behavioral indicators between adjacent windows to form a difference vector, performs standard deviation calculation on the difference vector, obtains the standard deviation change value of each vector group, and arranges the results in time order to generate a sequence of standard deviation changes.

[0093] The processor sets up a sliding window covering three unit time windows, sliding forward one unit at a time. Within each sliding window, the processor extracts the three-dimensional behavioral indicator vectors corresponding to two adjacent unit time windows and performs vector subtraction, subtracting the indicator vector of the previous window from the indicator vector of the subsequent window to obtain a difference vector. This difference vector reflects the drastic changes in the behavioral indicators over a short period. Next, the processor calculates the standard deviation of the three elements (distance change, angular velocity change, and delay change) in this difference vector to comprehensively quantify the volatility of the behavioral pattern within the sliding window. The processor arranges the calculated standard deviation values ​​according to the starting time of the sliding window, forming a sequence of standard deviation changes reflecting the consistency and volatility of the behavioral patterns.

[0094] The continuity mutation determination submodule extracts each standard deviation change value in sequence according to the time window based on the standard deviation change value sequence, compares it with the set path continuity threshold value, marks the window number where the standard deviation change value is greater than the path continuity threshold value, constructs an abnormal path segment number set by combining the numbers and adds index time range information to generate behavioral path continuity mutation segments.

[0095] The specific method for setting the path coherence threshold is as follows: perform distribution statistics on the standard deviation changes of all time windows in the standard deviation change sequence, obtain the mean and standard deviation of the standard deviation change sequence, and use the sum of the mean and standard deviation as the path coherence threshold.

[0096] The process of setting the path coherence threshold involves the processor statistically analyzing the distribution of all standard deviation changes throughout the sequence, calculating their arithmetic mean and standard deviation, and summing them to obtain the threshold. Subsequently, the processor iterates through the standard deviation change sequence item by item, comparing each value with the path coherence threshold. When the standard deviation change of a certain item exceeds the threshold, it indicates a drastic fluctuation in the behavioral pattern near that moment, constituting a coherence abrupt change. The processor records the time window number corresponding to this abrupt change and merges consecutively occurring abrupt change numbers into a segment. The processor adds specific start and end time range index information to each aberrant segment, ultimately generating behavioral path coherence abrupt change segments. These segments indicate potentially discontinuous or aberrant areas during the execution of tactical actions.

[0097] Specifically, such as Figure 2 , 6 As shown, the motion offset determination module includes:

[0098] The task instruction sequence extraction submodule extracts the tactical instruction code and execution timestamp corresponding to each paragraph based on the abrupt paragraphs of behavioral path coherence. It constructs the action execution sequence according to the chronological order and records the index position of each instruction in the path. It extracts the standard action chain sequence table from the standard tactical task structure and aligns the index numbers to generate a standard comparison action sequence set.

[0099] The processor reads the time range of the abrupt change segment and retrieves all tactical instructions falling within that time range from the tactical mission log database. For each retrieved instruction, the processor extracts its specific tactical instruction code (e.g., advance, turn, stop) and precise execution timestamp. The processor arranges these instructions in chronological order of their timestamps, constructing the actual action execution sequence within the abrupt period and recording the index position of each instruction in the overall mission path. Simultaneously, the processor reads pre-set tactical mission standard structure data and extracts a standard action chain sequence list that should theoretically be executed for that mission phase. The processor aligns the actual execution sequence with the standard action chain index numbers using time or phase identifiers, thereby generating a standard comparison action sequence set containing a comparison between "actual execution" and "standard requirements."

[0100] The offset feature calculation submodule matches each execution instruction with the corresponding number in the standard action chain according to the index position based on the standard comparison action sequence set. It calculates the index difference of the same instruction item in the two sequences and records it as the sequence offset. It extracts the number span of adjacent actions in the standard sequence as the jump amplitude. It counts the offset value and jump amplitude value of each segment to obtain the action offset feature sequence.

[0101] The processor iterates through each actual instruction in the sequence, searching for its corresponding standard number in the standard action chain based on its instruction code. The processor calculates the difference between the actual instruction's position index in the sequence and its theoretical position number in the standard action chain, recording this difference as the sequence offset. This value reflects whether the action execution is ahead or behind schedule. Simultaneously, the processor analyzes the number span between two adjacent actions in the standard chain, subtracting the standard number of the preceding action from the standard number of the following action, and uses this difference as the jump magnitude. If the jump magnitude is greater than one, it indicates a missed action; if it is negative, it indicates action rollback. The processor statistically analyzes all sequence offsets and jump magnitudes within each mutation segment, forming an action offset feature sequence used to quantitatively assess the degree of disorder in action execution.

[0102] The action anomaly segment identification submodule calls the action offset feature sequence, compares the sequence number offset value item by item to see if it is greater than the action tolerance limit threshold, determines whether the jump amplitude value exceeds the instruction cycle judgment boundary threshold, filters the action sequence segments that meet both conditions at the same time, extracts and combines the segment number and time range into an anomaly identification structure, and generates tactical action execution anomaly segments.

[0103] The specific method for setting the action tolerance limit threshold is as follows: based on the set of number differences between all consecutive action numbers in the tactical instruction coding sequence, calculate the sum of the median of the set of number differences and the standard deviation of the set, and use the obtained value as the action tolerance limit threshold.

[0104] The specific method for setting the instruction cycle determination boundary threshold is as follows: extract the theoretical interval time set between adjacent actions in the standard action chain sequence table, calculate the time value corresponding to the seventh quantile in the set, and use the time value as the instruction cycle determination boundary threshold;

[0105] First, the processor sets an action tolerance threshold: it collects the differences between all adjacent action numbers in the tactical instruction encoding sequence, forming a set of number differences, calculates the median and standard deviation of this set, and adds the median and standard deviation as the threshold. Second, it sets an instruction cycle judgment boundary threshold: the processor extracts the theoretical interval time between adjacent actions specified in the standard action chain, sorts these time values, and takes the 70th percentile as the threshold. During the judgment process, the processor checks each segment in the action offset feature sequence, determining whether its sequence offset is greater than the action tolerance threshold, and simultaneously determining whether the execution interval time corresponding to its jump amplitude exceeds the instruction cycle judgment boundary threshold. Only when both conditions are met simultaneously does the processor consider the segment a true tactical execution anomaly. The processor extracts the segment numbers and time ranges that meet these conditions, combines them to generate an anomaly identifier structure, and outputs the tactical action execution anomaly segment.

[0106] Specifically, such as Figure 2 , 7 As shown, the tactical effectiveness assessment module includes:

[0107] The task phase index calculation submodule calls the duration value, the tactical phase identifier value, and the standard duration value of the corresponding phase in the standard task flow for each abnormal segment in the tactical action execution abnormal segment. It classifies all abnormal segments by phase, calculates the ratio between the abnormal duration and the standard duration in each phase, and establishes an index mapping by phase number to generate a phase execution delay ratio sequence.

[0108] The processor accesses a pre-stored standard task flow database to obtain the standard duration reference value for each corresponding stage. Then, the processor calculates the actual duration of each abnormal segment, i.e., the end time minus the start time. The processor categorizes all abnormal segments according to the tactical stage identifier. For all abnormal segments within the same tactical stage, the durations are summed to obtain the total abnormal duration for that stage. Next, the processor performs a ratio calculation, dividing the total abnormal duration of the stage by the standard duration value of that stage to obtain the stage execution delay ratio. The processor establishes a key-value mapping relationship between the calculated ratio values ​​and the tactical stage numbers, ultimately generating a stage execution delay ratio sequence. This sequence visually reflects the degree of progress delay caused by abnormal actions in each task stage.

[0109] The sequence perturbation distance extraction submodule extracts the abnormal action sequence and standard action sequence under each stage based on the set of tactical stage numbers determined in the stage execution delay ratio sequence. It performs dynamic time warping according to the time alignment rules, calculates the cumulative distance value under the action alignment path, and records the perturbation metric corresponding to each stage as the index value to obtain the tactical sequence perturbation distance sequence.

[0110] The processor extracts the actual abnormal action sequences and their corresponding standard action sequences within the extraction phase. Since the two sequences may differ in duration and number of actions, the processor employs Dynamic Time Warping (DTW). This process constructs a distance matrix, where rows and columns correspond to time points of the two sequences, and matrix elements represent the Euclidean distance between the action features at the corresponding time points. The processor searches the matrix for the optimal path from the bottom left to the top right corner, minimizing the sum of distances for all elements along the path. This minimum path distance accumulation is the perturbation metric, eliminating the influence of timeline distortion and purely measuring the difference between the action execution order and form and the standard flow. The processor records the calculated perturbation metric as an index value, resulting in the tactical sequence perturbation distance sequence.

[0111] The effectiveness state mapping construction submodule extracts two corresponding indicator values ​​for each tactical stage based on the tactical sequence perturbation distance sequence and the stage execution delay ratio sequence, compares them with the preset state judgment threshold, identifies whether each indicator exceeds the limit, marks the exceeding state with Boolean value, binds the judgment result with the corresponding tactical stage number, constructs a structured result table, and generates a tactical stage effectiveness state mapping table.

[0112] The specific method for setting the state judgment threshold is as follows: based on the distribution of the perturbation distance values ​​at each stage in the tactical sequence perturbation distance sequence, extract the mean and standard deviation of all perturbation distance values, and sum the mean and standard deviation as the sequence perturbation distance threshold.

[0113] Based on the statistical results of the ratio values ​​of each stage in the stage execution delay ratio sequence, the 90th digit of the delay ratio value distribution is extracted as the stage execution delay ratio threshold.

[0114] The sequence perturbation distance threshold and the stage execution delay ratio threshold are respectively used as two components of the preset state judgment threshold;

[0115] The index value corresponding to each tactical phase in the tactical sequence perturbation distance sequence and the phase execution delay ratio sequence must be compared with the two components mentioned above.

[0116] When either indicator value in the two comparisons is greater than the corresponding threshold, a Boolean value is generated to mark the result as logically true.

[0117] First, the decision thresholds are determined. For the tactical sequence perturbation distance sequence, the processor calculates the arithmetic mean and standard deviation of all phase perturbation distance values, and adds them together as the sequence perturbation distance threshold. For the phase execution delay ratio sequence, the processor sorts all ratio values ​​and selects the 90th decimal place as the phase execution delay ratio threshold. During the decision phase, the processor iterates through each tactical phase, extracting its corresponding sequence perturbation distance value and execution delay ratio value. The processor compares these two values ​​with the two thresholds mentioned above: if the sequence perturbation distance value is greater than the sequence perturbation distance threshold, or the phase execution delay ratio value is greater than the phase execution delay ratio threshold, the processor marks the performance status of that tactical phase as logically true (i.e., there is a performance problem). Finally, the processor binds this Boolean result to the tactical phase number, constructing a structured tactical phase performance status mapping table.

[0118] Please see Figure 8 The intelligent positioning effectiveness monitoring application method for tactical training is executed based on the aforementioned intelligent positioning effectiveness monitoring device for tactical training, and includes the following steps:

[0119] S1: Collect linear acceleration and angular velocity data for each time slice in the inertial navigation device, calculate the difference between the rate of change of velocity and the critical threshold of the rate of change of velocity, and determine whether the difference is lower than the critical threshold of the rate of change of velocity. Determine whether the angle of change of direction within the time period is lower than the heading change threshold, filter the navigation time periods that meet the dual threshold conditions, and generate a set of inertial navigation stable drift interval identifiers.

[0120] S2: Based on the inertial navigation stable drift interval identifier set, obtain the inertial navigation estimated position value and the position value provided by the GNSS positioning device within the corresponding time period of the identifier set, calculate the Euclidean distance difference sequence of the two types of positions at the same time point, and count the linear fitting slope of the difference sequence, extract the interval that can be constructed for linear compensation, and generate the inertial navigation position correction parameter sequence.

[0121] S3: Call the inertial navigation position correction parameter sequence, extract the average travel distance, angular velocity and control response delay time, construct the behavior continuity vector group of the three indicators under the corresponding time window, calculate the standard deviation change between the vector groups, mark the window segment where the path change standard deviation change exceeds the path coherence threshold, and generate the behavior path coherence abrupt change segment.

[0122] S4: Based on the abrupt change in the continuity of the behavior path, combined with the pre-set standard action chain sequence table in the tactical mission, perform position comparison and index distance calculation, filter out segments that meet the condition that the offset is greater than the action tolerance limit and the jump range exceeds the instruction cycle judgment boundary, and generate tactical action execution abnormal segments.

[0123] S5: Invoke the duration, number of segments and tactical phase identifier of the abnormal segment in the execution of tactical actions, calculate the ratio of the abnormal segment duration to the standard duration, obtain the execution delay ratio, compare the dynamic time warping distance value between the abnormal action sequence and the standard action sequence as the sequence perturbation index, determine whether the phase has entered the abnormal state zone, and generate a tactical phase effectiveness state mapping table.

[0124] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An intelligent positioning-type performance monitoring device for tactical training, characterized in that, The device includes: The inertial navigation drift identification module collects linear acceleration and angular velocity data for each time slice in the inertial navigation device, calculates the difference between the rate of change of velocity and the critical threshold of the rate of change of velocity, and determines whether the difference is lower than the critical threshold of the rate of change of velocity. It also determines whether the angle of change of direction within the time period is lower than the heading change threshold, filters navigation time periods that meet the dual threshold conditions, and generates a set of inertial navigation stable drift interval identifiers. The dynamic position correction module, based on the inertial navigation stable drift interval identifier set, obtains the estimated position value of the inertial navigation and the position value provided by the GNSS positioning device within the corresponding time period of the identifier set, calculates the Euclidean distance difference sequence of the two types of positions at the same time point, and counts the linear fitting slope of the difference sequence, extracts the interval that can be constructed for linear compensation, and generates the inertial navigation position correction parameter sequence. The behavior path monitoring module calls the inertial navigation position correction parameter sequence, extracts the average travel distance, angular velocity and control response delay time, constructs a behavior continuity vector group of the three indicators under the corresponding time window, calculates the standard deviation change between the vector groups, marks the window segment where the path change standard deviation change exceeds the path coherence threshold, and generates abrupt behavior path coherence segments. The action offset judgment module, based on the behavior path continuity abrupt segment and combined with the pre-set standard action chain sequence table in the tactical mission, performs position comparison and index distance calculation, filters out segments that meet the condition that the offset is greater than the action tolerance limit and the jump amplitude exceeds the instruction cycle judgment boundary, and generates tactical action execution abnormal segments. The tactical effectiveness evaluation module calls the duration, number of segments and tactical phase identifier of the abnormal segments in the tactical action execution abnormal segment, calculates the ratio of the abnormal segment duration to the standard duration, obtains the execution delay ratio, compares the dynamic time warping distance value of the abnormal action sequence and the standard action sequence as the sequence perturbation index, determines whether the phase has entered the abnormal state area, and generates a tactical phase effectiveness state mapping table.

2. The intelligent positioning-type performance monitoring device for tactical training according to claim 1, characterized in that, The inertial navigation stable drift interval identifier set includes an interval start point set, an interval end point set, and an interval duration set. The inertial navigation position correction parameter sequence includes a position difference fitting slope set, a compensation offset set, and a synchronization time point index set. The behavior... The path coherence mutation segment includes a mutation window number group, a window span group, and a window coherence discrete group. The tactical action execution anomaly segment includes an anomaly action number group, an anomaly segment duration group, and a sequence offset amplitude group. The tactical phase effectiveness state mapping table includes a phase execution delay ratio group, a phase sequence disturbance index group, and a phase state identifier group.

3. The intelligent positioning-type performance monitoring device for tactical training according to claim 1, characterized in that, The inertial navigation drift recognition module includes: The inertial data acquisition submodule collects linear acceleration and angular velocity data for each time slice in the inertial navigation device. Based on the acquisition time sequence, it forms multiple sets of acceleration vector sequences and angular velocity vector sequences. It calculates the change of each set of vectors in the time axis direction and constructs a linear velocity change sequence. At the corresponding position of each time slice, it performs velocity value difference processing to obtain the linear velocity change rate values ​​of two adjacent time slices and constructs a velocity change rate data stream under a continuous time axis, generating a velocity change rate sequence value. The speed change determination submodule, based on the speed change rate sequence value, calls each value of the speed change rate in the continuous time slice, calculates the difference between it and the set speed change rate critical threshold, performs a judgment on the magnitude of the difference and zero value, filters the set of time slices that meet the condition that the difference is less than zero value, obtains the direction angle data corresponding to the time slice, calculates the adjacent time difference value of the direction angle and judges whether the angle change value is less than the heading change threshold, removes the time slices that do not meet the angle change determination condition, and obtains a set of low speed change direction stable time slices; The stable interval filtering submodule calls the set of stable time slices in the low-speed direction, divides the time slice set into continuous time periods according to the principle of time slice continuity, records the start index, end index and duration of each time interval, sorts each time interval according to the start and end index, constructs a continuous navigation time period interval set structure and identifies its sequence number, and establishes an inertial navigation stable drift interval identifier set.

4. The intelligent positioning-type performance monitoring device for tactical training according to claim 3, characterized in that, The specific method for setting the critical threshold of the rate of change of velocity is as follows: obtain the rate of change of velocity values ​​within multiple consecutive time slices, calculate the sum of the arithmetic mean and standard deviation of all the rate of change of velocity values, and use it as the critical threshold of the rate of change of velocity. The heading change threshold is set by obtaining the difference sequence between the heading angle in each time slice and the heading angle in the adjacent time slice, and calculating the angle value of the 70th decimal place in the heading angle difference sequence as the heading change threshold.

5. The intelligent positioning-type performance monitoring device for tactical training according to claim 1, characterized in that, The dynamic position correction module includes: The coordinate difference calculation submodule obtains the estimated position coordinates of the inertial navigation system and the actual position coordinates provided by the GNSS positioning module within each drift interval based on the inertial navigation stable drift interval identifier set. It extracts two position coordinates at the same time point according to the timestamp alignment rule, calculates the Euclidean distance value in three-dimensional space, and arranges them in time slice order to generate continuous position error data and establishes the Euclidean distance difference sequence value. The position drift trend extraction submodule calls the Euclidean distance difference sequence value, performs least squares linear fitting on the continuous error value in each time interval, obtains the slope value of the corresponding fitted line, extracts the slope set corresponding to each interval, compares each slope value with the set position drift change rate threshold, filters the time intervals that meet the condition that the slope is greater than the threshold, and obtains the drift trend interval set value. The compensation interval construction submodule extracts the estimated inertial navigation coordinates and error growth slope values ​​corresponding to the start time index and end time index based on the drift trend interval set value, establishes a linear compensation structure based on the current coordinates and slope values, constructs a compensation correction parameter list according to the segment number, and generates an inertial navigation position correction parameter sequence.

6. The intelligent positioning-type performance monitoring device for tactical training according to claim 1, characterized in that, The behavior path monitoring module includes: The behavior data extraction submodule calls the inertial navigation position correction parameter sequence. In the corrected inertial navigation trajectory, it divides the unit time window according to the time slice order, extracts the displacement value, angular velocity sequence value and command response timestamp in each window, calculates the moving distance value, average angular velocity and control response delay time value in each window, and combines and arranges them according to the window number to generate a three-dimensional behavior index sequence set. The continuous vector construction submodule constructs a continuous vector group across windows based on the three-dimensional behavioral indicator sequence set in a time sliding window manner. It extracts the corresponding values ​​of the three behavioral indicators between adjacent windows to form a difference vector, calculates the standard deviation of the three elements of the difference vector, obtains the standard deviation change value of each vector group, and arranges the results in time order to generate a standard deviation change sequence. The continuity mutation determination submodule extracts each standard deviation change value in the order of time windows based on the standard deviation change sequence, compares it with the set path continuity threshold, marks the window number where the standard deviation change is greater than the path continuity threshold, constructs an abnormal path segment number set by combining the numbers and adds index time range information to generate behavioral path continuity mutation segments.

7. The intelligent positioning-type performance monitoring device for tactical training according to claim 1, characterized in that, The action offset determination module includes: The task instruction sequence extraction submodule extracts the tactical instruction code and execution timestamp corresponding to each segment based on the behavioral path coherence abrupt segment, constructs the action execution sequence according to the chronological order and records the index position of each instruction in the path, extracts the standard action chain sequence table from the standard tactical task structure and aligns the index numbers to generate a standard comparison action sequence set. The offset feature calculation submodule matches each execution instruction with the corresponding number in the standard action chain according to the index position of the standard action sequence set. It calculates the index difference of the same instruction item in the two sequences and records it as the sequence offset. It extracts the number span of adjacent actions in the standard sequence as the jump amplitude. It counts the offset value and jump amplitude value of each segment to obtain the action offset feature sequence. The abnormal action segment identification submodule calls the action offset feature sequence, compares the sequence number offset value with the action tolerance limit threshold, determines whether the jump amplitude value exceeds the instruction cycle judgment boundary threshold, filters the action sequence segments that meet both conditions, extracts and combines the segment number and time range into an abnormal identification structure, and generates tactical action execution abnormal segments.

8. The intelligent positioning-type performance monitoring device for tactical training according to claim 1, characterized in that, The tactical effectiveness assessment module includes: The task phase index calculation submodule calls the duration value, the tactical phase identifier value, and the standard duration value of the corresponding phase in the standard task flow for each segment in the abnormal tactical action execution segment. It classifies all abnormal segments by phase, calculates the ratio between the abnormal duration and the standard duration in each phase, and establishes an index mapping by phase number to generate a phase execution delay ratio sequence. The sequence perturbation distance extraction submodule extracts the abnormal action sequence and standard action sequence under each stage based on the determined set of tactical stage numbers in the stage execution delay ratio sequence, performs dynamic time warping according to the time alignment rules, calculates the cumulative distance value under the action alignment path, and records the perturbation measure corresponding to each stage as an index value to obtain the tactical sequence perturbation distance sequence. The performance status mapping construction submodule extracts two corresponding indicator values ​​for each tactical stage based on the tactical sequence perturbation distance sequence and the stage execution delay ratio sequence, compares them with preset state judgment thresholds, identifies whether each indicator exceeds the limit, marks the exceeding state with Boolean values, binds the judgment result with the corresponding tactical stage number, constructs a structured result table, and generates a tactical stage performance status mapping table.

9. The intelligent positioning-type performance monitoring device for tactical training according to claim 8, characterized in that, The specific method for setting the state judgment threshold is as follows: based on the distribution of the perturbation distance values ​​at each stage in the tactical sequence perturbation distance sequence, the mean and standard deviation of all perturbation distance values ​​are extracted, and the sum of the mean and standard deviation is superimposed as the sequence perturbation distance threshold. Based on the statistical results of the ratio values ​​of each stage in the stage execution delay ratio sequence, the 90th percentile of the delay ratio value distribution is extracted as the stage execution delay ratio threshold. The sequence perturbation distance threshold and the stage execution delay ratio threshold are respectively used as two components of the preset state judgment threshold; The index value corresponding to each tactical phase in the tactical sequence perturbation distance sequence and the phase execution delay ratio sequence must be compared with the two components mentioned above. And when either indicator value in the two comparisons is greater than the corresponding threshold, a Boolean value is generated to mark the result as logically true.

10. An intelligent positioning-based performance monitoring application method for tactical training, characterized in that, The intelligent positioning-type performance monitoring device for tactical training as described in any one of claims 1-9 shall be executed. Includes the following steps: S1: Collect linear acceleration and angular velocity data for each time slice in the inertial navigation device, calculate the difference between the rate of change of velocity and the critical threshold of the rate of change of velocity, and determine whether the difference is lower than the critical threshold of the rate of change of velocity. Determine whether the angle of change of direction within the time period is lower than the heading change threshold, filter the navigation time periods that meet the dual threshold conditions, and generate a set of inertial navigation stable drift interval identifiers. S2: Based on the inertial navigation stable drift interval identifier set, obtain the inertial navigation estimated position value and the position value provided by the GNSS positioning device within the corresponding time period of the identifier set, calculate the Euclidean distance difference sequence of the two types of positions at the same time point, and count the linear fitting slope of the difference sequence, extract the interval that can be constructed for linear compensation, and generate the inertial navigation position correction parameter sequence. S3: Call the inertial navigation position correction parameter sequence, extract the average travel distance, angular velocity and control response delay time, construct the behavior continuity vector group of the three indicators under the corresponding time window, calculate the standard deviation change between the vector groups, mark the window segment where the path change standard deviation change exceeds the path coherence threshold, and generate the behavior path coherence abrupt change segment. S4: Based on the behavioral path coherence abrupt segment, combined with the standard action chain sequence table preset in the tactical task, perform position comparison and index distance calculation, filter out segments that meet the condition that the offset is greater than the action tolerance limit and the jump amplitude exceeds the instruction cycle judgment boundary, and generate tactical action execution abnormal segments. S5: Invoke the duration, number of segments and tactical phase identifier of the abnormal segment in the execution abnormal segment of the tactical action, calculate the ratio of the abnormal segment duration to the standard duration, obtain the execution delay ratio, compare the dynamic time warping distance value of the abnormal action sequence and the standard action sequence as the sequence perturbation index, determine whether the phase has entered the abnormal state area, and generate a tactical phase effectiveness state mapping table.

Images