Backpack guest operation track automatic generation method, device and system and medium

Abstract

The invention relates to the technical field of operation tracks, and provides a knapsack guest operation track automatic generation method, device and system and a medium. According to the implementation scheme, non-line-of-sight error compensation is carried out on an original multi-mode observation data set of the backpack guest, and a compensation track fragment data set is obtained; on the basis of the compensation trajectory fragment data set, constructing a gait trajectory of the knapsack guest to obtain each structured trajectory fragment; integrating each structured trajectory fragment to obtain a balanced trajectory fragment stream; performing fault probability calculation on each first risk fragment in the balanced trajectory fragment flow to obtain a predicted fault probability sequence; and according to the balanced trajectory fragment flow and the predicted fault probability sequence, a work trajectory report of the backpack guest is generated, and the work trajectory report comprises a path curve graph and a risk point list. According to the embodiment of the invention, the potential fault risk in the operation process of the knapsack guest can be quantitatively evaluated.

Description

Technical Field

[0001] This invention relates to the field of work trajectory technology, and in particular to a method, apparatus, system and medium for automatically generating work trajectories for backpackers. Background Technology

[0002] Backpackers performing tasks in complex environments (such as inspection robots) typically rely on multimodal sensors (including vision, inertial, localization, and energy consumption monitoring) to collect motion trajectories and status information during operations. However, due to factors such as obstacle obstruction, multipath propagation, and non-line-of-sight conditions, sensor observation data often suffers from positioning drift or trajectory interruptions, resulting in significant discontinuities and error accumulation in the acquired trajectory data.

[0003] Furthermore, existing methods for managing backpacker work routes mostly focus on reconstructing the geometric shape of the route itself, lacking effective means of perceiving and assessing potential fault risks caused by environmental interference, equipment malfunctions, and other factors during the work process. Currently, risk assessment often relies on the qualitative experience of back-end personnel or simple threshold alarms, making it difficult to conduct in-depth correlation and quantitative analysis of discrete and sporadic abnormal signals, let alone proactively predict the probability of route distortion, interruptions, and other faults occurring in specific road sections or time periods. Summary of the Invention

[0004] This invention provides a method, apparatus, system, and medium for automatically generating backpacker work trajectories, which can solve at least one of the above-mentioned technical problems.

[0005] In a first aspect, embodiments of the present invention provide a method for automatically generating backpacker work trajectories, including: Non-line-of-sight error compensation was performed on the original multimodal observation dataset of backpackers to obtain a compensated trajectory segment dataset; Based on the compensated trajectory fragment dataset, the backpacker's gait trajectory is constructed to obtain various structured trajectory fragments; The structured trajectory segments are integrated to obtain a balanced trajectory segment flow; The failure probability is calculated for each first risk segment in the balanced trajectory segment stream to obtain a predicted failure probability sequence. Based on the balanced trajectory segment stream and the predicted failure probability sequence, a work trajectory report for the backpacker is generated, wherein the work trajectory report includes a path curve and a list of risk points.

[0006] Secondly, embodiments of the present invention provide an automatic backpacker work trajectory generation device, comprising: The error compensation module is used to perform non-line-of-sight error compensation on the backpacker's original multimodal observation dataset to obtain a compensated trajectory segment dataset. The trajectory construction module is used to construct the backpacker's gait trajectory based on the compensated trajectory fragment dataset to obtain various structured trajectory fragments; The trajectory segment integration module is used to integrate the various structured trajectory segments to obtain a balanced trajectory segment flow; The fault probability calculation module is used to calculate the fault probability of each first risk segment in the balanced trajectory segment stream to obtain a predicted fault probability sequence. The generation module is used to generate a work trajectory report for the backpacker based on the balanced trajectory segment flow and the predicted failure probability sequence, wherein the work trajectory report includes a path curve and a list of risk points.

[0007] Thirdly, embodiments of the present invention also provide an automatic backpacker work trajectory generation system, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method described in any one of the embodiments of the present invention.

[0008] Fourthly, embodiments of the present invention also provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the method described in any one of the embodiments of the present invention.

[0009] This invention employs a non-line-of-sight error compensation method on the original multimodal observation dataset of backpackers. This corrects positioning errors caused by environmental occlusion or signal reflection, resulting in a continuous and high-precision compensated trajectory segment dataset. Based on this dataset, the backpackers' gait trajectories are constructed to obtain structured trajectory segments. By integrating these structured trajectory segments, a balanced trajectory segment stream is generated, eliminating temporal misalignment and displacement abrupt changes between multi-source data, resulting in a smoother and more continuous overall trajectory sequence. Furthermore, by calculating the failure probability of each first-risk segment in the balanced trajectory segment stream, the probabilistic relationship between trajectory anomalies and potential risks can be derived, yielding a predicted failure probability sequence characterizing risk intensity. Finally, a work trajectory report is generated based on the balanced trajectory segment stream and the predicted failure probability sequence. The path curves and risk point lists in the work trajectory report enable quantitative expression and visualization of high-risk spatiotemporal segments during the work process. This allows for a quantitative assessment of potential failure risks during backpacker operations.

[0010] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0011] The accompanying drawings are provided for a better understanding of this solution and do not constitute a limitation of the invention. Wherein: Figure 1 This is a flowchart of a method for automatically generating backpacker work trajectories according to an embodiment of the present invention; Figure 2 This is a structural block diagram of a backpacker's work trajectory automatic generation device according to an embodiment of the present invention; Figure 3 This is a schematic block diagram of an electronic device used to implement the methods of embodiments of the present invention. Detailed Implementation

[0012] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0013] It should be noted that the "backpacker" referred to in this invention refers to a mobile work unit equipped with multiple sensing devices in a specific work scenario. This can be a single worker, an inspection robot, or an autonomous device carrying a mobile terminal. These sensing devices include, for example, inertial measurement units (IMUs), Global Navigation Satellite System (GNSS) positioning modules, lidar, and multi-source sensors (such as environmental sensors like magnetometers and barometers, and foot pressure sensors). During operation, the backpacker continuously records its movement trajectory, posture changes, environmental images, and energy consumption status through these sensing devices. For example, in a power line inspection scenario, the backpacker could be a human inspector carrying a positioning backpack, equipped with a GNSS receiver, an IMU, an infrared camera, and a barometer; in an unmanned warehouse scenario, the backpacker could be an autonomous mobile robot performing material handling tasks, with its backpack module including lidar and an ultra-wideband positioning module.

[0014] Figure 1 This is a flowchart of a method for automatically generating backpacker work trajectories according to an embodiment of the present invention.

[0015] like Figure 1As shown, the method for automatically generating backpacker work routes may include: S110 performs non-line-of-sight error compensation on the backpacker's original multimodal observation dataset to obtain a compensated trajectory segment dataset; S120, based on the compensated trajectory fragment dataset, constructs the gait trajectory of backpackers to obtain various structured trajectory fragments; S130, integrate the various structured trajectory segments to obtain a balanced trajectory segment flow; S140, calculate the failure probability of each first risk segment in the balanced trajectory segment flow to obtain the predicted failure probability sequence; S150, based on the balanced trajectory segment flow and the predicted failure probability sequence, generates a backpacker's work trajectory report, which includes a path curve and a list of risk points.

[0016] For example, the original multimodal observation dataset refers to a collection of data containing various types of observation information collected by backpackers during their work, including but not limited to: positioning modal data: latitude and longitude coordinate data from GNSS; inertial modal data: acceleration, angular velocity, and attitude angle data from an inertial measurement unit (IMU); visual modal data: image frame sequences from a visual camera or infrared camera; ranging modal data: distance measurement data from a LiDAR (Light Detection and Range) or Ultra Wide Band (UWB) positioning module; and energy consumption modal data: such as equipment current, voltage, or power consumption information.

[0017] Specifically, the original multimodal observation dataset mainly comes from: airborne / human-borne sensors: synchronous output data from sensor modules such as GNSS, IMU, LiDAR, cameras, and barometers installed on backpacks or work equipment; edge computing terminals: field edge gateways used to collect and cache multi-source data in real time, such as industrial laptops or embedded computing boxes; remote monitoring systems: real-time work data streams uploaded from backpackers via wireless communication, such as fifth-generation mobile communication technology (5G), wireless Fidelity (Wi-Fi), or long-range low-power wireless communication technology (LoRa); and historical databases: cloud servers storing historical task records. For example, in power line inspection, GNSS modules collect location information, IMU modules record attitude and gait frequency, and infrared cameras record surface temperature images of equipment. All data is fused by edge computing terminals and stored in the task database to form the original multimodal observation dataset.

[0018] For example, a compensated trajectory segment dataset refers to a set of corrected and smoothed trajectory segments obtained after non-line-of-sight error compensation processing of the original multimodal observation data. Each trajectory segment contains spatial coordinates, attitude, velocity, and time information corrected by multimodal fusion within a continuous time window, which can be used for subsequent trajectory construction and risk analysis. For instance, assuming that the GNSS signal drifts due to obstruction by tall buildings during a certain period, the drift error of that segment can be corrected by IMU integration calculation and visual feature matching. The corrected path data is a "compensated trajectory segment".

[0019] For example, in a power equipment inspection scenario, the "backpacker" (or "backpacker") acts as a field inspector or inspection robot, moving through substations and transmission line corridors to collect multi-source data. Due to obstructions from trees, high-voltage towers, and other structures, GNSS signals are prone to non-line-of-sight (LOS) errors. Through LOS error compensation in step S110, the actual inspection path can be restored, ensuring accurate fault location. In a tunnel or indoor building surveying scenario, the "backpacker" acts as a surveying robot equipped with Simultaneous Localization and Mapping (SLAM) equipment. In underground tunnels, GNSS signals are lost; trajectory compensation is achieved through IMU and LiDAR fusion to obtain a continuous spatial path. In a disaster search and rescue scenario, the "backpacker" acts as a rescue robot. When moving through ruins or mountainous environments, a multimodal perception system compensates for positioning anomalies caused by obstructed line-of-sight, thereby accurately reconstructing the search and rescue trajectory and marking high-risk areas.

[0020] According to the above implementation method, by performing non-line-of-sight error compensation on the original multimodal observation dataset of backpackers, the positioning deviation caused by signal obstruction or reflection can be effectively corrected, improving the spatial accuracy of trajectory data. Furthermore, by constructing gait trajectories and forming structured segments based on the compensated trajectory segment dataset, temporal alignment and spatial fusion of different modal information can be achieved, ensuring the continuity and consistency of trajectory information. By integrating and equalizing the structured trajectory segments, a complete trajectory segment stream is obtained, providing a stable data foundation for subsequent risk identification. Based on the calculation of the failure probability of the first risk segment of the balanced trajectory segment stream, the potential abnormal risks in the trajectory can be quantitatively assessed. Finally, by combining the predicted failure probability sequence, a work trajectory report containing a path curve and a list of risk points is generated, realizing the visualization of trajectory information and risk indicators.

[0021] In one implementation, non-line-of-sight error compensation is performed on the original multimodal observation dataset of backpackers to obtain a compensated trajectory segment dataset. This includes: identifying anomalies in each modal signal in the original multimodal observation dataset to obtain the abnormal time periods of each modal signal; extracting abnormal coordinate points corresponding to the abnormal time periods from the original multimodal observation dataset based on the abnormal time periods of each modal signal to obtain each abnormal coordinate point; performing trajectory extension prediction on each abnormal coordinate point using a state estimation algorithm to obtain each compensated coordinate point; and replacing each abnormal coordinate point based on each compensated coordinate point to update the original multimodal observation dataset to obtain the compensated trajectory segment dataset.

[0022] For example, the original multimodal observation dataset includes time-series signal data synchronously acquired by multiple sensors, such as: inertial measurement unit (IMU) signals, including acceleration and angular velocity; satellite positioning signals (GPS or BeiDou), including latitude and longitude, positioning accuracy factors (e.g., horizontal accuracy factors); visual or lidar (LiDAR) signals, including pose estimation results, depth information, etc.; wireless signal strength signals, visual frame sequence signals, geomagnetic signals, and barometric pressure signals, etc. In complex scenarios, some modal signals may abruptly change or be lost due to occlusion, reflection, signal attenuation, or device drift. Anomaly detection algorithms (such as sliding window standard deviation detection, Z-score detection, or anomaly recognition models based on long short-term memory networks) can be used to identify anomalous segments of the signal in the time dimension. For example, when the positioning accuracy factor of the GPS signal is higher than 5.0 for 3 consecutive seconds, this time period can be determined as an anomalous period.

[0023] For example, the method for determining abnormal time periods can also be any of the following: First, for wireless signal strength signals, a moving average filtering algorithm is applied to smooth the signal strength within a continuous observation window, and the mean deviation within the smoothed window is calculated. When the mean signal strength deviates from the normal fluctuation range (e.g., normalized fluctuation range between 0.25 and 0.75), the time window is determined to be an abnormal time period. Second, for visual frame sequence signals, an edge detection algorithm is used, such as the Canny Edge Detection Operator (Canny) or the Sobel Edge Detection Operator (Sobel). Edge extraction is performed on adjacent frames according to the selected edge detection algorithm, and the inter-frame brightness change rate is calculated. If a sharp drop in the brightness change rate is detected within a short time window, it indicates that the visual scene is occluded or the illumination has changed significantly, and the time period corresponding to the frame sequence segment can be recorded as an abnormal time period. Third, state estimation models are constructed for inertial measurement signals, geomagnetic signals, and barometric pressure signals, respectively, and Kalman filtering is used for trajectory prediction. The system calculates the expected observation value at the current moment based on historical state variables (position, velocity, acceleration, attitude angle) and compares it with the actual observation value. When the Euclidean distance or attitude angle difference between the predicted trajectory and the observed trajectory exceeds a preset threshold, the signal for that time period is determined to be abnormal. For example, in geomagnetic signals, if the expected magnetic declination is ±2° and the actual observed deviation reaches ±8°, it indicates interference from a metal structure, and that time period is marked as an abnormal period.

[0024] For example, an anomaly coordinate point refers to a spatial coordinate data (such as latitude and longitude or three-dimensional position coordinates) output by the positioning module during an abnormal period that deviates significantly from the actual location. For instance, at t=5 seconds, the GPS output location is (x=102.35, y=30.22), where x is the horizontal coordinate and y is the vertical coordinate. However, the location calculated by the IMU should be (x=102.33, y=30.21). The deviation exceeds 1.5 meters, so this coordinate point is marked as an anomaly coordinate point. The extraction process can be achieved through index matching: the system retrieves coordinate records with timestamps within the range specified in the original multimodal observation dataset based on the start and end times of the abnormal period and marks them as the set of anomaly coordinate points.

[0025] For example, the compensation coordinate point refers to the corrected coordinate value calculated by the trajectory prediction model to replace the abnormal coordinate point. This compensation point is calculated based on the state information of the normal trajectory before and after the abnormal point to ensure the smoothness and continuity of the trajectory.

[0026] In this example, the state estimation algorithm can employ methods such as Kalman filtering, extended Kalman filtering, or particle filtering. The algorithm models historical coordinates, velocity, acceleration, and other state variables, and combines this with the observation model to predict the reasonable location of anomalies. For example, if the GPS coordinates are abnormal at time t=5 seconds, the system uses the trajectory states (position and velocity) at t=4s and t=6s as input to the Kalman filter, predicting the optimal estimated location at t=5s as (x=102.334, y=30.215), which is then the compensated coordinate point.

[0027] For example, outlier coordinates are replaced with corresponding compensation coordinates to generate updated trajectory data. During the replacement process, to avoid discontinuities between the compensation points and the original trajectory, interpolation smoothing (such as spline interpolation or weighted moving average) can be used for trajectory reconstruction. The replaced trajectory is segmented and saved as a compensation trajectory fragment dataset, where each fragment contains information such as timestamp, compensation coordinate sequence, modality type, and confidence level. For example, the outlier coordinates at time t=5s (x=102.35, y=30.22) in the original trajectory are replaced with compensation coordinates (x=102.334, y=30.215), and the updated trajectory fragments are: [(t=4s, 102.33, 30.21), (t=5s, 102.334, 30.215), (t=6s, 102.337, 30.218)]. This creates a continuous and reliable compensation trajectory fragment dataset.

[0028] According to the above implementation method, firstly, the abnormal time periods affected by interference in each modal signal are identified, and the corresponding abnormal coordinate points are extracted. Then, a state estimation algorithm is used to predict the trajectory of these abnormal points, generating compensating coordinate points. Finally, the abnormal coordinate points are replaced with the compensating coordinate points, thereby updating the original dataset and forming a continuous and reliable compensated trajectory segment dataset. This effectively reduces trajectory interruptions and deviations caused by non-line-of-sight interference, signal loss, or sensor errors, significantly improving the integrity and continuity of trajectory data, and providing high-quality input for subsequent gait reconstruction, risk analysis, and work report generation.

[0029] In one implementation, the gait trajectory of a backpacker is constructed based on a compensated trajectory fragment dataset to obtain various structured trajectory fragments. This includes: acquiring a compensated trajectory fragment dataset, which includes a compensated coordinate sequence, an original coordinate sequence, modal type, and timestamp information; expanding the original coordinate sequence according to the time axis to obtain a continuous time observation sequence; performing differential calculation on the compensated coordinate sequence and the continuous time observation sequence to obtain displacement vectors between adjacent time references; aligning each displacement vector with the current observation values ​​from multiple source sensors using coordinate transformation to obtain an alignment result; establishing a joint index based on the alignment result, modal type, and timestamp information to obtain a time-series fusion index set; fusing and connecting the compensated coordinate sequence and the corresponding trajectory data of the gait trajectory based on the correspondence of each index in the time-series fusion index set to obtain a coherent path coordinate point sequence; segmenting the path coordinate point sequence to obtain various short-time trajectory fragments; and encapsulating each short-time trajectory fragment to obtain various structured trajectory fragments.

[0030] For example, a time-continuous observation sequence refers to a continuous time series formed by filling in time gaps caused by differences in sampling frequency or lost observations in the original coordinate sequence through time interpolation or resampling algorithms. Specifically, methods such as linear interpolation, spline interpolation, or Kalman prediction can be used to fill in time intervals in the original sequence where adjacent time intervals exceed a threshold (e.g., 100 milliseconds), so that each time step on the time axis corresponds to a coordinate point. For example, if there is a data gap between t=1.0s and t=1.3s in the original coordinate sequence, estimated coordinate points at t=1.1s and t=1.2s can be inserted based on the displacement trend of the two adjacent points, thereby forming a time-continuous observation sequence.

[0031] For example, the displacement vector refers to the coordinate difference between adjacent time points, used to reflect the trend of motion change. The system calculates the coordinate difference between adjacent time points for the compensated coordinate sequence and the continuous time observation sequence respectively, to obtain the displacement vector under the corresponding time reference. For example, when the backpacker moves from position (3.0, 1.2, 0.5) to (3.4, 1.5, 0.6), the obtained displacement vector is (0.4, 0.3, 0.1).

[0032] In this example, the coordinate difference can be calculated using Euclidean distance or vector difference. For example: In the formula, Let be the backpacker's displacement vector between adjacent time bases; For backpackers at all times The three-dimensional coordinates of the points; , and The backpacker's three-dimensional coordinates at the current time reference point The three components in the spatial coordinate system represent the position along the X-axis, Y-axis, and Z-axis, respectively. , and The backpacker's three-dimensional coordinates for the next time reference point Position in the X, Y, and Z axes.

[0033] For example, coordinate transformation can specifically employ quaternion rotation matrices, homogeneous coordinate transformations, or spatial alignment based on multi-sensor extrinsic calibration results. For instance, the displacement vector in the IMU coordinate system can be transformed to the visual camera coordinate system using a rotation matrix to achieve a unified representation between modalities. Specifically, for example, when the coordinates of the visual modality at time t=5.0s are (4.2, 2.0, 0.5), and the predicted coordinates of the IMU modality at the same time are (4.15, 1.98, 0.52), after coordinate transformation and alignment, a joint index item is formed: {timestamp:5.0, modality:'fusion', position:(4.18, 1.99, 0.51)}, where timestamp represents the sampling time of the observation point on the time axis; modality represents the modality type, indicating which sensing modality or fusion method the data point comes from, such as vision, inertial (IMU), fusion, etc.; 'fusion' indicates that the point is the result of multi-source sensor fusion calculation; and position represents the spatial coordinates, that is, the backpacker's three-dimensional spatial position at the current timestamp, in Cartesian coordinate form.

[0034] For example, a temporal fusion index set refers to an index set formed by matching and associating multimodal observation data under a unified time axis and coordinate system. Each index item in the index set records a timestamp, the corresponding modality type, the coordinate transformation matrix (or attitude parameters), and the coordinate alignment result (fused coordinates).

[0035] For example, the path coordinate point sequence refers to a continuous set of coordinate points arranged in chronological order after multimodal fusion and compensation processing, reflecting the backpacker's movement path. The system fuses coordinate data from different modalities (such as visual and inertial) based on the timestamp matching relationship in the temporal fusion index set. For example, when the time difference between two modalities is less than 50 milliseconds, a weighted average or complementary filtering method can be used to calculate the final fused coordinates. Specifically, if the visual modality coordinates are (5.0, 3.2, 0.4) and the IMU modality coordinates are (5.1, 3.1, 0.5), then the fusion result... The weights can be set according to actual needs, for example, 0.6 and 0.4. By connecting these fused coordinates in sequence, a smooth and continuous sequence of path coordinate points can be obtained.

[0036] For example, a short-time trajectory segment refers to a small segment of trajectory unit into which a long-term path coordinate sequence is divided according to temporal or spatial characteristics. The division criteria may include: a fixed time window (e.g., dividing every 5 seconds); a fixed path length (e.g., dividing every 10 meters); and state change detection (e.g., points of sudden velocity change or directional inflection). For instance, if a backpacker's path is 500 meters long, it can be divided into 10-meter segments, resulting in 50 short-time trajectory segments, each reflecting a smooth travel trajectory.

[0037] For example, a structured trajectory segment refers to a structured data unit formed by adding statistical features and environmental identification information of the trajectory to a short-time trajectory segment. The content of the structured data unit may include: average velocity, direction angle, step frequency; modal combination (visual + inertial + wireless signal, etc.); start and end timestamps, start and end coordinates; anomaly annotation (whether it contains compensation points). For example, a structured trajectory segment can be represented as: {start time: 12.500 seconds, end time: 17.500 seconds, modal type: ["camera", "inertial measurement unit"], average velocity: 1.3 meters per second, heading angle: 45 degrees, segment type: "compensated segment"}.

[0038] According to the above implementation method, by performing multi-stage fusion and structured processing on the compensation trajectory segment dataset, the steps of time axis expansion, displacement difference calculation, coordinate transformation alignment, time series index establishment and trajectory fusion segmentation are completed in sequence, thereby achieving high-precision alignment and continuous trajectory construction of different modal data in time and space, so as to provide high-quality input data for subsequent trajectory visualization.

[0039] In one implementation, integrating the structured trajectory segments to obtain a balanced trajectory segment stream includes: sorting the structured trajectory segments in ascending order according to their start timestamps to obtain a first trajectory segment sequence; connecting adjacent first trajectory segments in the trajectory segment sequence if they are discontinuous to obtain a second trajectory segment sequence; and performing smooth interpolation processing at the splicing points of the second trajectory segments in the second trajectory segment sequence to obtain a balanced trajectory segment stream.

[0040] For example, the first trajectory segment sequence refers to the set of trajectory segments obtained by arranging multiple structured trajectory segments in chronological order according to their start timestamps from earliest to latest. In this example, for the encapsulated structured trajectory segments (each segment contains a timestamp interval, coordinate sequence, and gait information), the start timestamps of each structured trajectory segment are stored in a time index table; the segments are sorted in ascending order using quicksort or a database query based on key-value indexes; the sorted results are verified for temporal continuity, and the time intervals between adjacent segments are marked. This step ensures the logical coherence of the trajectory segments in the time dimension, providing a foundation for subsequent trajectory splicing and smooth interpolation.

[0041] For example, if there are three structured trajectory segments: segment A: start time 07:55, end time 08:05; segment B: start time 08:10, end time 08:20; segment C: start time 07:40, end time 07:50; then the ascending sort result is {segment C, segment A, segment B}, and the corresponding first trajectory segment sequence time order is 07:40-08:20.

[0042] For example, the second trajectory segment sequence refers to the sequence obtained by connecting adjacent segments in the first trajectory segment sequence that have time gaps or time overlaps.

[0043] For example, a balanced trajectory segment flow refers to a temporally and spatially continuous trajectory flow data set formed after smoothing interpolation processing of each splicing point in the second trajectory segment sequence. Smoothing interpolation processing is performed at each splicing point in the connected second trajectory segment sequence. Its purpose is to eliminate trajectory discontinuities caused by differences in sampling frequency, time drift, or sensor noise when splicing different segments. Specifically, smoothing interpolation can be implemented based on any of the following methods: First, spline interpolation: fitting the connection points of adjacent trajectory segments using a cubic spline function to make the first and second derivatives of the trajectory continuous at the splicing points. Second, weighted moving average method: smoothing the coordinate points within the splicing window according to time weights to reduce abrupt changes. Third, Kalman filtering: performing state estimation and prediction updates on the trajectory point sequence to suppress noise abrupt changes at the splicing points.

[0044] For example, suppose the second trajectory segment sequence contains: Segment 1: coordinate sequence [(X1, Y1), (X2, Y2), (X3, Y3)]; Segment 2: coordinate sequence [(X4, Y4), (X5, Y5), (X6, Y6)], where there is a positional difference of 0.8 meters between (X3, Y3) and [(X4, Y4)]. Cubic spline interpolation is used to generate transition points (X3.1, Y3.1), (X3.2, Y3.2), and (X3.3, Y3.3) to achieve a smooth connection between the two segments, resulting in a temporally and spatially continuous balanced trajectory segment flow.

[0045] In one embodiment, when two adjacent first trajectory segments in the first trajectory segment sequence are discontinuous, connecting the two adjacent first trajectory segments to obtain a second trajectory segment sequence includes: obtaining the timestamps of two adjacent first trajectory segments, wherein the timestamps of the two adjacent first trajectory segments include the end timestamp of the third trajectory segment and the start timestamp of the fourth trajectory segment; if the start timestamp of the fourth trajectory segment is greater than the end timestamp of the third trajectory segment, then inserting data points into the gap between the third and fourth trajectory segments to connect the third and fourth trajectory segments; if the start timestamp of the fourth trajectory segment is less than the end timestamp of the third trajectory segment, then performing weighted fusion on the coordinate point data of the third and fourth trajectory segments in the time overlap interval to connect the third and fourth trajectory segments.

[0046] For example, the sequence of first trajectory segments is traversed, and the timestamps of two adjacent first trajectory segments are extracted in turn, which is also the end timestamp of the previous segment (the third trajectory segment). and the start timestamp of the next segment (the fourth trajectory segment) For example, if the time of the third trajectory segment is [08:00, 08:10] and the time of the fourth trajectory segment is [08:12, 08:25], then: the end timestamp of the third trajectory segment. =08:10; Start timestamp of the fourth trajectory segment =08:12. Since the time difference between the two is 2 minutes, there is a time gap between the two trajectories.

[0047] For example, if the start timestamp of the fourth trajectory segment is greater than the end timestamp of the third trajectory segment, the system generates a virtual connection point at the gap between the third and fourth trajectory segments using an interpolation method, and inserts this virtual connection point as a data point into the sequence. The interpolation method may include, but is not limited to, the following traditional implementations: linear interpolation: calculating linearly changing trajectory points within the interval based on the boundary coordinates of the two segments; spline interpolation: generating a transition point sequence by fitting a cubic spline curve when high smoothness of the spatial curve is required; inertial compensation interpolation: predicting the direction of travel and velocity by combining acceleration and angular velocity data from the inertial measurement unit (IMU), generating an interpolated trajectory that conforms to the laws of motion.

[0048] For example, suppose the boundary coordinates corresponding to the end timestamp of the third trajectory segment (08:10) are (10.0, 5.0); and the boundary coordinates corresponding to the start timestamp of the fourth trajectory segment (08:12) are (13.0, 5.5). Using linear interpolation, a virtual point (11.5, 5.25) is generated at 08:11 and inserted into the trajectory to connect the third and fourth trajectory segments, thus forming a second trajectory segment sequence of a continuous path {(10.0, 5.0), (11.5, 5.25), (13.0, 5.5)}.

[0049] For example, if the start timestamp of the fourth trajectory segment is less than the end timestamp of the third trajectory segment, then the coordinate point data of the third and fourth trajectory segments in the time overlap interval are weighted and fused to remove duplication and smooth the transition. The weighted fusion can be based on the following calculation method: In the formula, The position of the merged trajectory point, where, The coordinates are on the X-axis (horizontal coordinate axis). The coordinates are on the Y-axis (the vertical coordinate axis); These are the original observation coordinates of the third trajectory segment; Here are the original observation coordinates of the fourth trajectory segment; where, and Let X be the coordinates of the corresponding trajectory segment on the X-axis. and This represents the Y-coordinate of the corresponding trajectory segment; and The weighting coefficients for the corresponding trajectory segments can be preset based on expert experience, for example... Set it to 0.6. Set it to 0.4.

[0050] According to the above implementation method, this step selects different trajectory connection strategies by comparing the timestamp relationship between two adjacent trajectory segments (the third trajectory segment and the fourth trajectory segment): when there is a time gap, missing data is filled by interpolation and other methods to achieve temporal continuity of the trajectory; when there is time overlap, overlapping segments are smoothly merged by weighted fusion to eliminate redundant data.

[0051] In one implementation, the failure probability of each first risk segment in the balanced trajectory segment stream is calculated to obtain a predicted failure probability sequence. This includes: identifying each first risk segment in the entire time domain of the balanced trajectory segment stream based on a preset detection time window; and for each first risk segment, performing the following weight calculation steps: within the detection time window, statistically analyzing the metric values ​​of each interference factor in the first risk segment within the detection time window, and determining the interruption of the balanced trajectory segment stream within the detection time window based on the ratio of the number of each interruption signal in the balanced trajectory segment stream within the sliding detection window to the duration of the detection time window. Signal index density; using Pearson correlation analysis, the correlation coefficients between each metric and the interrupt signal index density are calculated, and each correlation coefficient is subjected to min-max normalization mapping to obtain the influence weight sequence; based on the influence weight sequence, the metric values ​​of each interference factor in the first risk segment are weighted and summed to obtain the segment fault indication value of the first risk segment; using the interrupt signal index density as prior information, the segment fault indication value of each first risk segment is derived using the empirical Bayes update method to obtain the fault probability value of each segment; the fault occurrence probability values ​​are concatenated according to the timestamp order to obtain the predicted fault probability sequence.

[0052] For example, a detection time window length is set, such as 10 seconds or 1 minute, and the entire balanced trajectory segment stream is traversed. For each window, abnormal indicators or interference factors within that window are calculated (such as the number of GPS signal loss times, abnormal IMU acceleration values, visual signal frame loss rate, etc.). If the abnormal indicators within the window exceed a preset threshold (exemplary thresholds: more than 3 GPS loss times, IMU acceleration deviation greater than 0.5 m / s²), then the trajectory segment corresponding to that time period is marked as the first risk segment.

[0053] For example, if in a certain balanced trajectory segment stream, the GPS signal is lost three times between the 30th and 40th seconds, and the IMU acceleration deviation is 0.6 m / s², then the trajectory segment corresponding to that time period is identified as the first risk segment.

[0054] For example, within each first risk segment, the metrics of various interference factors (such as GPS loss, IMU anomaly, and visual frame loss) are counted, and the number of interruption signals within the sliding detection window is also counted. This count is divided by the detection window duration to obtain the interruption signal index density. This density reflects the degree to which the trajectory continuity is disrupted.

[0055] For example, if the detection window length is 10 seconds, and the GPS is lost 3 times and the IMU malfunctions 2 times within the first risk segment, that is, the total number of interrupt signals is 5, then the interrupt signal index density is 0.5.

[0056] For example, Pearson correlation analysis is used to calculate the correlation between various interference factor metrics (such as GPS signal-to-noise ratio, heart rate standard deviation, etc.) and the interruption signal index density (0.05 times / second). The purpose of this step is to explore which interference factors are more strongly associated with signal interruption. For instance, the correlation coefficient between GPS signal-to-noise ratio and interruption density is calculated to be -0.8, and the correlation coefficient between heart rate standard deviation and interruption density is 0.6. Since the correlation coefficients range from [-1, 1], a minimum-maximum normalization mapping is needed to linearly scale them to the interval in order to unify them into weights representing importance. Assuming that the maximum value among all correlation coefficients is 0.7 and the minimum value is -0.9, then for the correlation coefficient of -0.8, its normalized influence weight is [(-0.8) - (-0.9)] / [0.7 - (-0.9)] ≈ 0.0625. After this processing, each interference factor corresponds to an influence weight, collectively forming an influence weight sequence.

[0057] For example, the segment fault indication value is obtained by weighting and summing the metric values ​​of each interference factor within the first risk segment with their corresponding weights. For instance, the result of multiplying the number of GPS losses (3) by a weight of 1.0 is 3; the result of multiplying the number of IMU anomalies (2) by a weight of 0.75 is 1.5; therefore, the segment fault indication value is the sum of 3 and 1.5, which is 4.5.

[0058] For example, the interruption signal index density provides an initial judgment about the likelihood of a failure (prior probability), while the segment failure indication value is a value calculated based on the specific circumstances of the current segment (likelihood). The Empirical Bayesian method combines these two, utilizing statistical regularities learned from all historical first-risk segment data to refine and update the prior probability, thereby obtaining a more accurate posterior probability. This posterior probability is the final segment failure probability value. It represents the probability that the segment will indeed fail, considering the generality of signal interruptions and the specific manifestations of various current interference factors.

[0059] For example, after calculating the corresponding segment failure probability value for all identified first-risk segments according to the obstacles given in the previous example, these probability values ​​are sorted according to their respective timestamps in the equilibrium trajectory segment stream. For example, the segment failure probability at 10:01 is 0.65, at 10:03 it is 0.72, and at 10:05 it is 0.68. By concatenating these values ​​in chronological order [(10:01, 0.65), (10:03, 0.72), (10:05, 0.68), ...], a time-varying predicted failure probability sequence is formed.

[0060] According to the above implementation method, this step identifies the first risk segment in the balanced trajectory segment flow within a preset time window, counts the measurement values ​​of each interference factor and calculates the influence weight by combining the interruption signal index density, and then obtains the segment failure probability based on weighted summation empirical Bayes update, and finally generates a time-series-based predicted failure probability sequence, thereby realizing the quantitative assessment of potential failure risks in the operation trajectory, providing a reliable basis for risk visualization, trajectory optimization and safety management, and significantly improving the accuracy of trajectory analysis.

[0061] In one implementation, a backpacker's work trajectory report is generated based on an equilibrium trajectory segment stream and a predicted failure probability sequence. The work trajectory report includes a path graph and a risk point list. The process includes: extracting a sequence of spatial coordinate points for the backpacker at each timestamp from the equilibrium trajectory segment stream, and drawing a path graph based on the connectivity of these spatial coordinate points; marking time points in the predicted failure probability sequence where the failure probability value is greater than a preset risk threshold as risk points; aligning the risk points with the spatial coordinates in the equilibrium trajectory segment stream in time to determine the spatial location of each risk point in the path graph; generating a risk point list based on the spatial location of each risk point in the path graph and the failure probability value of the risk points; and determining the work trajectory report based on the path graph and the risk point list.

[0062] For example, all data points in the balanced trajectory segment stream are traversed, each containing a precise timestamp and corresponding spatial coordinate information (e.g., longitude and latitude). These spatial coordinate points are extracted in chronological order to form an ordered sequence of coordinate points. Subsequently, this sequence of coordinate points is visualized in a two-dimensional or three-dimensional coordinate system using a graphics drawing tool, and adjacent coordinate points are connected by straight lines or smooth curves to generate a path curve that visually displays the backpacker's complete movement trajectory from the starting point to the finish line.

[0063] For example, the predicted failure probability sequence is processed to identify risk points. Specifically, a risk threshold (e.g., 0.7) is set, and the failure probability value corresponding to each time point in the predicted failure probability sequence is judged sequentially. If the failure probability value at a certain time point is greater than the risk threshold, then that time point is marked as a risk point.

[0064] For example, if the predicted failure probability sequence is [(10:01, 0.65), (10:03, 0.72), (10:05, 0.68)], and the threshold is 0.7, then time point 10:03 is identified as a risk point. Through this process, the specific time points at which all potential risks occur can be screened out.

[0065] For example, to map risk points to specific spatial locations, a time alignment operation is used to map risk points in the predicted failure probability sequence to timestamps in the balanced trajectory segment stream. Since both share a unified time base, the spatial coordinates corresponding to the risk points can be obtained by finding data points whose timestamps perfectly match those of the risk points.

[0066] For example, the spatial coordinates corresponding to the timestamp 10:03 are (121.45°E, 31.22°N), where 121.45°E represents 121.45 degrees east longitude and 31.22°N represents 31.22 degrees north latitude. Based on this, the precise location of the risk point on the path curve can be determined.

[0067] For example, the risk point list serves as a structured data set used to systematically record the temporal, spatial, and quantitative information of various potential failure risk events along a backpacker's work trajectory. Specifically, the risk point list can be presented in tabular form, with each row representing a risk point and including information such as "risk point number," "occurrence time," "longitude," "latitude," and "failure probability value."

[0068] For example, risk point number 1 is recorded as follows: occurrence time is 10:03, longitude is 121.45°E, latitude is 31.22°N, and failure probability value is 0.72. This list provides users with detailed quantitative information on risk events.

[0069] For example, the path curve diagram is integrated and summarized with a structured list of risk points to form a backpacker's work trajectory report. In the report, the path curve diagram serves as the core display, accurately marking the locations of risk points, and is accompanied by a complete list of risk points. This allows users to intuitively understand the spatial distribution of risk points and obtain temporal and quantitative risk information for each risk point. In this way, the work trajectory is visualized and risk is quantitatively analyzed.

[0070] According to the above implementation method, by extracting the spatial coordinate sequence of backpackers at various timestamps from the balanced trajectory segment stream and plotting the path curve, and simultaneously identifying high-risk time points based on the predicted failure probability sequence and aligning them with the spatial coordinates to generate a risk point list, a work trajectory report is finally formed. In this way, the work trajectory is visualized and the potential failure risks are quantitatively analyzed.

[0071] Figure 2 This is a structural block diagram of a backpacker's work trajectory automatic generation device according to an embodiment of the present invention.

[0072] like Figure 2 As shown, the backpacker's work trajectory automatic generation device may include: Error compensation module 510 is used to perform non-line-of-sight error compensation on the backpacker's original multimodal observation dataset to obtain a compensated trajectory segment dataset; The trajectory construction module 520 is used to construct the backpacker's gait trajectory based on the compensated trajectory fragment dataset to obtain various structured trajectory fragments. The trajectory segment integration module 530 is used to integrate the various structured trajectory segments to obtain a balanced trajectory segment flow; The fault probability calculation module 540 is used to calculate the fault probability of each first risk segment in the balanced trajectory segment stream to obtain a predicted fault probability sequence. The generation module 550 is used to generate a work trajectory report for the backpacker based on the balanced trajectory segment flow and the predicted fault probability sequence, wherein the work trajectory report includes a path curve and a list of risk points.

[0073] In one embodiment, the error compensation module includes: An abnormal signal identification unit is used to identify anomalies in each modal signal in the original multimodal observation dataset and obtain the abnormal time periods of each modal signal. The extraction unit is used to extract the abnormal coordinate points corresponding to the abnormal time periods from the original multimodal observation dataset based on the abnormal time periods of each of the modal signals, so as to obtain each abnormal coordinate point; The trajectory extension prediction unit is used to perform trajectory extension prediction on each abnormal coordinate point through a state estimation algorithm to obtain each compensation coordinate point. The update unit is used to replace each of the abnormal coordinate points based on each of the compensation coordinate points in order to update the original multimodal observation dataset and obtain the compensation trajectory segment dataset.

[0074] In one implementation, the trajectory construction module includes: The dataset acquisition unit is used to acquire the compensation trajectory segment dataset, wherein the compensation trajectory segment dataset includes a compensation coordinate sequence, an original coordinate sequence, a modality type, and timestamp information; An extended processing unit is used to extend the original coordinate sequence according to the time axis to obtain a time-continuous observation sequence; The differential calculation unit is used to perform differential calculations on the compensated coordinate sequence and the time-continuous observation sequence to obtain the displacement vector between each adjacent time reference. The joint index building unit is used to perform coordinate transformation and alignment on each of the displacement vectors and each current observation value of the multi-source sensor to obtain the alignment result, and to build a joint index based on the alignment result, the mode type and the timestamp information to obtain a time-series fusion index set; The fusion connection unit is used to fuse and connect the compensation coordinate sequence and the trajectory data corresponding to the gait trajectory based on the correspondence of each index in the temporal fusion index set, so as to obtain a coherent path coordinate point sequence. The segmentation processing unit is used to segment the path coordinate point sequence to obtain various short-time trajectory segments; The encapsulation unit is used to encapsulate each of the short-time trajectory segments to obtain each of the structured trajectory segments.

[0075] In one embodiment, the trajectory segment integration module includes: An ascending sorting unit is used to sort each structured trajectory segment in ascending order according to the start timestamp of each structured trajectory segment to obtain a first trajectory segment sequence. A connection unit is used to connect two adjacent first trajectory segments in the first trajectory segment sequence when two adjacent first trajectory segments are not continuous, to obtain a second trajectory segment sequence. The smooth interpolation processing unit is used to perform smooth interpolation processing on the splicing points of each second trajectory segment in the second trajectory segment sequence to obtain the balanced trajectory segment stream.

[0076] In one embodiment, the connection unit includes: The trajectory segment acquisition subunit is used to acquire the timestamps of two adjacent first trajectory segments, wherein the timestamps of the two adjacent first trajectory segments include the end timestamp of the third trajectory segment and the start timestamp of the fourth trajectory segment; An insertion subunit is used to insert data points into the gap between the third trajectory segment and the fourth trajectory segment if the start timestamp of the fourth trajectory segment is greater than the end timestamp of the third trajectory segment, so as to connect the third trajectory segment and the fourth trajectory segment. The weighted fusion subunit is used to perform weighted fusion on the coordinate point data of the third trajectory segment and the fourth trajectory segment in the time overlap interval if the start timestamp of the fourth trajectory segment is less than the end timestamp of the third trajectory segment, so as to connect the third trajectory segment and the fourth trajectory segment.

[0077] In one embodiment, the fault probability calculation module includes: The first risk segment identification unit is used to identify the entire time domain of the balanced trajectory segment flow based on a preset detection time window, and obtain each first risk segment. The weight calculation unit performs the following weight calculation steps for each of the first risk segments: The statistical subunit is used to count the measurement values ​​of each interference factor in the first risk segment within the detection time window, and to determine the interruption signal index density of the balanced trajectory segment flow within the detection time window based on the ratio of the number of each interruption signal in the balanced trajectory segment flow within the sliding detection window to the duration of the detection time window. The correlation coefficient calculation subunit is used to calculate the correlation coefficient between each of the metric values ​​and the index density of the interruption signal using the Pearson correlation analysis method, and to perform minimum-maximum normalization mapping on each correlation coefficient to obtain the influence weight sequence. The weighted summation subunit is used to perform weighted summation of the measurement values ​​of each interference factor in the first risk segment based on the influence weight sequence, so as to obtain the segment fault indication value of the first risk segment. The derivation unit is used to derive the segment fault indication value of each of the first risk segments using the interrupt signal index density as prior information and the empirical Bayes update method to obtain the fault probability value of each segment. A concatenation unit is used to concatenate the various failure occurrence probability values ​​in the order of timestamps to obtain the predicted failure probability sequence.

[0078] In one embodiment, the generation module includes: The extraction unit is used to extract the sequence of spatial coordinate points of the backpacker at each time stamp from the balanced trajectory segment stream, and draw a path curve based on the connection relationship of each spatial coordinate point; A marking unit is used to mark time points in the predicted fault probability sequence where the fault probability value is greater than a preset risk threshold as risk points, so as to obtain each of the risk points; A time alignment unit is used to time align the risk points with the spatial coordinates in the balanced trajectory segment flow to determine the spatial position of each risk point in the path curve diagram. The risk point list generation unit is used to generate a risk point list based on the spatial location of each risk point in the path curve diagram and the failure probability value of the risk point. The operation trajectory report determination unit is used to determine the operation trajectory report based on the path curve and the risk point list.

[0079] The specific functions and examples of each module and submodule of the system in this embodiment of the invention can be found in the relevant descriptions of the corresponding steps in the above method embodiments, and will not be repeated here.

[0080] The acquisition, storage, and application of user personal information involved in the technical solution of this invention all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0081] This invention also provides an automatic backpacker work trajectory generation system, comprising: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method described in any one of the embodiments of the present invention.

[0082] The beneficial effects of the backpacker work trajectory automatic generation system in this embodiment of the invention are equivalent to the beneficial effects of the backpacker work trajectory automatic generation method described above, and will not be repeated here.

[0083] This invention also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the method described in any one of the embodiments of this invention.

[0084] The beneficial effects of the storage medium of the present invention are equivalent to the beneficial effects of the above-described method for automatically generating backpacker work trajectories, and will not be repeated here.

[0085] Figure 3 A schematic block diagram of an example electronic device 800 that can be used to implement embodiments of the present invention is shown. Electronic device 800 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic device 800 may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0086] like Figure 3As shown, the electronic device 800 includes a computing unit 801, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. The RAM 803 may also store various programs and data required for the operation of the electronic device 800. The computing unit 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0087] Multiple components in electronic device 800 are connected to I / O interface 805, including: input unit 806, such as keyboard, mouse, etc.; output unit 807, such as various types of displays, speakers, etc.; storage unit 808, such as disk, optical disk, etc.; and communication unit 809, such as network card, modem, wireless transceiver, etc. Communication unit 809 allows electronic device 800 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0088] The computing unit 801 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as the method for automatically generating backpacker work tracks. For example, in some embodiments, the method for automatically generating backpacker work tracks can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as storage unit 808. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of the method for automatically generating backpacker work tracks described above can be performed. Alternatively, in other embodiments, the computing unit 801 may be configured, by any other suitable means (e.g., by means of firmware), to perform a method for automatically generating backpacker work tracks.

[0089] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0090] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0091] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0092] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0093] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0094] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0095] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0096] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for automatically generating backpacker work trajectories, characterized in that, include: Non-line-of-sight error compensation was performed on the original multimodal observation dataset of backpackers to obtain a compensated trajectory segment dataset; Based on the compensated trajectory fragment dataset, the backpacker's gait trajectory is constructed to obtain various structured trajectory fragments; The structured trajectory segments are integrated to obtain a balanced trajectory segment flow; The failure probability is calculated for each first risk segment in the balanced trajectory segment stream to obtain a predicted failure probability sequence. Based on the balanced trajectory segment stream and the predicted failure probability sequence, a work trajectory report for the backpacker is generated, wherein the work trajectory report includes a path curve and a list of risk points.

2. The method according to claim 1, characterized in that, The process of performing non-line-of-sight error compensation on the original multimodal observation dataset of backpackers yields a compensated trajectory segment dataset, including: Anomaly identification is performed on each modal signal in the original multimodal observation dataset to obtain the abnormal time periods of each modal signal; Based on the abnormal time periods of each modal signal, the abnormal coordinate points corresponding to the abnormal time periods are extracted from the original multimodal observation dataset to obtain each abnormal coordinate point; The trajectory extension prediction of each abnormal coordinate point is performed by the state estimation algorithm to obtain each compensation coordinate point; Based on each of the compensation coordinate points, each of the abnormal coordinate points is replaced to update the original multimodal observation dataset, thereby obtaining a compensation trajectory segment dataset.

3. The method according to claim 1, characterized in that, Based on the compensated trajectory fragment dataset, the backpacker's gait trajectory is constructed to obtain various structured trajectory fragments, including: Obtain the compensation trajectory segment dataset, wherein the compensation trajectory segment dataset includes a compensation coordinate sequence, an original coordinate sequence, a modality type, and timestamp information; The original coordinate sequence is expanded based on the time axis to obtain a time-continuous observation sequence. The displacement vector between each adjacent time reference is obtained by performing differential calculation on the compensated coordinate sequence and the time continuous observation sequence. Each displacement vector is aligned with each current observation value of the multi-source sensor by coordinate transformation to obtain an alignment result. A joint index is established based on the alignment result, the mode type, and the timestamp information to obtain a time-series fusion index set. Based on the correspondence of each index in the temporal fusion index set, the compensation coordinate sequence and the trajectory data corresponding to the gait trajectory are fused and connected to obtain a coherent path coordinate point sequence. The path coordinate point sequence is segmented to obtain various short-time trajectory segments; Each of the short-time trajectory segments is encapsulated to obtain each of the structured trajectory segments.

4. The method according to claim 1, characterized in that, The process of integrating the various structured trajectory segments to obtain a balanced trajectory segment flow includes: Based on the start timestamp of each structured trajectory segment, the structured trajectory segments are sorted in ascending order to obtain the first trajectory segment sequence; If two adjacent first trajectory segments are discontinuous in the first trajectory segment sequence, the two adjacent first trajectory segments are connected to obtain the second trajectory segment sequence. Smooth interpolation is performed at the splicing points of each second trajectory segment in the second trajectory segment sequence to obtain the balanced trajectory segment stream.

5. The method according to claim 4, characterized in that, In the case where two adjacent first trajectory segments in the first trajectory segment sequence are discontinuous, connecting the two adjacent first trajectory segments to obtain the second trajectory segment sequence includes: Obtain the timestamps of two adjacent first trajectory segments, wherein the timestamps of the two adjacent first trajectory segments include the end timestamp of the third trajectory segment and the start timestamp of the fourth trajectory segment; If the start timestamp of the fourth trajectory segment is greater than the end timestamp of the third trajectory segment, then data points are inserted into the gap between the third trajectory segment and the fourth trajectory segment to connect the third trajectory segment and the fourth trajectory segment; If the start timestamp of the fourth trajectory segment is less than the end timestamp of the third trajectory segment, then the coordinate data of each coordinate point in the time overlap interval of the third trajectory segment and the fourth trajectory segment are weighted and fused to connect the third trajectory segment and the fourth trajectory segment.

6. The method according to claim 1, characterized in that, The step of calculating the failure probability of each first risk segment in the balanced trajectory segment stream to obtain a predicted failure probability sequence includes: Based on a preset detection time window, the entire time domain of the balanced trajectory segment flow is identified to obtain each of the first risk segments; For each of the first risk segments, the following weight calculation steps are performed: Within the detection time window, the measurement values ​​of each interference factor in the first risk segment within the detection time window are statistically analyzed, and the index density of the interruption signal in the balanced trajectory segment stream within the detection time window is determined based on the ratio of the number of each interruption signal in the balanced trajectory segment stream within the sliding detection window to the duration of the detection time window. The correlation coefficients between each metric and the index density of the interruption signal are calculated using Pearson correlation analysis, and the correlation coefficients are subjected to min-max normalization mapping to obtain the influence weight sequence. Based on the influence weight sequence, the measurement values ​​of each interference factor in the first risk segment are weighted and summed to obtain the segment fault indication value of the first risk segment. Using the interrupt signal index density as prior information, the fault indication value of each first risk segment is derived by using the empirical Bayes update method to obtain the fault probability value of each segment. The predicted fault probability sequence is obtained by concatenating the various fault occurrence probability values ​​in the order of their timestamps.

7. The method according to claim 1, characterized in that, The backpacker's work trajectory report is generated based on the balanced trajectory segment stream and the predicted failure probability sequence. The work trajectory report includes a path graph and a list of risk points, including: From the balanced trajectory segment stream, extract the sequence of spatial coordinate points of the backpacker at each timestamp, and draw a path curve based on the connection relationship of each spatial coordinate point; The time points in the predicted fault probability sequence where the fault probability value is greater than a preset risk threshold are marked as risk points to obtain each of the risk points; The risk points are time-aligned with the spatial coordinates in the balanced trajectory segment flow to determine the spatial location of each risk point in the path curve diagram. A list of risk points is generated based on the spatial location of each risk point in the path curve diagram and the failure probability value of each risk point. Based on the path curve and the risk point list, the operation trajectory report is determined.

8. A device for automatically generating backpacker work routes, characterized in that, include: The error compensation module is used to perform non-line-of-sight error compensation on the backpacker's original multimodal observation dataset to obtain a compensated trajectory segment dataset. The trajectory construction module is used to construct the backpacker's gait trajectory based on the compensated trajectory fragment dataset to obtain various structured trajectory fragments; The trajectory segment integration module is used to integrate the various structured trajectory segments to obtain a balanced trajectory segment flow; The fault probability calculation module is used to calculate the fault probability of each first risk segment in the balanced trajectory segment stream to obtain a predicted fault probability sequence. The generation module is used to generate a work trajectory report for the backpacker based on the balanced trajectory segment flow and the predicted failure probability sequence, wherein the work trajectory report includes a path curve and a list of risk points.

9. A system for automatically generating backpacker work routes, characterized in that, include: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-7.

Images