Inertial navigation method for underground coal mine and inspection robot

Abstract

The invention relates to an inertial navigation method for an underground coal mine, which belongs to the technical field of inertial navigation, and comprises the following steps: initializing an inspection robot, and controlling the inspection robot to traverse on a road of a target coal mine to obtain acceleration parameters of a time sequence; extracting the acceleration parameter of the time sequence to obtain a plurality of pieces of feature data, and obtaining a feature position corresponding to each piece of feature data; the inspection robot is controlled to traverse each feature position N times, N sets of acceleration parameters of each feature position are obtained, and N is a positive integer; performing association based on the N groups of acceleration parameters of each feature position; if the association succeeds, obtaining a bumping feature of the corresponding feature position; if the association fails, deleting the corresponding feature position; constructing a self-correction topology network based on the feature position, wherein the self-correction topology network is used for carrying out auxiliary verification on the bump feature; and carrying out self-correction on the inspection robot based on the self-correction topology network and the bumping characteristics.

Description

Technical Field

[0001] The present invention belongs to the technical field of inertial navigation, and relates to an inertial navigation method and an inspection robot for underground coal mines. Background Art

[0002] Inertial navigation in underground coal mines mainly relies on built-in inertial measurement units to achieve positioning and navigation in the underground coal mine environment without satellite signals. Its principle is as follows: Based on Newton's laws of motion, the acceleration and angular velocity measured by the inertial measurement unit are integrated to calculate the position, velocity, and direction. And a reference system is constructed based on the initial position, so as to achieve an approximate positioning and navigation with the initial position as the reference system. However, due to the existence of tiny errors in the integral calculation of inertial sensors, the errors will accumulate after long-term use, resulting in the accuracy being less than the preset value. Therefore, it is necessary to intermittently correct the inertial sensors. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide an inertial navigation method and an inspection robot for underground coal mines to solve the technical problems raised in the background art.

[0004] To achieve the above purpose, the present invention provides the following technical solutions:

[0005] On the one hand, the present invention provides an inertial navigation method for underground coal mines, including the following steps:

[0006] Initialize the inspection robot and control the inspection robot to traverse the roads in the target coal mine shaft to obtain sequential acceleration parameters;

[0007] Extract the sequential acceleration parameters to obtain a number of characteristic data, and obtain the characteristic positions corresponding to each characteristic data;

[0008] Control the inspection robot to traverse each characteristic position N times to obtain N groups of acceleration parameters for each characteristic position, where N is a positive integer;

[0009] Correlate based on the N groups of acceleration parameters for each characteristic position: If the correlation is successful, the bump characteristic of the corresponding characteristic position is obtained; if the correlation fails, the corresponding characteristic position is deleted;

[0010] Construct a self-correcting topology network based on the characteristic positions, and the self-correcting topology network is used to assist in verifying the bump characteristics;

[0011] Self-correct the inspection robot based on the self-correcting topology network and the bump characteristics.

[0012] Further, the extraction of the sequential acceleration parameters to obtain a number of characteristic data includes:

[0013] Establish an acceleration parameter threshold and initialize the generation of a sliding time window;

[0014] Based on the sliding time window, slice and extract the acceleration parameters of the time series to obtain sliced parameters, and compare the acceleration parameters at each moment in the sliced parameters with the acceleration parameter threshold. If there are no less than a preset number of acceleration parameters ≥ the acceleration parameter threshold, then use the corresponding sliced parameters as feature data and obtain the feature positions of the corresponding feature data;

[0015] Among them, the feature position includes the coordinates of each acceleration parameter within the corresponding sliding time window, and the coordinates of the acceleration parameter are calculated by the inertial measurement unit preset in the inspection robot.

[0016] Furthermore, the N - time traversal of each feature position includes:

[0017] For the i - th feature position, each selects times to perform forward traversal and reverse traversal on the i - th feature position to obtain the corresponding N groups of acceleration parameters; among them, both the forward traversal and the reverse traversal are manually defined; N is an even number; i is the feature position index and is a positive integer.

[0018] Furthermore, the association based on the N groups of acceleration parameters of each feature position includes:

[0019] For the i - th feature position, reverse the groups of acceleration parameters obtained by reverse traversal, including: obtaining the central moment of the j - th group of acceleration parameters obtained by reverse traversal, and based on the central moment, symmetrically exchange the moments of each acceleration parameter in the j - th group with respect to the central moment to obtain the reverse acceleration parameters of the j - th group of acceleration parameters obtained by reverse traversal, where j is a positive integer;

[0020] Combine and associate the groups of reverse acceleration parameters with the groups of acceleration parameters obtained by forward traversal. The combination and association include: comparing the acceleration parameters at each moment in the N groups of acceleration parameters with the acceleration parameter threshold, extracting the acceleration parameters greater than or equal to the acceleration parameter threshold in each group of acceleration parameters as matching parameters, and performing time association and space association on the matching parameters;

[0021] Initialize the generation of M spatio - temporal alignment points, allocate N matching parameters to each spatio - temporal alignment point, and each matching parameter comes from different groups of acceleration parameters respectively, and each matching parameter can only be allocated to one spatio - temporal alignment point, where M is a positive integer;

[0022] The time correlation includes: obtaining the maximum time and the minimum time corresponding to the N matching parameters at each spatio-temporal alignment point to calculate the time difference, and cumulatively summing the time differences of the M spatio-temporal alignment points to obtain the alignment cost of the time error;

[0023] The space correlation includes: obtaining the maximum acceleration parameter and the minimum acceleration parameter corresponding to the N matching parameters at each spatio-temporal alignment point to calculate the space difference, and cumulatively summing the space differences of the M spatio-temporal alignment points to obtain the alignment cost of the space error;

[0024] If the alignment cost of the time error is less than or equal to the corresponding time threshold, and the alignment cost of the space error is less than or equal to the corresponding space threshold, the correlation is successful, otherwise the correlation fails.

[0025] Furthermore, based on the genetic algorithm, the alignment costs of the time error and the space error are minimized to obtain the minimized alignment cost of the time error and the minimized alignment cost of the space error, and the correlation result is obtained based on the minimized alignment cost of the time error and the minimized alignment cost of the space error;

[0026] The correlation result includes: correlation success and correlation failure.

[0027] Furthermore, the process of obtaining the bump feature includes:

[0028] If the i-th feature position is successfully correlated, the matching parameters of the i-th feature position with respect to each spatio-temporal alignment point among the M spatio-temporal alignment points are obtained, and the spatial mean and the time mean of the matching parameters of each spatio-temporal alignment point are calculated;

[0029] The bump feature is encoded and represented by a matrix of size 2×M as: where S1…SM represent the spatial means of the 1st to M-th spatio-temporal alignment points, T1…TM represent the time means of the 1st to M-th spatio-temporal alignment points, and T1…TM are sorted in chronological order, and S1…SM are sorted correspondingly according to T1…TM.

[0030] Furthermore, the construction of the self-correcting topological network based on the feature positions includes:

[0031] Mapping each feature position to a topological node;

[0032] If the i-th feature position and the k-th feature position are adjacent, a topological edge is constructed between the i-th topological node and the k-th topological node to obtain the self-correcting topological network; where, i≠k, adjacent means that there is no g-th feature position on the shortest path between the i-th feature position and the k-th feature position, i≠k≠g, and k, g are positive integers.

[0033] Further, the self - calibration of the inspection robot based on the self - correcting topology network and the bump characteristics includes:

[0034] Step A1: During the target time period, based on the inspection robot, obtain the time - series acceleration parameters, and extract the first target shard parameters from the time - series acceleration parameters based on a sliding time window; reverse the first target shard parameters to obtain the second target shard parameters;

[0035] Based on the acceleration parameter threshold, extract the first target matching parameters and the second target matching parameters from the first target shard parameters and the second target shard parameters respectively, and construct the corresponding first target bump characteristics and second target bump characteristics based on the first target matching parameters and the second target matching parameters; calculate the similarity between the first target bump characteristics and the second target bump characteristics and the bump characteristics at the k - th feature position. If the similarity of any one of the first target bump characteristics and the second target bump characteristics and the bump characteristics at the k - th feature position is less than or equal to the preset similarity threshold, then obtain a similarity verification;

[0036] Step A2: Obtain the feature positions traversed by the inspection robot historically during the target time period, and extract the feature position of the last traversal; if there is a topological edge constructed between the feature position of the last traversal and the topological node corresponding to the k - th feature position, then obtain an auxiliary verification;

[0037] Step A3: If both the similarity verification and the auxiliary verification are successful, then obtain the estimated coordinates of the inertial measurement unit of the inspection robot, and obtain the starting coordinates and ending coordinates of the k - th feature position;

[0038] If the similarity between the first target bump characteristics and the bump characteristics at the k - th feature position is less than or equal to the preset similarity threshold, then correct the estimated coordinates to the ending coordinates;

[0039] If the similarity between the second target bump characteristics and the bump characteristics at the k - th feature position is less than or equal to the preset similarity threshold, then correct the estimated coordinates to the starting coordinates;

[0040] Among them, for the first feature position within the target time period, the inspection robot only needs to pass the similarity verification.

[0041] Further, the similarity of the bump characteristics is calculated based on the Euclidean distance.

[0042] On the other hand, the present invention provides an inertial navigation inspection robot for coal mine underground, including: a controller, a memory, a computer program stored on the memory and executable by the controller, and a data bus for realizing the connection and communication between the controller and the memory. When the computer program is executed by the controller, it implements any one of the inertial navigation methods for coal mine underground described above.

[0043] The beneficial effects of the present invention are as follows: By utilizing the inherent unevenness of the underground roads in coal mines to capture and extract acceleration features, combined with multiple traversal acquisitions, spatio-temporal matching, genetic algorithm optimization, and self-correcting topology network construction, the correction of the cumulative inertial navigation integration error is achieved. Specifically, during the operation of the inspection robot, the system automatically identifies fixed feature points and uses these feature points as reference coordinates. Once a deviation between the actually measured coordinates and the reference coordinates is detected, error correction is immediately performed, thereby greatly improving the positioning accuracy and navigation stability, saving costs and labor input, and at the same time enhancing the robustness and adaptive ability of the system in an environment without satellite signals.

[0044] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be learned from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be described in detail preferably in conjunction with the drawings, where:

[0046] Figure 1 is a flowchart of an inertial navigation method for underground coal mines of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] The following specific examples illustrate the embodiments of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention schematically. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0048] It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention schematically. Therefore, only the components related to the present invention are shown in the diagrams, rather than being drawn according to the number, shape, and size of the components during actual implementation. The type, quantity, and ratio of each component during actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0049] In the following description, a large number of details are explored to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0050] As Figure 1 shown, an inertial navigation method for underground coal mines includes:

[0051] Step 1, initialize the inspection robot and control the inspection robot to traverse the roads in the target coal mine to obtain time-series acceleration parameters;

[0052] Step 2, extract a number of feature data from the time-series acceleration parameters and obtain the feature positions corresponding to each feature data;

[0053] Step 3, control the inspection robot to traverse each feature position N times to obtain N groups of acceleration parameters for each feature position, where N is a positive integer;

[0054] Step 4, correlate based on the N groups of acceleration parameters for each feature position:

[0055] If the correlation is successful, the bump feature of the corresponding feature position is obtained;

[0056] If the correlation fails, the corresponding feature position is deleted;

[0057] Step 5, construct a self-correcting topology network based on the feature positions, and the self-correcting topology network is used to assist in verifying the bump feature;

[0058] Step 6, perform self-calibration on the inspection robot based on the self-correcting topology network and the bump feature.

[0059] It should be noted that when the inspection robot runs for the first time, due to the small error, the coordinate data collected during its traversal of the target coal mine roadway is relatively accurate and can be used as a reference basis for subsequent coordinate correction. There are many narrow and unstable roadways in the coal mine. When the robot passes through these roadways, obvious acceleration changes will occur. Therefore, during the initial traversal process, such roadways can be screened out as characteristic positions. However, for some wider roadways, the acceleration changes when the robot passes through may be unstable - sometimes obvious and sometimes not significant. For this reason, the system uses N forward and reverse traversals to verify each candidate characteristic position multiple times, and screens out those roadways that can stably trigger the same acceleration change as the true characteristic positions. Subsequently, an auxiliary verification is performed on these characteristic positions by constructing a topological network. During subsequent operations, once it is detected that the acceleration parameter indicates that a certain characteristic position has been reached, the current estimated coordinate is replaced with the accurate coordinate of that characteristic position, thereby correcting the cumulative error.

[0060] In an embodiment of the present invention, a number of characteristic data are extracted from the acceleration parameters of the time series, including:

[0061] An acceleration parameter threshold is established, and a sliding time window is initialized and generated;

[0062] Based on the sliding time window, the acceleration parameters of the time series are extracted in slices to obtain slice parameters, and the acceleration parameters at each moment in the slice parameters are compared with the acceleration parameter threshold. If there are no less than a preset number of acceleration parameters ≥ the acceleration parameter threshold, the corresponding slice parameters are used as characteristic data, and the characteristic positions of the corresponding characteristic data are obtained;

[0063] Among them, the characteristic position includes the coordinates of each acceleration parameter within the corresponding sliding time window, and the coordinates of the acceleration parameter are calculated by the inertial measurement unit preset in the inspection robot.

[0064] Specifically, an acceleration threshold is set according to the actual situation, and this threshold is used to determine whether the acceleration at a certain moment is significant enough. The continuous time-series acceleration data is segmented using a sliding time window. The sliding time window is a time slice with a fixed duration, and each window contains a segment of continuous acceleration data. The acceleration data of the entire time series is sliced in an overlapping manner according to the sliding window. The step size of the sliding time window is set manually, and the acceleration data segment within each time window (i.e., the slicing parameter) is obtained. For each data segment, the acceleration parameters at each moment are compared with the preset threshold. If there are no less than the preset number of acceleration data reaching or exceeding the threshold in this segment, it is considered that there is a significant acceleration change within this segment, and this segment is marked as "feature data". For each sliding window determined to be feature data, the corresponding coordinate information at each moment is further extracted. These coordinate information are calculated by the inertial measurement unit (IMU) preset in the inspection robot. These coordinates constitute the "feature position" of the feature data, that is, the spatial position describing the acceleration anomaly during this period of time.

[0065] In an embodiment of the present invention, each feature position is traversed N times, including:

[0066] For the i-th feature position, each selects times to perform forward traversal and reverse traversal on the i-th feature position to obtain the corresponding N groups of acceleration parameters; among them, both the forward traversal and the reverse traversal are defined manually; N is an even number; i is the feature position index and is a positive integer.

[0067] Specifically, for each feature position (such as the i-th feature position), it is required that the inspection robot traverses this position N times. This means that when the robot passes through this feature position, it does not pass through only once, but repeats multiple times to obtain multiple independent groups of acceleration data. Each traversal includes two ways: forward driving and reverse driving. These two traversal methods are both predefined manually to ensure that consistent acceleration characteristics can be captured in different driving directions, thereby enhancing the reliability and symmetry of the data. By obtaining N groups of acceleration data, subsequent spatio-temporal correlation and matching can be performed on these data to verify whether the acceleration change at this feature position is consistent. Only when the data of multiple traversals all meet the preset matching criteria, this feature position is considered stable and reliable for subsequent positioning correction.

[0068] In an embodiment of the present invention, the N groups of acceleration parameters based on each feature position are correlated, including:

[0069] For the i-th feature position, reverse the group of acceleration parameters obtained by reverse traversal, including:

[0070] Obtain the central moment of the acceleration parameters traversed in reverse for the j-th group. Based on the central moment, symmetrically swap the moments of each acceleration parameter in the j-th group with respect to the central moment to obtain the reverse acceleration parameters of the acceleration parameters traversed in reverse for the j-th group, where j is a positive integer;

[0071] Combine and correlate the group of reverse acceleration parameters with the acceleration parameters traversed in the forward direction. The combination and correlation include: comparing the acceleration parameters at each moment of the N groups of acceleration parameters with the acceleration parameter threshold, extracting the acceleration parameters greater than or equal to the acceleration parameter threshold in each group of acceleration parameters as matching parameters, and performing time correlation and spatial correlation on the matching parameters; Initialize and generate M spatio-temporal alignment points, and assign N matching parameters to each spatio-temporal alignment point. Moreover, each matching parameter comes from the acceleration parameters of different groups respectively, and each matching parameter can only be assigned to one spatio-temporal alignment point, where M is a positive integer;

[0072] Time correlation includes: obtaining the maximum moment and the minimum moment corresponding to the N matching parameters of each spatio-temporal alignment point to calculate the time difference, and cumulatively summing the time differences of the M spatio-temporal alignment points to obtain the alignment cost of the time error;

[0073] Spatial correlation includes: obtaining the maximum acceleration parameter and the minimum acceleration parameter corresponding to the N matching parameters of each spatio-temporal alignment point to calculate the spatial difference, and cumulatively summing the spatial differences of the M spatio-temporal alignment points to obtain the alignment cost of the spatial error;

[0074] If the alignment cost of the time error is less than or equal to the corresponding time threshold, and the alignment cost of the spatial error is less than or equal to the corresponding spatial threshold, then the correlation is successful; otherwise, the correlation fails.

[0075]

[0076] ​Specifically, for the i-th feature position, forward and reverse acceleration data are obtained from multiple traversals. To align the forward and reverse data, first, the acceleration parameters of the reverse traversal are reversed. The specific method is as follows: for the j-th group of data in the reverse traversal, first determine the central time of this group of data, and then exchange the times of each acceleration data in this group with the central time as the axis of symmetry to obtain the "reverse acceleration parameter". This step ensures that the forward and reverse data can match in time, facilitating subsequent correlation comparison. Combine the reverse acceleration parameter after time reversal with the forward traversal data. During the combination process, compare the parameters of each time in each group of acceleration data with a preset acceleration threshold, and extract the acceleration data greater than or equal to the threshold as the "matching parameter". The purpose of this step is to screen out those data points that show significant acceleration changes in each traversal. Initialize M spatio-temporal alignment points, and assign matching parameters from different traversal groups (a total of N) to each alignment point, ensuring that each alignment point contains data from each independent traversal. For each alignment point, take the maximum time and the minimum time among its N matching parameters, calculate the time difference, and then sum up the time differences of all M alignment points to obtain a total time error cost. Similarly, for each alignment point, take the maximum and minimum acceleration parameters among its N matching parameters, calculate the space difference, and sum up the space differences to obtain the space error cost. Finally, determine whether the accumulated time error and space error are both less than or equal to the preset time and space thresholds. If both conditions are met, it is considered that the association of this feature position is successful, indicating that this position shows stable and consistent acceleration characteristics in multiple traversals; otherwise, the association fails, and this feature position may be unreliable and will be excluded.

[0077] In one embodiment of the present invention, the alignment cost of the time error and the alignment cost of the space error are minimized based on a genetic algorithm to obtain the minimized alignment cost of the time error and the minimized alignment cost of the space error, and the association result is obtained based on the minimized alignment cost of the time error and the minimized alignment cost of the space error;

[0078] The association result includes: association success and association failure.

[0079] Specifically, by traversing the collected matching parameters multiple times, the time error and space error of each spatio-temporal alignment point are calculated, and the overall error cost is accumulated. The genetic algorithm is used to minimize and optimize these time error costs and space error costs. By simulating the process of natural selection and evolution, the genetic algorithm finds the best (i.e., the smallest) combination of error costs, so that the alignment error reaches the lowest level. The minimized spatio-temporal alignment errors obtained by the genetic algorithm can more accurately reflect the matching degree of the data at each feature position in different traversals. The lowest error cost indicates that the corresponding matching parameters are consistent and stable in space and time, reflecting the repeatability and reliability of the feature position in different traversals. Based on the alignment cost of the minimized time error and space error, the system determines the data association result of the feature position. If the optimized error values all meet the preset error threshold, it is determined that the association of the feature position is "successful association"; otherwise, it is considered that the data matching is not ideal and is determined to be "association failure".

[0080] In an embodiment of the present invention, the process of obtaining the jolt feature includes:

[0081] If the i-th feature position is successfully associated, the matching parameters of the i-th feature position with respect to each of the M spatio-temporal alignment points are obtained, and the spatial mean value and time mean value of the matching parameters of each spatio-temporal alignment point are calculated;

[0082] The jolt feature is encoded and represented by a matrix of size 2×M as: Where S1…SM represent the spatial mean values of the 1st to M-th spatio-temporal alignment points, T1…TM represent the time mean values of the 1st to M-th spatio-temporal alignment points, and T1…TM are sorted in chronological order, and S1…SM are sorted correspondingly according to T1…TM.

[0083] Specifically, for each successfully associated feature position (the i-th feature position), the matching parameters of each alignment point are respectively obtained from the M spatio-temporal alignment points corresponding to this position. For each spatio-temporal alignment point, the spatial mean value and time mean value of these matching parameters are calculated. The spatial mean value reflects the average characteristics of the collected data at this alignment point in space, and the time mean value reflects the time distribution of the collected data. The above calculated mean values are arranged in chronological order to form a matrix of size 2×M. The first row of the matrix represents the spatial mean values of the 1st to M-th spatio-temporal alignment points, and the second row represents the corresponding time mean values. This matrix encoding method condenses the information originally scattered on multiple alignment points into a compact and structured feature representation, which is convenient for subsequent feature matching and comparison.

[0084] In an embodiment of the present invention, constructing a self-correcting topology network based on feature positions includes:

[0085] Map each feature position to a topological node;

[0086] If the i-th feature position and the k-th feature position are adjacent, construct a topological edge between the i-th topological node and the k-th topological node to obtain a self-correcting topological network; where i≠k, adjacent means that there is no g-th feature position on the shortest path between the i-th feature position and the k-th feature position, i≠k≠g, and both k and g are positive integers.

[0087] Specifically, map the feature positions to topological nodes in the network respectively. These nodes represent the key reference points with significant acceleration changes in the underground coal mine road. For any two different feature positions (such as the i-th and the k-th), if they are determined to be adjacent in space, establish a topological edge between the corresponding topological nodes. Here, "adjacent" is defined as: there is no other feature position on the shortest path between these two feature positions (that is, there is no g-th feature position, where i≠k≠g). This rule ensures that connections are only established between feature positions with direct spatial connections, avoiding redundancy and interference.

[0088] In an embodiment of the present invention, self-correction of the inspection robot is performed based on the self-correcting topological network and the bump feature, including:

[0089] Step 61, within the target time period, based on the inspection robot, obtain the time-series acceleration parameters, and extract the first target shard parameters from the time-series acceleration parameters based on a sliding time window; reverse the first target shard parameters to obtain the second target shard parameters;

[0090] Based on the acceleration parameter threshold, extract the first target matching parameters and the second target matching parameters from the first target shard parameters and the second target shard parameters respectively, and construct the corresponding first target bump feature and the second target bump feature based on the first target matching parameters and the second target matching parameters respectively; calculate the similarity between the first target bump feature and the second target bump feature and the bump feature of the k-th feature position. If the similarity of any one of the first target bump feature and the second target bump feature and the bump feature of the k-th feature position is less than or equal to the preset similarity threshold, then obtain a similarity verification.

[0091] Step 62, obtain the feature positions traversed by the inspection robot historically within the target time period, and extract the feature position of the last traversal; if there is a topological edge constructed between the feature position of the last traversal and the topological node corresponding to the k-th feature position, then obtain an auxiliary verification.

[0092] Step 63, if both the similarity verification and the auxiliary verification are successful, obtain the estimated coordinates of the inertial measurement unit of the inspection robot, and obtain the starting coordinates and the ending coordinates of the k-th feature position.

[0093] If the similarity between the first target bump feature and the bump feature at the k-th feature position is less than or equal to a preset similarity threshold, the estimated coordinate is corrected to the end coordinate;

[0094] If the similarity between the second target bump feature and the bump feature at the k-th feature position is less than or equal to a preset similarity threshold, the estimated coordinate is corrected to the start coordinate;

[0095] Among them, for the first feature position within the target time period, the inspection robot only needs to pass the similarity verification.

[0096] Specifically, within the predetermined target time period, the inspection robot first collects continuous time-series acceleration data. The first target shard parameters are extracted from these data using a sliding time window; then the first target shard parameters are processed in reverse order to obtain the second target shard parameters. Matching parameters are respectively extracted from these two shards according to the preset acceleration parameter threshold, and then the corresponding first target bump feature and second target bump feature are constructed. The similarity between these two target bump features and the bump feature at the k-th feature position extracted from the feature positions stored in the history is calculated. If the similarity between any target bump feature and the historical bump feature is less than or equal to the preset similarity threshold, it is determined that the similarity verification is passed. The system obtains the feature positions traversed by the inspection robot in the target time period history and extracts the feature position at the last traversal. If the feature position at the last traversal is connected to the k-th feature position through a topological edge in the self-correcting topology network, it is regarded as passing the auxiliary verification. When both the similarity verification and the auxiliary verification are successful, the system will obtain the current position information estimated by the inertial measurement unit, as well as the start coordinate and end coordinate of the k-th feature position. If the similarity between the first target bump feature and the bump feature at the k-th feature position is less than or equal to the threshold, the current estimated coordinate is corrected to the end coordinate of the feature position; if the similarity between the second target bump feature and the bump feature at the k-th feature position is less than or equal to the threshold, the current estimated coordinate is corrected to the start coordinate of the feature position.

[0097] In the target mine, since most roads may randomly trigger jolt features, relying solely on similarity verification (i.e., comparing the similarity of the jolt features based on the currently collected data and the jolt features at the pre-stored feature positions) may lead to misjudgment: some accidental acceleration mutations may have a similarity lower than the preset threshold with the jolt features at any feature position, thus wrongly triggering correction. For this reason, the auxiliary verification introduces the spatial association of the self-correcting topology network, that is, it is required that the currently detected feature position must be connected to the feature position recorded in the previous traversal through a topological edge in the network. Such a dual-verification mechanism can effectively eliminate the mis-matching caused by random triggering, ensuring that only real and continuous feature positions are used for coordinate correction, thereby improving the accuracy and robustness of the overall calibration.

[0098] In an embodiment of the present invention, the similarity of the jolt features is calculated based on the Euclidean distance.

[0099] An inertial navigation inspection robot for underground coal mines includes: a controller, a memory, a computer program stored on the memory and executable by the controller, and a data bus for realizing the connection and communication between the controller and the memory. When the computer program is executed by the controller, it realizes any one of the inertial navigation methods for underground coal mines described above.

[0100] In the above embodiments, the reference to "this embodiment" in the specification means that the specific features, structures or characteristics described in combination with the embodiments are included in at least some embodiments, but not necessarily all embodiments. Multiple occurrences of "this embodiment" do not necessarily all refer to the same embodiment.

[0101] In the above embodiments, although the present invention has been described in combination with specific embodiments of the present invention, according to the previous description, many substitutions, modifications and deformations of these embodiments will be obvious to those of ordinary skill in the art. For example, other storage structures (such as dynamic RAM (DRAM)) can be used in the embodiments discussed. The embodiments of the present invention are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims.

[0102] This embodiment also provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it realizes any one of the methods in this embodiment.

[0103] This embodiment also provides an electronic terminal, including: a processor and a memory;

[0104] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the terminal executes any one of the methods in this embodiment.

[0105] For the computer-readable storage medium in this embodiment, those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to a computer program. The foregoing computer program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including those of the above method embodiments; and the foregoing storage medium includes: various media such as ROM, RAM, magnetic disks, or optical discs that can store program codes.

[0106] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver, and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store a computer program, the communication interface is used for communication, and the processor and the transceiver are used to run the computer program so that the electronic terminal executes each step of the above method.

[0107] In this embodiment, the memory may include a random access memory (Random Access Memory, abbreviated as RAM), and may also include a non-volatile memory, such as at least one disk memory.

[0108] The above-mentioned processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, abbreviated as CPU), a network processor (Network Processor, abbreviated as NP), etc.; it may also be a digital signal processor (Digital Signal Processing, abbreviated as DSP), an application specific integrated circuit (Application Specific Integrated Circuit, abbreviated as ASIC), a field-programmable gate array (Field-Programmable Gate Array, abbreviated as FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0109] The present invention can be used in many general-purpose or special-purpose computing system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, small computers, large computers, distributed computing environments including any of the above systems or devices, and so on.

[0110] The present invention may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including storage devices.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention may be modified or equivalently replaced without departing from the spirit and scope of the present technical solution, and they should all be covered by the scope of the claims of the present invention.

Claims

1. An inertial navigation method for underground coal mines, characterized in that: Including the following steps: Initialize the inspection robot and control the inspection robot to traverse the roads in the target coal mine shaft to obtain time-series acceleration parameters; Extract the time-series acceleration parameters to obtain a number of feature data, and obtain the feature positions corresponding to each feature data; Control the inspection robot to traverse each feature position N times to obtain N groups of acceleration parameters for each feature position, where N is a positive integer; Correlate based on the N groups of acceleration parameters for each feature position: If the correlation is successful, obtain the bump feature of the corresponding feature position; If the correlation fails, delete the corresponding feature position; Construct a self-correcting topology network based on the feature positions, and the self-correcting topology network is used to assist in verifying the bump feature; Perform self-correction on the inspection robot based on the self-correcting topology network and the bump feature.

2. The inertial navigation method for underground coal mines according to claim 1, wherein: The extraction of the time-series acceleration parameters to obtain a number of feature data includes: Establish an acceleration parameter threshold and initialize a sliding time window; Extract the time-series acceleration parameters in slices based on the sliding time window to obtain slice parameters, and compare the acceleration parameters at each moment in the slice parameters with the acceleration parameter threshold. If there are no less than a preset number of acceleration parameters ≥ the acceleration parameter threshold, use the corresponding slice parameter as the feature data and obtain the feature position of the corresponding feature data; Among them, the feature position includes the coordinates of each acceleration parameter within the corresponding sliding time window, and the coordinates of the acceleration parameter are calculated by the inertial measurement unit preset in the inspection robot.

3. The inertial navigation method for underground coal mines according to claim 1, characterized in that: The traversing each feature position N times includes: For the i-th feature position, select times to perform forward traversal and backward traversal on the i-th feature position to obtain the corresponding N sets of acceleration parameters; where the forward traversal and backward traversal are both manually defined; N is an even number; i is the feature position index and is a positive integer.

4. The inertial navigation method for underground coal mines according to claim 1, wherein: The correlating based on the N groups of acceleration parameters for each feature position includes: For the i-th feature position, perform reverse traversal on the group of acceleration parameters in reverse order, including: obtaining the central moment of the acceleration parameters of the j-th group in reverse traversal, and symmetrically swapping the moments of each acceleration parameter in the j-th group with respect to the central moment to obtain the reverse acceleration parameters of the acceleration parameters of the j-th group in reverse traversal, where j is a positive integer; Combine the set of reverse acceleration parameters with the set of acceleration parameters traversed forward for combined association, which includes: comparing the acceleration parameter at each moment in the N sets of acceleration parameters with the acceleration parameter threshold, extracting the acceleration parameters greater than or equal to the acceleration parameter threshold in each set of acceleration parameters as matching parameters, and performing time association and spatial association on the matching parameters; Initialize and generate M spatio-temporal alignment points, and assign N matching parameters to each spatio-temporal alignment point. Each matching parameter comes from a different group of acceleration parameters, and each matching parameter can only be assigned to one spatio-temporal alignment point, where M is a positive integer; The time correlation includes: Obtain the maximum moment and the minimum moment corresponding to the N matching parameters of each spatio-temporal alignment point to calculate the time difference, and accumulate and sum the time differences of the M spatio-temporal alignment points to obtain the alignment cost of the time error; The space correlation includes: Obtain the maximum acceleration parameter and the minimum acceleration parameter corresponding to the N matching parameters of each spatio-temporal alignment point to calculate the space difference, and accumulate and sum the space differences of the M spatio-temporal alignment points to obtain the alignment cost of the space error; If the alignment cost of the time error is less than or equal to the corresponding time threshold, and the alignment cost of the space error is less than or equal to the corresponding space threshold, the correlation is successful, otherwise the correlation fails.

5. The inertial navigation method for underground coal mines according to claim 4, characterized in that: Minimize the alignment cost of the time error and the alignment cost of the space error based on the genetic algorithm to obtain the minimized alignment cost of the time error and the minimized alignment cost of the space error, and obtain the correlation result based on the minimized alignment cost of the time error and the minimized alignment cost of the space error; The correlation result includes: correlation success and correlation failure.

6. The inertial navigation method for underground coal mines according to claim 1, characterized in that: The process of obtaining the bump feature includes: If the association of the i-th feature position is successful, obtain the matching parameters of the i-th feature position with respect to each of the M spatio-temporal alignment points, and calculate the spatial mean and temporal mean of the matching parameters of each spatio-temporal alignment point; The jolt feature is encoded and represented by a 2×M matrix as follows: where S1…SM represent the spatial means of the 1st to Mth spatio-temporal alignment points, T1…TM represent the temporal means of the 1st to Mth spatio-temporal alignment points, and T1…TM are sorted in chronological order, and S1…SM are sorted correspondingly according to T1…TM.

7. The inertial navigation method for underground coal mines according to claim 1, characterized in that: The self-correcting topology network constructed based on the feature positions includes: Map each feature position to a topology node; If the i-th feature position and the k-th feature position are adjacent, construct a topology edge between the i-th topology node and the k-th topology node to obtain a self-correcting topology network; where i≠k, adjacent means that there is no g-th feature position on the shortest path between the i-th feature position and the k-th feature position, i≠k≠g, and k and g are both positive integers.

8. The inertial navigation method for underground coal mines according to claim 7, characterized in that: The self-calibration of the inspection robot based on the self-correcting topology network and the jitter feature includes: Step A1: During the target time period, obtain the time-series acceleration parameters based on the inspection robot, and extract the first target shard parameters from the time-series acceleration parameters based on a sliding time window; reverse the first target shard parameters to obtain the second target shard parameters; Extract the first target matching parameters and the second target matching parameters from the first target shard parameters and the second target shard parameters respectively based on the acceleration parameter threshold, and construct the corresponding first target jitter feature and second target jitter feature respectively based on the first target matching parameters and the second target matching parameters; calculate the similarity between the first target jitter feature and the second target jitter feature and the jitter feature of the k-th feature position. If the similarity between any one of the first target jitter feature and the second target jitter feature and the jitter feature of the k-th feature position is less than or equal to the preset similarity threshold, then obtain a similarity verification; Step A2: Obtain the feature positions traversed by the inspection robot historically during the target time period, and extract the feature position of the last traversal; if there is a topology edge constructed between the feature position of the last traversal and the topology node corresponding to the k-th feature position, then obtain an auxiliary verification; Step A3: If both the similarity verification and the auxiliary verification are successful, obtain the estimated coordinates of the inertial measurement unit of the inspection robot, and obtain the starting coordinates and ending coordinates of the k-th feature position; If the similarity between the first target jitter feature and the jitter feature of the k-th feature position is less than or equal to the preset similarity threshold, then correct the estimated coordinates to the ending coordinates; If the similarity between the second target jitter feature and the jitter feature of the k-th feature position is less than or equal to the preset similarity threshold, then correct the estimated coordinates to the starting coordinates; Among them, for the first feature position during the target time period, the inspection robot only needs to pass the similarity verification.

9. The inertial navigation method for underground coal mines according to claim 8, wherein: The similarity of the jitter feature is calculated based on the Euclidean distance.

10. An inertial navigation inspection robot for underground coal mines, comprising: A controller, a memory, a computer program stored on the memory and executable by the controller, and a data bus for realizing the connection and communication between the controller and the memory, characterized in that: when the computer program is executed by the controller, it realizes the inertial navigation method for underground coal mines as described in any one of claims 1-9.

Images