Underground tunnel sensing and mapping system and method for engineering machinery

By integrating sensors and edge computing processors on construction machinery equipment, map data can be processed and updated in real time, and communication stability is enhanced through wireless communication systems, the problems of instability in communication and incomplete map data in underground tunnel operations are solved, and the problems of instability in underground tunnel operations are improved. Operational safety and data processing efficiency of unmanned construction machinery.

CN118913241APending Publication Date: 2024-11-08JIANGSU XCMG STATE KEY LAB TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410852273.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-28
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The operating environment of underground tunnels is complex, the wireless communication of traditional engineering machinery is unstable, and the map data generated by SLAM technology is incomplete, which cannot meet the real-time operation needs of unmanned engineering machinery.

Method used

Using sensors, SLAM processors and edge computing processors to install on engineering machinery equipment, the environment data is processed in real time through edge computing, and the stereo map is generated and updated, and communication stability is enhanced through wireless communication systems, and map splicing is used to use multiple sensor data. and optimize.

Benefits of technology

It improves the real-time and efficiency of data processing, generates more realistic and consistent map data, and enhances the operational safety and communication stability of underground unmanned construction machinery.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118913241A_ABST
    Figure CN118913241A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of engineering machinery underground tunnel mapping, in particular to an engineering machinery underground tunnel sensing and mapping system and method. The engineering mechanical equipment, the sensor component and the SLAM processor are used for processing according to the initial environment data to obtain initial three-dimensional map data; the edge calculation processor is used for preprocessing the preliminary environment data; the data processing module is used for carrying out supplementary updating on the preliminary three-dimensional map data according to the supplementary environment data so as to obtain updated three-dimensional map data; the splicing module is also used for splicing to obtain spliced map data; and the control system feedback processor is used for transmitting the spliced map data to the vehicle end controller. According to the invention, the edge computing processor is installed on the engineering mechanical equipment working underground, so that the real-time performance and efficiency of data processing are remarkably improved compared with the mode of transmitting data to a cloud base station for processing, and the dependence on cloud computing resources is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of underground tunnel mapping for engineering machinery, and in particular to a system and method for sensing and mapping underground tunnels for engineering machinery. Background Art

[0002] With the acceleration of urbanization and the continuous advancement of infrastructure construction, the demand for underground engineering is growing. In underground tunnel operations, engineering machinery plays a vital role, including tasks such as excavation, transportation, and support. However, the complexity and unpredictability of the underground environment poses great challenges to the operation of engineering machinery, especially in wireless communication and real-time perception and mapping.

[0003] The underground tunnel working environment is characterized by narrow space, insufficient light, complex and changeable terrain, which places high demands on the operation of engineering machinery. Traditional engineering machinery mainly relies on manual operation, which has problems of low efficiency and poor safety. In recent years, unmanned engineering machinery has gradually become a research hotspot. By introducing automation and intelligent technology, the operating efficiency and safety can be improved. However, unmanned engineering machinery faces many technical difficulties when operating in underground environments.

[0004] First of all, the traditional underground tunnel mapping work requires uploading the data collected by the vehicle-mounted sensors to the cloud base station for processing. The real-time performance of data processing is poor, and it is prone to network congestion and delay problems, which cannot meet the real-time operation requirements of unmanned construction machinery. In addition, traditional SLAM technology mainly relies on sensor data from lidar, cameras, and inertial measurement units, and realizes environmental perception and map construction by fusing multiple sensor data. However, in the underground tunnel working environment, factors such as insufficient light and flying dust can significantly affect the performance of the sensor, resulting in incomplete perceived map data information, and the SLAM technology itself is not easy to extend and add other sensor data. Summary of the invention

[0005] The purpose of the present invention is to provide an underground tunnel perception and mapping system and method for engineering machinery, so as to solve the technical problems of unstable underground communication and incomplete data information in the three-dimensional map generated by SLAM technology (simultaneous localization and mapping) proposed in the above background technology. To solve the above problems, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a real-time perception and mapping system for underground tunnels, comprising: Engineering machinery equipment, used for traveling in the underground tunnel; A sensor component is installed on the engineering machinery equipment and is used to collect environmental data around the driving position of the engineering machinery equipment when the engineering machinery equipment is driving, wherein the environmental data includes preliminary environmental data and supplementary environmental data; A SLAM processor, installed on the engineering machinery equipment, is in communication connection with the sensor component, and is used to process the preliminary environmental data to obtain preliminary three-dimensional map data; an edge computing processor, installed on the engineering machinery equipment, communicating with the sensor component and the SLAM processor, for supplementing and updating the preliminary 3D map data according to the supplementary environmental data to obtain updated 3D map data; and for splicing the updated 3D map data obtained at different driving positions of the engineering machinery equipment to obtain spliced ​​map data; A control system feedback processor is communicatively connected to the edge computing processor and is used to transmit the stitched map data to the vehicle-side controller for controlling the driving of the engineering machinery equipment.

[0006] Optionally, the sensor components corresponding to the preliminary environmental data and the corresponding environmental data include: LiDAR, used to obtain scanning data around construction machinery and equipment; A camera is used to capture environmental image data around the construction machinery and equipment; Inertial measurement sensors are used to measure the attitude and motion state of engineering machinery and equipment to collect IMU attitude data; The sensor components corresponding to the supplementary environmental data and the corresponding environmental data at least include: Ultrasonic sensors are used to collect ultrasonic data around construction machinery and equipment.

[0007] Optionally, the edge computing processor is further used to pre-process the preliminary environmental data to obtain filtered, denoised and corrected pre-processed data.

[0008] Optionally, it also includes: A cloud processor wirelessly connected to the control system feedback processor; The control system feedback processor is also used to upload the stitched map data to a cloud processor for subsequent access and use.

[0009] Optionally, the cloud processor is communicatively connected to the control system feedback processor via a wireless communication system; the wireless communication system comprises: Wireless communication nodes are used to be arranged in the underground tunnel and form a grid network, wherein each node has a relay function; Dynamic communication links, used for communication connections between wireless communication nodes; Wired communication nodes are used to be arranged in underground tunnels to connect to wired communication nodes of cloud public network base stations.

[0010] In a second aspect, the present invention further provides a method for real-time perception and mapping of underground tunnels, which is performed using the real-time perception and mapping system for underground tunnels, and comprises: Acquire environmental data around the engineering machinery equipment; the environmental data includes preliminary environmental data for SLAM algorithm processing and supplementary environmental data for edge computing processor processing; Preprocessing the preliminary environmental data using an edge computing processor; The preliminary environment data after preprocessing is processed by using a SLAM algorithm to obtain preliminary three-dimensional map data; Using the edge computing processor to supplement and update the preliminary 3D map data with the supplementary environment data to obtain updated 3D map data; splicing updated three-dimensional map data obtained when the engineering machinery equipment is located at different positions to obtain spliced ​​map data; The stitched map data is transmitted to a vehicle-side controller for controlling the driving of the engineering machinery equipment.

[0011] Optionally, The sensor components corresponding to the preliminary environmental data and the corresponding environmental data include: LiDAR, used to obtain scanning data around construction machinery and equipment; A camera is used to capture environmental image data around the construction machinery and equipment; Inertial measurement sensors are used to measure the attitude and motion state of engineering machinery and equipment to collect IMU attitude data; The obtaining of preliminary three-dimensional map data by using the SLAM algorithm according to the preliminary environment data comprises: Acquiring the preliminary environmental data collected at different times; According to the preliminary environmental data, using a particle filter algorithm to predict the pose estimation result of the engineering machinery equipment; Preliminary three-dimensional map data of the engineering machinery equipment at its location is generated in real time according to the position estimation result and the environmental image data.

[0012] Optionally, the supplementary environmental data includes ultrasonic data; The supplementary updating of the preliminary three-dimensional map data according to the pre-processed supplementary environmental data includes: the ultrasonic data includes reflection signal data and reflection reception time difference data; identifying and detecting obstacles in the environment based on the reflected signal data; Determining the relative position of the obstacle and the engineering machinery vehicle according to the reflection reception time difference data; When a new obstacle is detected: Adding the collected obstacles to the preliminary 3D map data obtained by SLAM algorithm processing in the form of map marks to obtain updated 3D map data; and / or For an existing obstacle, when its position changes, the obstacle with the changed position is updated into the preliminary 3D map data to obtain updated 3D map data.

[0013] Optionally, the pose estimation result of the engineering machinery equipment is predicted using a particle filter algorithm according to the preliminary environmental data; wherein the propagation model of the particle filter algorithm is: ; ; for The state vector at the moment; for The state vector at the moment; the state vector represents the pose estimation result of the engineering machinery equipment at this moment, for The control vector at a certain moment represents the environment image data and IMU attitude data at that moment; is the control vector in the time period from 1 to t-1, representing the environmental image data and IMU attitude data in this time period: is the observation vector at time t, representing the environmental scan data at that time; is the observation vector in the time period 1~t-1, representing the environmental scanning data in this time period; Represented by assuming the state vector and the assumed weight vector The matrix composed of n represents the number of particles; is a hypothetical state vector, representing a position and posture estimation result of the engineering machinery equipment (101) in a state hypothesized at time t; is the assumed weight vector, representing The matching degree between the pose estimation result in and the observation vector at time t, the higher the matching degree, the higher the weight; The method of using a particle filter algorithm to predict and obtain the estimated position and posture result of the engineering machinery equipment includes: randomly generating a number of particles; according to and Get the particle distribution at time t-1. The particle distribution at time t-1 is: in, Respectively represent the hypothesized state vector and hypothesized weight vector at time t-1; according to Propagate each particle from time t-1 to time t to predict the particle distribution at time t ; according to The assumed weight vector for each particle Make corrections, The correction formula is: ; According to the correction results, the particles with low weights are discarded, the weights of the remaining particles after the discarding are normalized, and the particles are resampled according to the normalized weights to supplement the discarded particles; Substituting the resampled particles into the propagation model and iterating cyclically to obtain a final particle set; The final particle set prediction is used to obtain the position and posture estimation result of the engineering machinery equipment at time t.

[0014] Optionally, before the updated three-dimensional map data are spliced ​​to obtain spliced ​​map data, map optimization is required; the map optimization uses an optimization algorithm, Levenberg-Marquardt algorithm, to optimize the map structure and eliminate accumulated errors.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention installs the edge computing processor (edge ​​computing unit, edge cloud) on engineering machinery equipment working underground, which significantly improves the real-time and efficiency of data processing compared to transmitting data to a cloud base station for processing, and reduces dependence on cloud computing resources.

[0016] 2. The present invention can also expand more sensors on mechanical equipment by arranging edge computing processors. Compared with the traditional technology that only generates three-dimensional maps through SLAM algorithms, more sensors can be used to make the map data contain more useful information, so as to more realistically reflect the underground environment and ensure high-precision and high-consistency map generation, so as to improve the operational safety of unmanned construction machinery and equipment in underground work.

[0017] 3. The present invention builds a wireless communication system underground, including multiple wireless communication nodes, dynamic communication links and wired communication nodes, which enhances the stability and coverage of communication and solves the communication problems in underground environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a schematic diagram of the underground tunnel real-time perception and mapping system provided in Example 1; Figure 2It is a schematic diagram of the wireless communication system of the underground tunnel real-time perception and mapping system in Example 1, and the cloud public network base station; Figure 3 It is a flow chart of a method for real-time perception and mapping of underground tunnels in Example 2. DETAILED DESCRIPTION

[0019] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention. Example 1

[0020] Combination Figure 1 This embodiment provides a real-time perception and mapping system for underground tunnels, which includes: The engineering machinery equipment 101, as a carrier of the other components of the whole system except the wireless communication system 104, is used for traveling in the underground tunnel; A wireless communication system 104, configured to be arranged in the underground tunnel; Specifically, the wireless communication system 104 used in this embodiment is to solve the problem of unstable wireless communication in underground environments. Since most passenger cars work on the ground, there is no need to arrange a wireless communication system 104 for the vehicle separately. The engineering machinery vehicle in this embodiment needs to work underground, and there is no commercial wireless communication system 104 arranged underground. Therefore, this embodiment arranges the wireless communication system 104 in the underground tunnel in an ad hoc manner, combined with Figure 2 The wireless communication system 104 includes: wireless communication nodes 202, which are arranged in underground tunnels and form a grid network, in which each node has a relay function; dynamic communication links 203, which are used for communication connections between nodes; used to transmit data, using adaptive frequency hopping technology, real-time adjustment according to network status to ensure reliable data transmission; wired communication nodes 201, which are used to connect public network base stations and wired communication nodes 201, and are deployed in slender tunnels with regular terrain to ensure long-distance communication reliability.

[0021] The specific layout process is: Node deployment: multiple wireless communication nodes 202 are arranged in the underground tunnel to form a grid network. Each node has a relay function to ensure that the communication covers the entire operation area. The communication protocol uses adaptive frequency hopping technology and multi-channel communication to reduce the impact of electromagnetic interference and improve the reliability of the communication link. A dynamic routing algorithm is introduced to adjust the data transmission path in real time according to the node status and network topology to avoid network congestion and data loss. At the same time, each node synchronizes data regularly to ensure the clock consistency and data consistency of the entire network.

[0022] A sensor component 103 is installed on the engineering machinery equipment 101 and is used to collect environmental data around the engineering machinery equipment 101 when the engineering machinery equipment 101 is traveling, wherein the environmental data includes preliminary environmental data for SLAM (Simultaneous Localization and Mapping) algorithm processing and supplementary environmental data for edge computing processing; A laser radar, used to obtain environmental scanning data around the engineering machinery equipment 101; A camera, used to capture image data of the environment around the engineering machinery equipment 101; An inertial measurement sensor, used to measure the attitude and motion state of the engineering machinery equipment 101 to collect IMU attitude data; The ultrasonic sensor is used to collect ultrasonic data around the engineering machinery equipment 101.

[0023] A SLAM processor is installed on the engineering machinery equipment 101 and is connected to the sensor component 103 by wired communication; it is used to obtain preliminary three-dimensional map data according to the processed preliminary environmental data; The edge computing processor 102 is installed on the engineering machinery equipment 101, is connected to the sensor component 103 and the SLAM processor by wired communication, and is used to supplement and update the preliminary three-dimensional map data according to the supplementary environmental data to obtain updated three-dimensional map data; and is also used to splice the updated three-dimensional map data at different positions of the engineering machinery equipment 101 to obtain spliced ​​map data; Specifically, the edge computing unit in this embodiment can be directly installed on the engineering machinery equipment 101 to perform real-time calculations without uploading all data to the cloud for calculation, which reduces the pressure on the cloud and avoids the situation where the cloud wireless communication is disconnected. Because the edge computing unit is installed on the vehicle, it can communicate directly with the sensor component 103 and the SLAM processor through wires.

[0024] Moreover, since the SLAM processor can only obtain preliminary three-dimensional map data through lidar, cameras, and inertial measurement sensors, the sensor data that can be acquired and processed is relatively limited, and the information obtained is also relatively limited.

[0025] On the one hand, the edge computing unit of this embodiment can reduce the pressure of cloud computing, and on the other hand, it can also enable further expansion of vehicle-mounted sensors. Although the newly added sensor data is difficult to process by the SLAM processor, it can be directly processed by the edge computing unit, and then the new information obtained after processing is supplemented and updated to the three-dimensional map data obtained by the SLAM processor, so that the final three-dimensional map data can contain more useful information obtained through sensor data, so as to facilitate the underground unmanned intelligent operation of engineering machinery equipment 101.

[0026] The edge computing processor 102 (edge ​​computing unit) is the core component of this system, which is mainly responsible for processing sensor data. The edge computing processor 102 can not only pre-process the preliminary environmental data to obtain filtered, denoised and corrected pre-processed data, and use it to fuse the pre-processed data in the particle filtering algorithm so that the SLAM algorithm can reconstruct a three-dimensional map based on the preliminary environmental data, but can also perform the same processing on the supplementary environmental data to obtain useful information through sensor data.

[0027] In this embodiment, the edge computing processor 102 uses a multi-core processor to support parallel computing and meet high real-time requirements. It is equipped with a large-capacity memory for storing sensor data and generated updated stereo map data and spliced ​​map data, and optimizes the hardware and software architecture to ensure long-term stable operation in complex underground environments.

[0028] In addition, the edge computing processor 102 in this embodiment is directly deployed on the engineering machinery, close to the sensor component 103 and the wireless communication module of the wireless communication system 104, which can reduce data transmission delay and improve the response speed of the system.

[0029] The control system feedback processor 105 is connected to the edge computing processor 102 by wired communication, and is used to feed back the stitched map data to the vehicle-side controller for controlling the driving of the engineering machinery equipment 101.

[0030] In this embodiment, cloud-side and vehicle-side communication is mainly carried out through a control system feedback processor 105, which is installed on the engineering machinery equipment 101. It can directly feed back the edge computing unit supplement and the stitched map data to the vehicle-side processor used for unmanned vehicle control through wired communication, and can also be uploaded to the cloud for storage through the wireless communication network 104 built by the underground self-organizing network. Example 2

[0031] This embodiment provides a method for real-time perception and mapping of underground tunnels, which is performed using the real-time perception and mapping system for underground tunnels described in Embodiment 1, and includes: Step S1, obtaining environmental data around the engineering machinery equipment 101; the environmental data includes preliminary environmental data and supplementary environmental data; The sensor component 103 and the corresponding environmental data corresponding to the preliminary environmental data include: A laser radar, used to obtain environmental scanning data around the engineering machinery equipment 101; A camera, used to capture image data of the environment around the engineering machinery equipment 101; An inertial measurement sensor, used to measure the attitude and motion state of the engineering machinery equipment 101 to collect IMU attitude data; The sensor component 103 corresponding to the supplementary environmental data and the corresponding environmental data at least include: The ultrasonic sensor is used to collect ultrasonic data around the engineering machinery equipment 101.

[0032] Step S2: pre-processing the preliminary environmental data using an edge computing processor (102); Among them, the preprocessing is used to obtain the preprocessed data after filtering, denoising and correction.

[0033] Step S3, obtaining preliminary three-dimensional map data by using the SLAM algorithm according to the pre-processed preliminary environmental data; Step S3 specifically includes: Step S31, obtaining the preliminary environmental data collected at different times; Step S32: predicting the position and posture estimation result of the engineering machinery equipment 101 by using a particle filter algorithm according to the preliminary environmental data; Step S33: Generate preliminary three-dimensional map data of the location of the engineering machinery equipment 101 in real time according to the position estimation result and the environmental image data.

[0034] In step S32, the particle filter algorithm is used to predict the pose estimation result of the engineering machinery equipment 101. The propagation model of the particle filter algorithm is: ; ; in, for The state vector at the moment; for The state vector at the moment; the state vector represents the pose estimation result of the engineering machinery equipment at this moment, for The control vector at a certain moment represents the environment image data and IMU attitude data at that moment; is the control vector in the time period from 1 to t-1, representing the environmental image data and IMU attitude data in this time period: is the observation vector at time t, representing the environmental scan data at that time; is the observation vector in the time period 1~t-1, representing the environmental scanning data in this time period; Represented by assuming the state vector and the assumed weight vector The matrix composed of n represents the number of particles; is a hypothetical state vector, representing a pose estimation result of the engineering machinery equipment 101 under a state assumed at time t; is the assumed weight vector, representing The matching degree between the pose estimation result in and the observation vector at time t, the higher the matching degree, the higher the weight; The pose estimation result of the engineering machinery equipment 101 predicted by the particle filter algorithm includes: Generate a number of particles randomly; according to and Get the particle distribution at time t-1. The particle distribution at time t-1 is: in, Respectively represent the hypothesized state vector and hypothesized weight vector at time t-1; according to Propagate each particle from time t-1 to time t to predict the particle distribution at time t ; according to The assumed weight vector for each particle Make corrections, The correction formula is: ; According to the correction results, the particles with low weights are discarded, the weights of the remaining particles after the discarding are normalized, and the particles are resampled according to the normalized weights to supplement the discarded particles; Substituting the resampled particles into the propagation model and iterating cyclically to obtain a final particle set; The final particle set prediction is used to obtain the position and posture estimation result of the engineering machinery equipment at time t.

[0035] In this embodiment, the preliminary three-dimensional map data is further optimized by a graph optimization algorithm. The graph optimization algorithm adopts the optimization algorithm Levenberg-Marquardt algorithm to optimize the structure of the graph and eliminate the accumulated error.

[0036] Step S4: using the edge computing processor 102 to supplement and update the preliminary 3D map data with the supplementary environment data to obtain updated 3D map data; Before the supplementary update, the supplementary environment data is first preprocessed by using an edge computing unit, and the preliminary 3D map data is supplemented and updated according to the preprocessed supplementary environment data to obtain updated 3D map data; The supplementary environmental data includes ultrasonic data; The process of ultrasonic data collection and transmission is as follows: first, install an ultrasonic sensor on the mobile engineering machinery 101 and ensure that it can cover the surrounding environment area to be sensed. Initialize the ultrasonic sensor, including setting the detection frequency, sampling interval, working mode, etc.

[0037] The ultrasonic sensor continuously transmits ultrasonic pulses to the surrounding environment and receives reflected signals according to the set frequency and sampling interval to obtain reflected signal data and reflection reception time difference data. According to the time difference between the transmission and reception of ultrasonic waves, the distance and relative angle information between the obstacle and the sensor are calculated.

[0038] The collected ultrasonic data is filtered to remove noise and interference signals, and the data is compressed and encoded to facilitate transmission to the edge computing processor 102 for processing.

[0039] The pre-processed ultrasonic data is transmitted to the edge computing unit through the wireless communication network 104. The stability and real-time performance of data transmission are ensured to meet the requirements of SLAM map optimization. The data transmission protocol uses MQTT for communication between the ultrasonic sensor and the edge computing unit, which can transmit data under low power consumption and network constraints.

[0040] The obstacle distance and relative angle information collected by the ultrasonic sensor are encoded and packaged according to the requirements of the selected transmission protocol. The data format should contain necessary information fields: timestamp, sensor ID, obstacle distance, obstacle angle, etc.

[0041] According to the needs of the application scenario, set the appropriate data transmission frequency. In terms of data compression and encryption, compress the ultrasonic data to reduce the amount of data transmitted and improve the transmission efficiency. At the same time, in order to ensure the security of the data, the weighted fractional Fourier technology is used to encrypt the data to ensure that it is not maliciously intercepted and tampered with during the transmission process.

[0042] Real-time monitoring of the quality of data transmission, including data transmission rate, packet loss rate, delay and other indicators. By monitoring these data, problems that arise during data transmission can be discovered and resolved in a timely manner to ensure data integrity and real-time performance.

[0043] After receiving the ultrasonic data, the edge computing unit decodes and decompresses it, extracts the distance and relative angle information of the obstacle, and supplements and integrates it with the existing SLAM map data.

[0044] Step S4 specifically includes: Step S41, the ultrasonic data includes reflection signal data and reflection reception time difference data; Step S42: identifying and detecting obstacles in the environment according to the reflected signal data; Step S43, determining the relative position of the obstacle and the engineering machinery vehicle according to the reflection reception time difference data; Step S44: When a new obstacle is detected: Adding the collected obstacles to the preliminary 3D map data obtained by SLAM algorithm processing in the form of map marks to obtain updated 3D map data; and / or For an existing obstacle, when its position changes, the obstacle with the changed position is updated into the preliminary 3D map data to obtain updated 3D map data.

[0045] For step S4, the following optimization strategy is adopted: Loop detection: By detecting whether the engineering machinery equipment 101 passes through a known area (i.e., loop), the accumulated error is eliminated and the accuracy of the map is improved. When a loop is detected, the known information can be used to perform global optimization on the map.

[0046] Local optimization: During the movement of the construction machinery 101, the local map is optimized in real time based on the newly acquired ultrasonic data and other sensor data. This can be done by adjusting the position and shape of obstacles in the local map to make it more consistent with the actual environment.

[0047] Multi-sensor fusion optimization: Combine data from different sensors (such as environmental scanning data from lidar, environmental image data from cameras, etc.) to perform multi-source information fusion optimization on the map to improve the accuracy and robustness of the map.

[0048] It can also adjust relevant parameters in the SLAM algorithm (such as the weight of sensor data fusion, map update frequency, etc.) according to actual application scenarios and needs to optimize the quality and performance of the map.

[0049] Step S5, splicing the updated 3D map data obtained when the engineering machinery equipment 101 is located at different positions to obtain spliced ​​map data; Step S6: transmitting the stitched map data to the vehicle-side controller for controlling the driving of the engineering machinery equipment 101.

[0050] For step S6, the optimized and stitched map data is stored on the server of the edge cloud for subsequent access and use.

[0051] If necessary, the optimized stitched map data is distributed to other devices or systems to achieve map sharing and collaborative use.

[0052] In addition, for the optimization process, the optimized updated 3D map data and the spliced ​​map data are evaluated, including map accuracy, consistency, real-time performance, etc. Based on the evaluation results, the processing algorithms of the ultrasonic sensor and the edge computing processor 102 are fed back and adjusted to further improve the effect of map optimization.

[0053] It is easy to understand that those skilled in the art can combine, split, reorganize, etc. the embodiments of the present application to obtain other embodiments based on one or several embodiments provided in the present application, and these embodiments do not exceed the protection scope of the present application.

[0054] The present invention and its implementation methods are described schematically above, and the description is not restrictive. The embodiments shown in the embodiments are only part of the implementation methods of the present invention, and the actual structure is not limited thereto. Therefore, if a person skilled in the art is inspired by the embodiments and designs a structure and an implementation method similar to the technical solution without creativity without departing from the purpose of the invention, they should all fall within the protection scope of the present invention.

Claims

1. A real-time perception and mapping system for underground tunnels, characterized in that: include: Engineering machinery equipment (101), used for traveling in the underground tunnel; A sensor component (103) is mounted on the engineering machinery equipment (101) and is used to collect environmental data around the driving position of the engineering machinery equipment (101) when the engineering machinery equipment (101) is driving, wherein the environmental data includes preliminary environmental data and supplementary environmental data; A SLAM processor, installed on the engineering machinery equipment (101), is in communication connection with the sensor component (103), and is used to process the preliminary environmental data to obtain preliminary three-dimensional map data; An edge computing processor (102) is installed on the engineering machinery equipment (101), is communicatively connected to the sensor component (103) and the SLAM processor, and is used to supplement and update the preliminary three-dimensional map data according to the supplementary environmental data to obtain updated three-dimensional map data; and is also used to splice the updated three-dimensional map data obtained when the engineering machinery equipment (101) is located at different driving positions to obtain spliced ​​map data; A control system feedback processor (105) is communicatively connected to the edge computing processor (102) and is used to transmit the spliced ​​map data to a vehicle-side controller for controlling the travel of the engineering machinery equipment (101).

2. The underground tunnel real-time perception and mapping system according to claim 1 is characterized in that: The sensor component (103) corresponding to the preliminary environmental data and the corresponding environmental data include: A laser radar for acquiring scanning data around the construction machinery equipment (101); A camera, used to capture image data of the environment around the engineering machinery equipment (101); An inertial measurement sensor, used to measure the attitude and motion state of the engineering machinery equipment (101) to collect IMU attitude data; The sensor component (103) corresponding to the supplementary environmental data and the corresponding environmental data at least include: An ultrasonic sensor is used to collect ultrasonic data around the engineering machinery equipment (101).

3. The underground tunnel real-time perception and mapping system according to claim 1, characterized in that: The edge computing processor (102) is also used to pre-process the preliminary environmental data to obtain filtered, denoised and corrected pre-processed data.

4. The underground tunnel real-time perception and mapping system according to claim 1, characterized in that: Also includes: A cloud processor wirelessly connected to the control system feedback processor (105); The control system feedback processor (105) is also used to upload the stitched map data to a cloud processor for subsequent access and use.

5. The underground tunnel real-time perception and mapping system according to claim 4 is characterized in that: The cloud processor is communicatively connected to the control system feedback processor (105) via a wireless communication system (104); the wireless communication system (104) comprises: Wireless communication nodes (202) are arranged in the underground tunnel to form a grid network, wherein each node has a relay function; A dynamic communication link (203) for communication connection between wireless communication nodes (202); The wired communication node (201) is used to be arranged in an underground tunnel to connect to the wired communication node (201) of a cloud public network base station.

6. A method for real-time perception and mapping of underground tunnels, characterized by: The method comprises: utilizing the underground tunnel real-time perception and mapping system according to any one of claims 1 to 5, comprising: Acquiring environmental data around the engineering machinery equipment (101); the environmental data comprising preliminary environmental data for SLAM algorithm processing and supplementary environmental data for edge computing processor (102) processing; Preprocessing the preliminary environmental data using an edge computing processor (102); The preliminary environment data after preprocessing is processed by using a SLAM algorithm to obtain preliminary three-dimensional map data; Using the edge computing processor (102) to supplement and update the preliminary 3D map data with the supplementary environment data, so as to obtain updated 3D map data; splicing updated three-dimensional map data obtained at different positions of the engineering machinery equipment (101) to obtain spliced ​​map data; The spliced ​​map data is transmitted to a vehicle-side controller for controlling the travel of the engineering machinery equipment (101).

7. The underground tunnel real-time perception and mapping method according to claim 6, characterized in that: The sensor component (103) corresponding to the preliminary environmental data and the corresponding environmental data include: A laser radar for acquiring scanning data around the construction machinery equipment (101); A camera, used to capture image data of the environment around the engineering machinery equipment (101); An inertial measurement sensor, used to measure the attitude and motion state of the engineering machinery equipment (101) to collect IMU attitude data; The obtaining of preliminary three-dimensional map data by using the SLAM algorithm according to the preliminary environmental data comprises: Acquiring the preliminary environmental data collected at different times; Predicting a pose estimation result of the engineering machinery equipment (101) using a particle filter algorithm according to the preliminary environmental data; Preliminary three-dimensional map data of the location of the engineering machinery equipment (101) is generated in real time according to the position estimation result and the environmental image data.

8. The underground tunnel real-time perception and mapping method according to claim 7, characterized in that: The supplementary environmental data includes ultrasonic data; The supplementary updating of the preliminary three-dimensional map data according to the pre-processed supplementary environmental data includes: the ultrasonic data includes reflection signal data and reflection reception time difference data; identifying and detecting obstacles in the environment based on the reflected signal data; Determining the relative position of the obstacle and the engineering machinery vehicle according to the reflection reception time difference data; When a new obstacle is detected: Adding the collected obstacles to the preliminary 3D map data obtained by SLAM algorithm processing in the form of map marks to obtain updated 3D map data; and / or For an existing obstacle, when its position changes, the obstacle with the changed position is updated into the preliminary 3D map data to obtain updated 3D map data.

9. The underground tunnel real-time perception and mapping method according to claim 7, characterized in that: The position and posture estimation result of the engineering machinery equipment (101) is obtained by predicting the preliminary environmental data using a particle filter algorithm, wherein the propagation model of the particle filter algorithm is: ; ; for The state vector at the moment; for The state vector at the moment; the state vector represents the pose estimation result of the engineering machinery equipment at this moment, for The control vector at a certain moment represents the environment image data and IMU attitude data at that moment; is the control vector in the time period from 1 to t-1, representing the environmental image data and IMU attitude data in this time period: is the observation vector at time t, representing the environmental scan data at that time; is the observation vector in the time period 1~t-1, representing the environmental scanning data in this time period; Represented by assuming the state vector and the assumed weight vector The matrix composed of n represents the number of particles; is a hypothetical state vector, representing a position and posture estimation result of the engineering machinery equipment (101) in a state hypothesized at time t; is the assumed weight vector, representing The matching degree between the pose estimation result in and the observation vector at time t, the higher the matching degree, the higher the weight; The prediction using the particle filter algorithm to obtain the estimated position and posture result of the engineering machinery equipment (101) comprises: Generate a number of particles randomly; according to and Get the particle distribution at time t-1. The particle distribution at time t-1 is: in, Respectively represent the hypothesized state vector and hypothesized weight vector at time t-1; according to Propagate each particle from time t-1 to time t to predict the particle distribution at time t ; according to The assumed weight vector for each particle Make corrections, The correction formula is: ; According to the correction results, the particles with low weights are discarded, the weights of the remaining particles after the discarding are normalized, and the particles are resampled according to the normalized weights to supplement the discarded particles; Substituting the resampled particles into the propagation model and iterating cyclically to obtain a final particle set; The final particle set prediction is used to obtain a position and posture estimation result of the engineering machinery equipment (101) at time t.

10. The underground tunnel real-time perception and mapping method according to claim 6, characterized in that: Also includes: Before the updated three-dimensional map data are spliced ​​to obtain spliced ​​map data, map optimization is required; the map optimization adopts the optimization algorithm Levenberg-Marquardt algorithm to optimize the map structure and eliminate the accumulated errors.