Autonomous navigation method and system for coal mine underground unmanned vehicle

Abstract

The invention discloses an autonomous navigation method and system for an underground coal mine unmanned vehicle, and relates to the technical field of vehicle navigation, and the method comprises the following steps: defining interference indexes, setting interference standards by considering a plurality of interference factors, and dividing an underground coal mine map into various types of interference areas by virtue of the interference indexes, so that the underground coal mine areas are divided; mark points are set for each type of interference area in a targeted mode, and a reliable basis is provided for follow-up vehicle position information. The position information of the vehicle is adjusted by identifying the mark point through the vehicle, the accurate current position of the vehicle is obtained by assisting the identification condition of the mark point, and the reliability of the position of the vehicle is ensured. According to the method, the multi-model uncertainty motion prediction is carried out on the dynamic obstacle, and whether the task path needs to be adjusted is judged according to the obstacle event, so that the reliability and adaptability of autonomous navigation of the underground coal mine unmanned vehicle are improved, and the complex traffic demand and navigation demand of the underground coal mine unmanned vehicle are met.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle navigation, and in particular to an autonomous navigation method and system for an unmanned vehicle in a coal mine. Background Art

[0002] The underground environment of coal mines is unique, characterized by cramped spaces, dark and damp conditions, low visibility, and complex and variable geological conditions, presenting numerous safety hazards. Traditional manual driving in this environment is not only inefficient but also prone to frequent accidents. With the advancement of technology, unmanned driving technology has become the key to solving the challenges of underground transportation. This technology integrates core technologies such as high-precision positioning, sensor fusion, environmental perception, path planning, and intelligent control. It uses multiple sensors, such as lidar, cameras, and millimeter-wave radar, to acquire environmental information and, combined with advanced algorithms, enables autonomous decision-making and navigation. In the absence of satellite positioning signals underground, alternative positioning technologies, such as laser SLAM, are required to ensure accurate vehicle positioning. Furthermore, the reliability of communication technology and vehicle control systems, as well as dynamic obstacle prediction and avoidance strategies, must be considered to achieve safe and efficient underground unmanned transportation and promote the intelligent transformation of coal mines.

[0003] In the existing technology, there are many interferences in the underground environment, which leads to inaccurate vehicle positioning, low real-time obstacle avoidance capability and path flexibility. These problems together lead to poor reliability and adaptability of autonomous navigation of unmanned vehicles in coal mines, which cannot meet the needs of the complex traffic environment in coal mines.

[0004] Therefore, how to improve the reliability and adaptability of autonomous navigation of unmanned vehicles in coal mines is a technical problem that needs to be solved. Summary of the Invention

[0005] The purpose of the present invention is to solve the problem of poor reliability and adaptability of autonomous navigation of unmanned vehicles in coal mines in the prior art, and to propose an autonomous navigation method for unmanned vehicles in coal mines, which includes: Obtain road data and underground interference data in the coal mine, establish an underground coal mine map based on the road data, mark the underground interference data at the corresponding position on the underground coal mine map, and define interference indicators; The coal mine map is divided into multiple interference areas based on interference indicators. The density of marker points in each interference area is set. Multiple marker points are selected in different interference areas based on the density of marker points. The vehicle position information is adjusted by the vehicle's recognition of the marker points. While the vehicle is traveling along the mission path, it performs multi-model uncertainty motion prediction on dynamic obstacles to achieve vehicle obstacle avoidance. At the same time, it detects obstruction events and determines whether the mission path needs to be adjusted based on the obstruction events to achieve autonomous vehicle navigation.

[0006] In some embodiments of the present application, underground interference data is marked at corresponding positions on the underground map of the coal mine, including: Downhole interference data includes physical obstruction interference data, electromagnetic interference data, environmental dynamic interference data, and multipath effect interference data; Analyze the impact of physical occlusion interference data, electromagnetic interference data, environmental dynamic interference data, and multipath effect interference data on vehicle signals, and obtain occlusion interference index, electromagnetic interference index, environmental interference index, and multipath effect interference index; The physical occlusion interference data and occlusion interference indicators are marked on the coal mine underground map and represented by a fixed layer; Mark the electromagnetic interference data, environmental dynamic interference data and multipath effect interference data as well as electromagnetic interference indicators, environmental interference indicators and multipath effect interference indicators on the coal mine underground map and use dynamic layers to represent them; Among them, physical obstruction interference data is static interference, and electromagnetic interference data, environmental dynamic interference data and multipath effect interference data are all dynamic interference.

[0007] In some embodiments of the present application, the interference index is defined, including: Construct interference index curves of the electromagnetic interference index, the environmental interference index, and the multipath effect interference index that change over time according to their respective preset periods; For the electromagnetic interference index, the operating status of the electrical equipment in the electromagnetic interference data is time-aligned with the interference index curve of the electromagnetic interference index. The part of the interference index curve corresponding to the normal operating state of the electrical equipment is used as the initial electromagnetic interference index curve segment. The part with large changes on the initial electromagnetic interference index curve segment is removed by changing the slope to obtain the stable range of the electromagnetic interference index. For the environmental interference index, the concentration of environmental substances in the environmental dynamic interference data is time-aligned with the interference index curve of the environmental interference index. The part of the interference index curve corresponding to the stable state of the environmental substance concentration is used as the initial environmental interference index curve segment. The part with large changes on the initial environmental interference index curve segment is removed by changing the slope to obtain the stable range of the environmental interference index. For the multipath effect interference index, the interference index curve of the multipath effect interference index is screened by slope change to obtain the stable range of the multipath effect interference index; The interference index is defined by combining the shielding interference index, the stable interval of the electromagnetic interference index, the stable interval of the environmental interference index and the stable interval of the multipath effect interference index.

[0008] In some embodiments of the present application, the interference index is defined by combining the shielding interference index, the stable interval of the electromagnetic interference index, the stable interval of the environmental interference index, and the stable interval of the multipath effect interference index, including: Counting the median value and mode in the stable interval of the electromagnetic interference index, the stable interval of the environmental interference index, and the stable interval of the multipath effect interference index, and determining a representative value based on the median value and the mode; The interference index is calculated based on the representative values ​​of the shielding interference index, the electromagnetic interference index, the environmental interference index and the multipath effect interference index.

[0009] In some embodiments of the present application, the coal mine underground map is divided into multiple types of interference areas by using interference indicators, the density of marker points in each type of interference area is set, and multiple marker points are selected in different interference areas according to the density of marker points, including: Cluster analysis is performed on the interference indicators in regional units, and regions with similar interference indicators are regarded as the same type of interference areas. The density of landmark points is determined according to the range of interference indicators of each type of interference area. According to the marking point principle, all the spare marking points are confirmed on the coal mine underground map, and the spare marking points are screened according to the marking point density to establish a matching relationship between the interference area and the marking point.

[0010] In some embodiments of the present application, the vehicle position information is adjusted by the vehicle's recognition of landmarks, including: Identify the location of the landmark point, calculate the distance and relative position between the current vehicle and the landmark point, and obtain a first inferred position of the current vehicle; Obtaining a second inferred position of the current vehicle through vehicle positioning calculation; The confidence levels of the first and second inferred positions are assigned according to the category of the interference area in which the current vehicle is located. The first and second inferred positions of the current vehicle are fused based on the confidence levels to output the position information of the current vehicle.

[0011] In some embodiments of the present application, multi-model uncertainty motion prediction is performed on dynamic obstacles, including: A list of dynamic obstacle categories in coal mines is established, and the motion of each category of dynamic obstacle is predicted using the IMM model. The IMM model is a combination of the CV model and the CTRV model. Different weights are assigned to the CV model and the CTRV model according to the Markov chain. The IMM model is updated with vehicle radar observation data to obtain the vehicle collision risk, thereby achieving vehicle obstacle avoidance.

[0012] In some embodiments of the present application, determining whether a task path needs to be adjusted based on an obstruction event includes: Obstruction events include road blockage events and other road obstruction events; When the obstruction event is another road obstruction event, the current vehicle replans the mission path and thus re-navigates; When the obstruction event is a road congestion event, the congestion of the road congestion section is predicted, the path maintenance cost is calculated, and the travel path is replanned. The path change cost is calculated, and the path maintenance cost and path change cost are compared to select a path, thereby navigating the current vehicle.

[0013] Correspondingly, the present application also provides an autonomous navigation system for unmanned vehicles in coal mines, comprising: The first module is used to obtain road data and underground interference data in the coal mine, establish an underground coal mine map based on the road data, mark the underground interference data at the corresponding position on the underground coal mine map, and define interference indicators; The second module is used to divide the coal mine underground map into multiple interference areas based on interference indicators, set the density of marker points in each interference area, select multiple marker points in different interference areas based on the marker point density, and adjust the vehicle position information through vehicle recognition of the marker points; The third module is used to perform multi-model uncertainty motion prediction on dynamic obstacles while the vehicle is traveling along the mission path, thus achieving vehicle obstacle avoidance. The fourth module is used to detect obstruction events and determine whether the task path needs to be adjusted based on the obstruction events to achieve autonomous vehicle navigation.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. Define interference indicators, taking multiple interference factors into account to set interference standards. Using these indicators, the coal mine map is divided into multiple interference areas, thereby dividing the coal mine area. Targeted landmarks are set for each interference area, providing a reliable basis for subsequent vehicle location information. Vehicle location information is adjusted through vehicle recognition of landmarks, and auxiliary landmark recognition is used to accurately determine the vehicle's current location, ensuring the reliability of vehicle location.

[0015] 2. Multi-model uncertainty motion prediction is performed on dynamic obstacles. The need to adjust the mission path is determined based on the obstruction event, which improves the reliability and adaptability of autonomous navigation of unmanned vehicles in coal mines and meets the complex traffic and navigation needs of unmanned vehicles in coal mines. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of an autonomous navigation method for an unmanned vehicle in a coal mine proposed by the present invention; Figure 2 This is a structural schematic diagram of an autonomous navigation system for unmanned vehicles in coal mines proposed by the present invention. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0018] Reference Figure 1 , an autonomous navigation method for an unmanned vehicle in a coal mine, comprising the following steps: Step S101: Obtain road data and underground interference data in the coal mine, establish an underground coal mine map based on the road data, mark the underground interference data at corresponding positions on the underground coal mine map, and define interference indicators.

[0019] In this embodiment, the underground coal mine environment is complex and highly variable, subject to numerous interference factors, including physical obstructions, electromagnetic interference, environmental dynamics, and multipath effects. Positioning is the prerequisite and foundation for vehicle navigation, placing extremely high demands on the navigation and positioning of unmanned vehicles. To achieve safe and efficient underground transportation, core technologies such as high-precision positioning, sensor fusion, environmental perception, path planning, and intelligent control must be integrated. Road and interference data collection: Sensors such as lidar and cameras are used to collect underground road data and interference data, including physical obstructions, electromagnetic interference, environmental dynamics, and multipath effects. These interference conditions are combined to define an interference index.

[0020] In some embodiments of the present application, underground interference data is marked at corresponding positions on the underground map of the coal mine, including: Downhole interference data includes physical obstruction interference data, electromagnetic interference data, environmental dynamic interference data, and multipath effect interference data; Analyze the impact of physical occlusion interference data, electromagnetic interference data, environmental dynamic interference data, and multipath effect interference data on vehicle signals, and obtain occlusion interference index, electromagnetic interference index, environmental interference index, and multipath effect interference index; The physical occlusion interference data and occlusion interference indicators are marked on the coal mine underground map and represented by a fixed layer; Mark the electromagnetic interference data, environmental dynamic interference data and multipath effect interference data as well as electromagnetic interference indicators, environmental interference indicators and multipath effect interference indicators on the coal mine underground map and use dynamic layers to represent them; Among them, physical obstruction interference data is static interference, and electromagnetic interference data, environmental dynamic interference data and multipath effect interference data are all dynamic interference.

[0021] In this embodiment, the data types include: underground interference data including physical obstruction interference data (such as fixed obstacles formed by protrusions on the tunnel wall and equipment stacking), electromagnetic interference data (electromagnetic field fluctuations generated by the operation of electrical equipment), environmental dynamic interference data (environmental factors that affect signal transmission, such as changes in dust concentration and unstable ventilation airflow), and multipath effect interference data (interference paths formed by reflection and refraction of signals in complex underground structures).

[0022] Impact Analysis: We analyze the impact of various types of interference data on vehicle signals (positioning, communication, and so on). For example, physical obstructions can directly block signal transmission paths, causing signal weakening or loss; electromagnetic interference can introduce noise or distortion; dynamic environmental interference can alter signal propagation conditions over time; and multipath interference can cause uncertainty in signal phase and amplitude. This analysis yields indicators for obstruction interference, electromagnetic interference, environmental interference, and multipath interference.

[0023] Fixed layer annotation: Physical occlusion interference data and corresponding occlusion interference indicators are annotated on the coal mine map using a fixed layer. Because physical occlusion interference data is static, its location and impact range remain relatively stable over time. Fixed layers clearly display this static information, making it a convenient reference for vehicles during route planning and navigation.

[0024] Dynamic layer annotation: Electromagnetic interference data, environmental dynamic interference data, multipath interference data, and their respective interference indicators are displayed on the map using dynamic layers. This dynamic interference data changes over time, device operating status, environmental conditions, and other factors. Dynamic layers update this information in real time, allowing vehicles to obtain the latest interference information and make appropriate navigation adjustments.

[0025] Understandably, physical obstruction interference is relatively stable and rarely changes, so a fixed layer is used and directly participates in the subsequent interference index calculation. Electromagnetic interference data, environmental dynamic interference data, and multipath effect interference data are in a state of flux and instability, so they are annotated using dynamic layers. Stable and reliable values ​​need to be screened out to participate in the interference index calculation.

[0026] In some embodiments of the present application, the interference index is defined, including: Construct interference index curves of the electromagnetic interference index, the environmental interference index, and the multipath effect interference index that change over time according to their respective preset periods; For the electromagnetic interference index, the operating status of the electrical equipment in the electromagnetic interference data is time-aligned with the interference index curve of the electromagnetic interference index. The part of the interference index curve corresponding to the normal operating state of the electrical equipment is used as the initial electromagnetic interference index curve segment. The part with large changes on the initial electromagnetic interference index curve segment is removed by changing the slope to obtain the stable range of the electromagnetic interference index. For the environmental interference index, the concentration of environmental substances in the environmental dynamic interference data is time-aligned with the interference index curve of the environmental interference index. The part of the interference index curve corresponding to the stable state of the environmental substance concentration is used as the initial environmental interference index curve segment. The part with large changes on the initial environmental interference index curve segment is removed by changing the slope to obtain the stable range of the environmental interference index. For the multipath effect interference index, the interference index curve of the multipath effect interference index is screened by slope change to obtain the stable range of the multipath effect interference index; The interference index is defined by combining the shielding interference index, the stable interval of the electromagnetic interference index, the stable interval of the environmental interference index and the stable interval of the multipath effect interference index.

[0027] In this embodiment, the electromagnetic interference data, the environmental dynamic interference data, and the multipath effect interference data have corresponding periods of different times, and thus different interference curves that vary with time are constructed.

[0028] Electromagnetic interference index curve: According to the preset period (such as once every two hours), the electromagnetic interference index value is continuously monitored and recorded, and a curve of the electromagnetic interference index changing over time is drawn with time as the horizontal axis and the electromagnetic interference index value as the vertical axis.

[0029] Environmental interference index curve: Also according to the preset period, monitor the concentration of environmental objects (such as dust concentration, harmful gas concentration, etc.) and record the corresponding environmental interference index values, and draw a curve of the environmental interference index changing over time.

[0030] Multipath effect interference index curve: According to the preset period, monitor the multipath effect of the signal and record the multipath effect interference index value, and draw a curve of the multipath effect interference index changing with time.

[0031] Electromagnetic interference (EMI) is caused by underground electrical equipment. Therefore, the interference curve is initially screened based on the operating status of the electrical equipment. The operating status of the electrical equipment in the EMI data is time-aligned with the EMI index curve. For example, the time points at which the electrical equipment is turned on, off, and in normal operation are recorded and matched with the corresponding time points on the EMI index curve. The time period during which the electrical equipment is in normal operation is identified, and the corresponding EMI index curve segment within this time period is used as the initial EMI index curve segment. The slope of each point on the initial EMI index curve segment is calculated. By setting a slope threshold (e.g., a slope with an absolute value greater than 0.34 indicates significant change), sections with significant slope changes are removed. The remaining curve segment represents the stable range of the EMI index.

[0032] Environmental interference primarily refers to the concentration of environmental substances (such as dust concentration and harmful gas concentration) present in the underground environment that affect signal transmission. The concentration of environmental substances in the dynamic environmental interference data is time-aligned with the environmental interference index curve. The time points of environmental substance concentration monitoring are recorded and correspond one-to-one with the time points on the environmental interference index curve. The time period during which the concentration of environmental substances is stable is determined, and the corresponding environmental interference index curve segment within this time period is used as the initial environmental interference index curve segment. The slope of the initial environmental interference index curve segment is calculated, and using the set slope threshold (e.g., a slope with an absolute value greater than 0.4 indicates a significant change) as the standard, the portion with a significant slope change is eliminated to obtain the stable range of the environmental interference index.

[0033] The multipath effect is the influence of interference paths formed by the reflection and refraction of signals in complex underground structures. Therefore, the curve stability is directly analyzed, and the slope of the multipath effect interference index curve is directly calculated. It is screened according to the set slope threshold (for example, the absolute value of the slope is greater than 0.2, which indicates a large change), and the part with a large slope change is eliminated. The remaining curve segment is the stable range of the multipath effect interference index.

[0034] In some embodiments of the present application, the interference index is defined by combining the shielding interference index, the stable interval of the electromagnetic interference index, the stable interval of the environmental interference index, and the stable interval of the multipath effect interference index, including: Counting the median value and mode in the stable interval of the electromagnetic interference index, the stable interval of the environmental interference index, and the stable interval of the multipath effect interference index, and determining a representative value based on the median value and the mode; The interference index is calculated based on the representative values ​​of the shielding interference index, the electromagnetic interference index, the environmental interference index and the multipath effect interference index.

[0035] In this embodiment, all data points within the stable range of the electromagnetic interference indicator are collected, and these data points are sorted in ascending order.

[0036] Calculate the median: If the number of data points is odd, the median is the data point in the middle after sorting; if the number of data points is even, the median is the average of the two middle data points after sorting.

[0037] Determine the mode: Count the frequency of each data point. The data point with the highest frequency is the mode. If multiple data points have the same highest frequency, these are all modes. You can select one or perform the desired processing (e.g., taking the average) based on your needs.

[0038] For each stability interval (electromagnetic interference, environmental interference, and multipath interference), the median and mode are considered to determine the representative value. A weighted average approach can be used, for example, assigning different weights to the median and mode (weights can be set based on actual needs and the characteristics of each indicator), then calculating the weighted average as the representative value. Alternatively, depending on the specific situation, the median or mode value that best reflects the actual interference situation can be selected as the representative value.

[0039] The interference index calculation formula is as follows: ; in, is the interference indicator, 、 、 、 are the interference weights of the shielding interference index, electromagnetic interference index, environmental interference index and multipath effect interference index, respectively. To block interference indicators, 、 、 are the representative values ​​of electromagnetic interference index, environmental interference index and multipath effect interference index respectively, for 、 、 The maximum value in for 、 、 The minimum value in is the first constant, It represents the correction of the sum of the four types of interference indicators by the average value of the fluctuating interference indicator. Since the electromagnetic interference indicator, environmental interference indicator and multipath effect interference indicator are in change, correction is made accordingly.

[0040] In step S102, the coal mine underground map is divided into multiple types of interference areas based on the interference index, the density of marker points in each type of interference area is set, multiple marker points are selected in different interference areas according to the marker point density, and the vehicle position information is adjusted by the vehicle's recognition of the marker points.

[0041] In some embodiments of the present application, the coal mine underground map is divided into multiple types of interference areas by using interference indicators, the density of marker points in each type of interference area is set, and multiple marker points are selected in different interference areas according to the density of marker points, including: Cluster analysis is performed on the interference indicators in regional units, and regions with similar interference indicators are regarded as the same type of interference areas. The density of landmark points is determined according to the range of interference indicators of each type of interference area. According to the marking point principle, all the spare marking points are confirmed on the coal mine underground map, and the spare marking points are screened according to the marking point density to establish a matching relationship between the interference area and the marking point.

[0042] In this embodiment, the coal mine underground map is divided into multiple small area units, and each area unit corresponds to an interference index value (the interference index value can be calculated by comprehensively considering the electromagnetic interference, environmental interference, multipath effect interference and shielding interference mentioned above).

[0043] Clustering algorithms (such as the K-means clustering algorithm) are used to perform cluster analysis on these regional units. The input of the clustering algorithm is the interference index value of each regional unit. Through iterative calculation, regional units with similar interference indexes are grouped together to form the same type of interference area.

[0044] The corresponding marker density is set based on the interference index range for each type of interference area. The higher the interference index, the more severe the interference in the area, and the greater the difficulty in vehicle positioning and navigation. Therefore, a higher marker density is required to provide more positioning reference information. Conversely, the lower the interference index, the lower the marker density can be appropriately reduced. For example, areas with an interference index in the range of 0-0.3 (after normalization) can be set as low-interference areas, with a marker density of 1 marker per 100 square meters; areas with an interference index in the range of 0.3-0.6 can be set as medium-interference areas, with a marker density of 1 marker per 50 square meters; areas with an interference index above 0.6 can be set as high-interference areas, with a marker density of 1 marker per 20 square meters.

[0045] On the underground coal mine map, all possible backup landmarks are preliminarily identified based on map features (such as tunnel intersections and key equipment locations) and actual needs. These backup landmarks can be artificial markers with obvious features (such as reflective strips and QR codes) or natural features (such as special textures on tunnel walls and protruding rocks).

[0046] The backup markers are screened based on the set marker density. For each interference area, the required number of markers is calculated based on its area and marker density. Then, from the backup markers, a number that meets the required number is selected randomly or according to a specific rule (such as uniform distribution). These markers are retained as valid markers for the interference area, and the redundant backup markers are discarded.

[0047] The filtered landmarks are associated with the corresponding interference areas to establish a matching relationship between the interference area and the landmarks. The landmark information contained in each interference area can be recorded in a database or table, including the landmark's location coordinates and feature description, so that the vehicle can quickly and accurately identify and utilize these landmarks during driving.

[0048] In some embodiments of the present application, the vehicle position information is adjusted by the vehicle's recognition of landmarks, including: Identify the location of the landmark point, calculate the distance and relative position between the current vehicle and the landmark point, and obtain a first inferred position of the current vehicle; Obtaining a second inferred position of the current vehicle through vehicle positioning calculation; The confidence levels of the first and second inferred positions are assigned according to the category of the interference area in which the current vehicle is located. The first and second inferred positions of the current vehicle are fused based on the confidence levels to output the position information of the current vehicle.

[0049] In this embodiment, in the scenario of unmanned vehicles operating underground in coal mines, due to the complex environment and numerous interferences, a single position inference method (vehicle self-positioning) cannot guarantee accurate positioning. By fusing the first inferred position obtained by landmark recognition with the second inferred position calculated by vehicle positioning, the reliability of vehicle position information can be effectively improved.

[0050] Landmark point recognition and first inferred position calculation Landmark recognition: The vehicle uses onboard sensors (such as cameras and lidar) to scan its surroundings and identify pre-set landmarks through image processing and feature matching. For example, a reflective stripe landmark can be identified by detecting the specific color, shape, and brightness of the stripe; a QR code landmark can be identified by decoding the information contained in the QR code.

[0051] Distance and relative position calculation: After identifying the landmark location, the distance and relative position between the current vehicle and the landmark are calculated based on the distance information (such as the distance measured by the lidar) and angle information (such as the angle captured by the camera) measured by the vehicle's sensors. For example, if the coordinates of the landmark on the map are known to be (x1, y1), and the vehicle's sensors measure the distance d and angle θ (relative to the vehicle's front), the vehicle's coordinates relative to the landmark can be calculated as (dcosθ, dsinθ), thereby obtaining the vehicle's first inferred position in the map coordinate system (xinferred1, yinferred1).

[0052] Vehicle positioning calculation and second inferred position acquisition Vehicle Positioning: The vehicle calculates its position using its own positioning system (such as an inertial navigation system (INS) or GPS. In underground coal mines, the INS may primarily rely on a combination of other positioning aids). The INS measures the vehicle's acceleration and angular velocity, integrating them to determine the vehicle's position and attitude. This positioning result is then corrected using other positioning aids (such as odometers) to produce a second estimated position (x-inferred2, y-inferred2).

[0053] Confidence Assignment: The confidence levels of the first and second estimated positions are assigned based on the type of interference region the vehicle is currently located in. The interference region type can be determined using the interference index classification method described above. Generally speaking, the more severe the interference, the more likely the accuracy of both landmark recognition and vehicle positioning will be affected, so a reasonable confidence level distribution is necessary. For example, in low-interference regions, where both landmark recognition and vehicle positioning accuracy are high, the same confidence level (e.g., 0.5) can be assigned to both the first and second estimated positions. In medium-interference regions, where landmark recognition may be subject to some interference and vehicle positioning error may increase, the confidence level of the first estimated position for landmark recognition can be appropriately lowered (e.g., 0.4) while the confidence level of the second estimated position for vehicle positioning can be increased (e.g., 0.6). In high-interference regions, where both landmark recognition and vehicle positioning accuracy are low, the confidence level of the first estimated position can be further lowered (e.g., 0.3) while the confidence level of the second estimated position can be increased (e.g., 0.7).

[0054] Position fusion: The first inferred position and the second inferred position of the current vehicle are fused with the assigned confidence level. The fusion method can adopt the weighted average method.

[0055] In step S103 , the vehicle performs multi-model uncertainty motion prediction on dynamic obstacles while traveling along the mission path to achieve vehicle obstacle avoidance.

[0056] In some embodiments of the present application, multi-model uncertainty motion prediction is performed on dynamic obstacles, including: A list of dynamic obstacle categories in coal mines is established, and the motion of each category of dynamic obstacle is predicted using the IMM model. The IMM model is a combination of the CV model and the CTRV model. Different weights are assigned to the CV model and the CTRV model according to the Markov chain. The IMM model is updated with vehicle radar observation data to obtain the vehicle collision risk, thereby achieving vehicle obstacle avoidance.

[0057] In this example, dynamic obstacles that may appear underground in a coal mine are categorized into various types, such as personnel, transport vehicles (e.g., mine cars), and small equipment (e.g., mobile drilling rigs). This creates a list of dynamic obstacle categories. Different types of dynamic obstacles exhibit distinct motion characteristics and behavior patterns, and categorization helps more accurately predict their movement.

[0058] Model combination: The IMM model is a combination of the CV model and the CTRV model. The CV model assumes that dynamic obstacles move in a straight line at a constant speed and is suitable for obstacles with relatively stable motion. The CTRV model assumes that dynamic obstacles move at a constant rotation rate and speed and can describe curved motions such as turns.

[0059] Weight assignment: Based on a Markov chain (pre-selecting and defining multiple movement states of a dynamic obstacle and predicting the state of the dynamic obstacle at the next moment through state transition probabilities), different initial weights are assigned to the CV model and the CTRV model. The initial weight setting can be based on prior knowledge of the common movement patterns of different types of dynamic obstacles. For example, for people, since they are more flexible in movement and turn frequently, the CTRV model can be assigned a higher initial weight (such as 0.7) and the CV model can be assigned a lower initial weight (such as 0.3). For transport vehicles, since their movement is relatively stable when traveling in straight lanes, the CV model can be assigned a higher initial weight (such as 0.6) and the CTRV model can be assigned a lower initial weight (such as 0.4).

[0060] Model prediction: The CV model and CTRV model are used to predict the motion of each type of dynamic obstacle. The CV model predicts the next state based on the current state (position, velocity); the CTRV model predicts the next state based on the current state (position, velocity, rotation rate).

[0061] Observation Update: The IMM model is updated with model probabilities based on vehicle radar observation data. Radar captures real-time information such as the position and velocity of dynamic obstacles. This observation data is compared with the predictions from the CV and CTRV models, and the degree of match between each model and the observations is calculated. Based on this match, the weights of the CV and CTRV models are updated using the Bayesian formula. For example, if the CTRV model's predictions better match the radar observations, its weight is increased, while the weight of the CV model is decreased.

[0062] Collision Risk Calculation: Based on the updated model weights and prediction results, the collision risk between the vehicle and the dynamic obstacle is calculated. The predicted position, speed, and direction of movement of the dynamic obstacle, as well as the vehicle's own motion state, are considered to determine the likelihood of a collision between the vehicle and the dynamic obstacle within a certain period of time. For example, a collision risk threshold can be set; when the calculated collision risk exceeds this threshold, a collision risk is considered present.

[0063] Obstacle Avoidance Decision-Making: If a collision risk exists, the vehicle will develop an appropriate obstacle avoidance strategy based on the magnitude and direction of the collision risk. This strategy may include deceleration, stopping, or changing direction. For example, if the collision risk primarily originates from the left side of the vehicle, the vehicle may adjust its direction to the right to avoid a dynamic obstacle.

[0064] In step S104 , at the same time, an obstruction event is detected, and it is determined whether the mission path needs to be adjusted according to the obstruction event to achieve autonomous vehicle navigation.

[0065] In some embodiments of the present application, determining whether a task path needs to be adjusted based on an obstruction event includes: Obstruction events include road blockage events and other road obstruction events; When the obstruction event is another road obstruction event, the current vehicle replans the mission path and thus re-navigates; When the obstruction event is a road congestion event, the congestion of the road congestion section is predicted, the path maintenance cost is calculated, and the travel path is replanned. The path change cost is calculated, and the path maintenance cost and path change cost are compared to select a path, thereby navigating the current vehicle.

[0066] In this embodiment, the obstruction event is an emergency event that affects the vehicle's travel and the path change.

[0067] Detection: Vehicle-mounted sensors (such as cameras, LiDAR, and ultrasonic sensors) monitor the vehicle's surroundings in real time to detect any obstacles that could hinder normal movement. For example, cameras can identify large obstacles or crowds ahead; LiDAR can measure the distance to surrounding objects to determine whether the road is blocked. Alternatively, this determination can be made based on the position, speed, and distance of all vehicles.

[0068] Classification: Detected obstruction events are classified into road congestion events and other road obstruction events. Road congestion events typically occur when vehicles or equipment occupy the road, preventing vehicles from passing normally. Other road obstruction events may include road collapse, water accumulation, construction areas, etc.

[0069] When an obstruction occurs due to other road obstructions, the vehicle must re-route the original mission path, as such events often make the original mission path unavailable. The vehicle utilizes built-in navigation algorithms and map information, combined with real-time sensor data, to search for a feasible path that avoids the obstruction and re-navigate along the newly planned route.

[0070] Congestion Prediction: When the impeding event is a road congestion, the system analyzes historical data, current vehicle sensor data, and communication information with other vehicles (if a connected vehicle system is available) to predict the congestion situation in the congested section of the road. For example, the system calculates the number of vehicles in the congested area and their speed to determine the severity and duration of the congestion.

[0071] Path maintenance cost calculation and correction: Basic Calculation: Route maintenance cost refers to the cost of continuing along the original mission path. Factors that may be considered include travel time (calculated based on current speed and remaining distance) and energy consumption (estimated based on the vehicle's powertrain and driving conditions).

[0072] Additional factors introduced to correct: Risk Factor: Considering the potential risks of driving on congested roads, such as the increased risk of collisions with surrounding vehicles, a risk factor can be introduced to adjust the path maintenance cost. The risk factor is determined based on factors such as congestion level and vehicle density. The more severe the congestion and the greater the vehicle density, the higher the risk factor and the correspondingly higher the path maintenance cost.

[0073] Time window factor: If a vehicle has strict time constraints on its mission, it is necessary to consider the potential delays caused by traveling on congested roads. A time window weight can be introduced. When traveling on congested roads is expected to cause the mission to exceed the specified time window, the time window weight is increased, and the path maintenance cost increases.

[0074] The revised path maintenance cost calculation formula is as follows: ; in, For the The vehicle in The corrected path maintenance cost of the congested paths (congestion events), For the The vehicle in The basic path maintenance cost of the congested paths (congestion events), 、 are the impact weights of risk factors and time factors respectively, 、 are the standardized impact values ​​of risk factors and time factors, For the The vehicle in The second constant of the congested path, It represents the modification of the basic path maintenance cost by risk factors and time factors.

[0075] Path Change Cost Calculation: Path change cost is the cost of rerouting a vehicle. Factors considered include the length of the reroute, the number of turns, and the impact of slope changes on driving difficulty and energy consumption.

[0076] Path selection: Compare the path maintenance cost and path change cost and select the path with the lower cost as the vehicle's route. If the path maintenance cost is lower than the path change cost, the vehicle continues along the original mission path; if the path change cost is lower than the path maintenance cost, the vehicle follows the re-planned route.

[0077] Correspondingly, this application also provides an autonomous navigation system for unmanned vehicles in coal mines, such as Figure 2 Shown, including, The first module is used to obtain road data and underground interference data in the coal mine, establish an underground coal mine map based on the road data, mark the underground interference data at the corresponding position on the underground coal mine map, and define interference indicators; The second module is used to divide the coal mine underground map into multiple interference areas based on interference indicators, set the density of marker points in each interference area, select multiple marker points in different interference areas based on the marker point density, and adjust the vehicle position information through vehicle recognition of the marker points; The third module is used to perform multi-model uncertainty motion prediction on dynamic obstacles while the vehicle is traveling along the mission path, thus achieving vehicle obstacle avoidance. The fourth module is used to detect obstruction events and determine whether the task path needs to be adjusted based on the obstruction events to achieve autonomous vehicle navigation.

[0078] Compared with the prior art, the present invention has the following beneficial effects: 1. Define interference indicators, taking multiple interference factors into account to set interference standards. Using these indicators, the coal mine map is divided into multiple interference areas, thereby dividing the coal mine area. Targeted landmarks are set for each interference area, providing a reliable basis for subsequent vehicle location information. Vehicle location information is adjusted through vehicle recognition of landmarks, and auxiliary landmark recognition is used to accurately determine the vehicle's current location, ensuring the reliability of vehicle location.

[0079] 2. Multi-model uncertainty motion prediction is performed on dynamic obstacles. The need to adjust the mission path is determined based on the obstruction event, which improves the reliability and adaptability of autonomous navigation of unmanned vehicles in coal mines and meets the complex traffic and navigation needs of unmanned vehicles in coal mines.

[0080] Through the above description of the embodiments, those skilled in the art will clearly understand that the present invention can be implemented via hardware or via software combined with a necessary general-purpose hardware platform. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product. This software product can be stored on a non-volatile storage medium (such as a CD-ROM, USB flash drive, or external hard drive) and includes instructions for enabling a computer device (such as a personal computer, server, or network device) to execute the methods described in various implementation scenarios of the present invention.

[0081] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0082] Those skilled in the art will appreciate that the modules in the system of the implementation scenario can be distributed in the system of the implementation scenario according to the implementation scenario description, or can be modified accordingly and located in one or more systems different from the implementation scenario. The modules of the above implementation scenario can be combined into one module or further divided into multiple submodules.

[0083] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. An autonomous navigation method for an unmanned vehicle in a coal mine, characterized in that: include, Obtain road data and underground interference data in the coal mine, establish an underground coal mine map based on the road data, mark the underground interference data at the corresponding position on the underground coal mine map, and define interference indicators; The coal mine map is divided into multiple interference areas based on interference indicators. The density of marker points in each interference area is set. Multiple marker points are selected in different interference areas based on the density of marker points. The vehicle position information is adjusted by the vehicle's recognition of the marker points. While the vehicle is traveling along the mission path, it performs multi-model uncertainty motion prediction on dynamic obstacles to achieve vehicle obstacle avoidance. At the same time, it detects obstruction events and determines whether the mission path needs to be adjusted based on the obstruction events to achieve autonomous vehicle navigation.

2. The autonomous navigation method for an unmanned vehicle in a coal mine according to claim 1, characterized in that: Mark the underground interference data at the corresponding position on the coal mine underground map, including: Downhole interference data includes physical obstruction interference data, electromagnetic interference data, environmental dynamic interference data, and multipath effect interference data; Analyze the impact of physical occlusion interference data, electromagnetic interference data, environmental dynamic interference data, and multipath effect interference data on vehicle signals, and obtain occlusion interference index, electromagnetic interference index, environmental interference index, and multipath effect interference index; The physical occlusion interference data and occlusion interference indicators are marked on the coal mine underground map and represented by a fixed layer; Mark the electromagnetic interference data, environmental dynamic interference data and multipath effect interference data as well as electromagnetic interference indicators, environmental interference indicators and multipath effect interference indicators on the coal mine underground map and use dynamic layers to represent them; Among them, physical obstruction interference data is static interference, and electromagnetic interference data, environmental dynamic interference data and multipath effect interference data are all dynamic interference.

3. The autonomous navigation method for an unmanned vehicle in a coal mine according to claim 2, characterized in that: Define interference indicators, including, Construct interference index curves of the electromagnetic interference index, the environmental interference index, and the multipath effect interference index that change over time according to their respective preset periods; For the electromagnetic interference index, the operating status of the electrical equipment in the electromagnetic interference data is time-aligned with the interference index curve of the electromagnetic interference index. The part of the interference index curve corresponding to the normal operating state of the electrical equipment is used as the initial electromagnetic interference index curve segment. The part with large changes on the initial electromagnetic interference index curve segment is removed by changing the slope to obtain the stable range of the electromagnetic interference index. For the environmental interference index, the concentration of environmental substances in the environmental dynamic interference data is time-aligned with the interference index curve of the environmental interference index. The part of the interference index curve corresponding to the stable state of the environmental substance concentration is used as the initial environmental interference index curve segment. The part with large changes on the initial environmental interference index curve segment is removed by changing the slope to obtain the stable range of the environmental interference index. For the multipath effect interference index, the interference index curve of the multipath effect interference index is screened by slope change to obtain the stable range of the multipath effect interference index; The interference index is defined by combining the shielding interference index, the stable interval of the electromagnetic interference index, the stable interval of the environmental interference index and the stable interval of the multipath effect interference index.

4. The autonomous navigation method for an unmanned vehicle in an underground coal mine according to claim 3, characterized in that: The interference index is defined by combining the shielding interference index, the stable interval of the electromagnetic interference index, the stable interval of the environmental interference index and the stable interval of the multipath effect interference index, including: Counting the median value and mode in the stable interval of the electromagnetic interference index, the stable interval of the environmental interference index, and the stable interval of the multipath effect interference index, and determining a representative value based on the median value and the mode; The interference index is calculated based on the representative values ​​of the shielding interference index, the electromagnetic interference index, the environmental interference index and the multipath effect interference index.

5. The autonomous navigation method for an unmanned vehicle in an underground coal mine according to claim 1, characterized in that: The coal mine underground map is divided into multiple interference areas based on the interference index, and the density of marker points in each interference area is set. According to the density of marker points, multiple marker points are selected in different interference areas, including: Cluster analysis is performed on the interference indicators in regional units, and regions with similar interference indicators are regarded as the same type of interference areas. The density of landmark points is determined according to the range of interference indicators of each type of interference area. According to the marking point principle, all the spare marking points are confirmed on the coal mine underground map, and the spare marking points are screened according to the marking point density to establish a matching relationship between the interference area and the marking point.

6. The autonomous navigation method for an unmanned vehicle in an underground coal mine according to claim 1, characterized in that: Adjust the vehicle's position information by identifying the landmarks. include, Identify the location of the landmark point, calculate the distance and relative position between the current vehicle and the landmark point, and obtain a first inferred position of the current vehicle; Obtaining a second inferred position of the current vehicle through vehicle positioning calculation; The confidence levels of the first and second inferred positions are assigned according to the category of the interference area in which the current vehicle is located. The first and second inferred positions of the current vehicle are fused based on the confidence levels to output the position information of the current vehicle.

7. The autonomous navigation method for an unmanned vehicle in an underground coal mine according to claim 1, characterized in that: Multi-model uncertainty motion prediction for dynamic obstacles, including, A list of dynamic obstacle categories in coal mines is established, and the motion of each category of dynamic obstacle is predicted using the IMM model. The IMM model is a combination of the CV model and the CTRV model. Different weights are assigned to the CV model and the CTRV model according to the Markov chain. The IMM model is updated with vehicle radar observation data to obtain the vehicle collision risk, thereby achieving vehicle obstacle avoidance.

8. The autonomous navigation method for an unmanned vehicle in an underground coal mine according to claim 1, characterized in that: Determine whether the task path needs to be adjusted based on the blocking event, including: Obstruction events include road blockage events and other road obstruction events; When the obstruction event is another road obstruction event, the current vehicle replans the mission path and thus re-navigates; When the obstruction event is a road congestion event, the congestion of the road congestion section is predicted, the path maintenance cost is calculated, and the travel path is replanned. The path change cost is calculated, and the path maintenance cost and path change cost are compared to select a path, thereby navigating the current vehicle.

9. An autonomous navigation system for unmanned vehicles in coal mines, characterized in that: include, The first module is used to obtain road data and underground interference data in the coal mine, establish an underground coal mine map based on the road data, mark the underground interference data at the corresponding position on the underground coal mine map, and define interference indicators; The second module is used to divide the coal mine underground map into multiple interference areas based on interference indicators, set the density of marker points in each interference area, select multiple marker points in different interference areas based on the marker point density, and adjust the vehicle position information through vehicle recognition of the marker points; The third module is used to perform multi-model uncertainty motion prediction on dynamic obstacles while the vehicle is traveling along the mission path, thus achieving vehicle obstacle avoidance. The fourth module is used to detect obstruction events and determine whether the task path needs to be adjusted based on the obstruction events to achieve autonomous vehicle navigation.

Images