Multi-agent way giving method and device based on right of way and unmanned vehicle

Abstract

The invention relates to the technical field of unmanned driving, and discloses a multi-agent way giving method and device based on right of way and an unmanned vehicle. The method comprises the following steps: according to a first prediction track and a second prediction track of a first unmanned vehicle, carrying out collision zone analysis on the first unmanned vehicle and a second unmanned vehicle to obtain a collision zone analysis result; when the collision zone analysis result is that the collision zone exists, determining a first position relation between the first unmanned vehicle and the collision zone and a second position relation between the second unmanned vehicle and the collision zone; according to the first position relation and the second position relation, action decision analysis is conducted on the first unmanned vehicle, a target action decision of the first unmanned vehicle is obtained, and the first unmanned vehicle is controlled to execute actions according to the target action decision. According to the invention, the problem of trajectory conflict of the vehicle in the open area and the intersection area can be solved, so that the safety is ensured and the operation efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of unmanned driving technology, and in particular to a multi-agent right-of-way-based method, device and unmanned vehicle. Background Art

[0002] The unmanned driving scenario in mines is a complex multi-agent interactive environment, in which safety and efficiency are two core goals. In this scenario, trajectory conflicts will inevitably occur between multiple autonomous driving vehicles, which requires the vehicle's decision-making module to determine the order of passage based on the passage priority between vehicles while ensuring safety, so as to achieve safe and efficient collaborative operation. At present, there are two main methods for dealing with multi-agent trajectory conflicts: centralized decision-making based on servers and networks, and methods based on map lane occupancy. Centralized decision-making relies on cloud servers to arbitrate the scene and detect and resolve conflicts by receiving information from all vehicles. The method based on map lane occupancy is to send the map route information occupied by the vehicle to surrounding vehicles after completing the global planning to avoid conflicts.

[0003] However, these existing technologies have obvious limitations. Centralized decision-making methods are heavily dependent on networks and servers, and may fail in areas with poor network coverage. There is also a large decision lag, which affects the real-time nature of vehicle decision-making. Map lane occupancy-based methods are overly dependent on pre-built map road networks, greatly limiting the flexibility and freedom of vehicle driving. These deficiencies make it difficult for existing methods to meet the needs of diverse scenarios in mining environments, and are unable to fully ensure safety while taking into account operational efficiency. Summary of the invention

[0004] The present application provides a multi-agent giving way method, device and unmanned vehicle based on road rights, which solves the trajectory conflict problem of multiple vehicles in open areas and intersections through real-time processing of vehicle driving status information, collision risk assessment, road right priority judgment and dynamic decision analysis, thereby ensuring safety and improving operating efficiency.

[0005] In the first aspect, the present application provides a multi-agent giving way method based on road right, and the multi-agent giving way method based on road right includes: obtaining driving status information of a second unmanned vehicle; wherein the driving status information at least includes a second predicted trajectory of the second unmanned vehicle; based on the first predicted trajectory of the first unmanned vehicle and the second predicted trajectory, performing a collision zone analysis on the first unmanned vehicle and the second unmanned vehicle to obtain a collision zone analysis result; when the collision zone analysis result is that a collision zone exists, determining a first position relationship between the first unmanned vehicle and the collision zone and a second position relationship between the second unmanned vehicle and the collision zone respectively; based on the first position relationship and the second position relationship, performing an action decision analysis on the first unmanned vehicle to obtain a target action decision of the first unmanned vehicle, and controlling the first unmanned vehicle to perform an action according to the target action decision.

[0006] In a second aspect, the present application provides an unmanned vehicle control device, comprising:

[0007] An acquisition module, configured to acquire driving state information of the second unmanned vehicle; wherein the driving state information at least includes a second predicted trajectory of the second unmanned vehicle;

[0008] A collision zone analysis module, configured to perform a collision zone analysis on the first unmanned vehicle and the second unmanned vehicle according to the first predicted trajectory of the first unmanned vehicle and the second predicted trajectory, and obtain a collision zone analysis result;

[0009] a position relationship determination module, configured to respectively determine a first position relationship between the first unmanned vehicle and the collision zone and a second position relationship between the second unmanned vehicle and the collision zone when the collision zone analysis result indicates that the collision zone exists;

[0010] A decision-making module is used to perform an action decision analysis on the first unmanned vehicle according to the first position relationship and the second position relationship, obtain a target action decision of the first unmanned vehicle, and control the first unmanned vehicle to perform an action according to the target action decision.

[0011] In a third aspect, the present application provides an unmanned vehicle, comprising: any of the unmanned vehicle control devices described above.

[0012] The fourth aspect of the present application provides a multi-agent giving way device based on road right, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory to enable the multi-agent giving way device based on road right to execute the above-mentioned multi-agent giving way method based on road right.

[0013] A fifth aspect of the present application provides a computer-readable storage medium, which stores instructions that, when executed on a computer, enable the computer to execute the above-mentioned multi-agent yielding method based on road rights.

[0014] In the technical solution provided by the present application, by obtaining the driving state information of the second unmanned vehicle and combining the first predicted trajectory of the first unmanned vehicle to perform collision zone analysis, the potential conflict area can be accurately identified, thereby improving the ability to predict dangerous situations. By determining the positional relationship between the two vehicles and the collision zone, the solution realizes the accurate modeling of complex traffic scenes and provides a reliable data basis for subsequent decision-making. The action decision analysis based on these positional relationships can make the target action decision that best suits the current situation for the first unmanned vehicle, which not only ensures safety but also takes into account operational efficiency. This dynamic decision-making mechanism based on real-time data greatly improves the system's ability to cope with complex and changing mining environments. At the same time, since the decision-making process is based on the vehicle's own perception and calculation, it reduces dependence on external networks and central servers, and improves the robustness and real-time performance of the system. Through accurate collision prediction, positional relationship analysis and intelligent decision-making, the safe and efficient coordinated operation of unmanned vehicles in mines is achieved, significantly improving the overall transportation efficiency and safety level. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0016] Figure 1 A schematic diagram of an embodiment of a multi-agent yielding method based on road rights in an embodiment of the present application;

[0017] Figure 2 A schematic diagram of the decision-making process in an embodiment of the present application;

[0018] Figure 3 This is a schematic diagram of an embodiment of a multi-agent right-of-way-based right-of-way device in an embodiment of the present application. DETAILED DESCRIPTION

[0019] Embodiments of the present application provide a multi-agent giving way method, device and unmanned vehicle based on road rights. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0020] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the multi-agent yielding method based on road rights includes:

[0021] Step S101, obtaining driving state information of a second unmanned vehicle; wherein the driving state information at least includes a second predicted trajectory of the second unmanned vehicle;

[0022] Step S102: performing collision zone analysis on the first unmanned vehicle and the second unmanned vehicle according to the first predicted trajectory and the second predicted trajectory of the first unmanned vehicle to obtain a collision zone analysis result;

[0023] Step S103: when the collision zone analysis result indicates that a collision zone exists, determining a first position relationship between the first unmanned vehicle and the collision zone and a second position relationship between the second unmanned vehicle and the collision zone respectively;

[0024] Step S104: perform action decision analysis on the first unmanned vehicle according to the first position relationship and the second position relationship, obtain a target action decision of the first unmanned vehicle, and control the first unmanned vehicle to perform an action according to the target action decision.

[0025] It is understandable that the executor of the present application may be an unmanned vehicle or a roadside device, etc., and the specifics are not limited here.

[0026] Specifically, in the unmanned driving scenario of mines, the key to achieving safe and efficient coordinated operation of multiple vehicles lies in accurately obtaining vehicle information, analyzing potential collision risks, and making reasonable action decisions. First, obtain the driving status information of the second unmanned vehicle, which includes at least the second predicted trajectory of the second unmanned vehicle. The predicted trajectory is based on the vehicle's current position, speed, acceleration and other parameters, combined with road conditions and mission objectives, and is generated by the trajectory prediction algorithm to predict the path that the vehicle may travel in the future. After obtaining the driving status information, the collision zone of the two vehicles is analyzed according to the first predicted trajectory of the first unmanned vehicle and the second predicted trajectory of the second unmanned vehicle. The collision zone analysis uses a trajectory intersection detection algorithm to determine whether there is a potential collision risk by comparing the intersection of the two predicted trajectories in the time and space dimensions. If the analysis results show that there is a collision zone, the first position relationship between the first unmanned vehicle and the collision zone, and the second position relationship between the second unmanned vehicle and the collision zone are further determined.

[0027] The determination of the positional relationship is achieved by calculating parameters such as the distance from the vehicle to the collision zone and the time to entry (TTE). The calculation of TTE takes into account the current speed of the vehicle and the distance from the front of the collision zone. Based on this information, the system can evaluate the risk level of each vehicle. The risk level is divided into three categories: Dangerous when TTE is less than or equal to 4.0 seconds, Attentive when TTE is between 4.0 and 7.0 seconds, and Safe when TTE is greater than 7.0 seconds. Based on these positional relationships and risk assessment results, the first unmanned vehicle is analyzed for action decisions. The decision-making process takes into account multiple factors, including the vehicle's safety level, business priority, and relative position relationship with other vehicles. A series of possible action options are generated, such as no restriction, starting to pass, maintaining the current speed, deceleration, stopping at the stop line, emergency stop, or virtual following. Finally, an optimal target action decision is selected, and the first unmanned vehicle is controlled to perform the action according to the decision. This process is dynamic, continuously monitoring vehicle status and environmental changes, and adjusting decisions in real time to ensure safety and efficiency.

[0028] It should be noted that the calculation formula for Time To Entry (TTE) is as follows:

[0029]

[0030] This is the calculated time to enter the collision zone (Time To Entry, TTE)

[0031] This is the current time t, the distance between other car k and collision area c k Indicates how far the other car is from the collision area. This is the current time t, his car k is heading towards the collision area c k This value indicates how fast the car is moving towards the collision zone.

[0032] (1) When the speed and distance When both are greater than 0: Indicates how long it takes for the vehicle to reach the collision area at its current speed.

[0033] (2) When the distance When less than or equal to 0: Indicates that the vehicle has reached or passed through the collision zone;

[0034] (3) In any other case: Indicates that the vehicle will not reach the collision area (for example, when the speed is 0, the vehicle stops and will not enter the collision area again).

[0035] For example, suppose there is a car k, which is 100 meters away from the collision zone c at time t, and is moving towards the zone at a speed of 20 meters per second. distance According to the formula: According to the risk level judgment, the vehicle risk level of this vehicle is Attentive.

[0036] For example, suppose that on a mining area transportation channel, the first unmanned vehicle A is traveling at a speed of 20 km / h, and its trajectory is predicted to reach an intersection in 30 seconds. At the same time, the second unmanned vehicle B is approaching the same intersection from the side at a speed of 15 km / h and is expected to arrive in 35 seconds. After the system performs a collision zone analysis, it is found that the trajectories of the two vehicles intersect at the intersection, forming a potential collision zone.

[0037] Through calculation, it is found that the TTE of car A to the collision zone is 30 seconds, which is within the safety level; the TTE of car B is 35 seconds, which is also within the safety level. However, considering that car A is carrying emergency supplies and has a higher business priority, the system has formulated a target action decision of "maintaining the current speed" for car A. At the same time, a deceleration instruction is sent to car B, requiring it to slow down to 5km / h before the virtual stop line 10 meters away from the intersection to ensure that car A passes the intersection safely. In this way, the driving of the two unmanned vehicles was successfully coordinated, which not only ensured safety, but also took into account efficiency and business needs.

[0038] In the embodiment of the present application, by obtaining the driving state information of the second unmanned vehicle and combining the first predicted trajectory of the first unmanned vehicle to perform collision zone analysis, the potential conflict area can be accurately identified, thereby improving the ability to predict dangerous situations. By determining the positional relationship between the two vehicles and the collision zone, the scheme realizes the accurate modeling of complex traffic scenes and provides a reliable data basis for subsequent decision-making. The action decision analysis based on these positional relationships can make the target action decision that best suits the current situation for the first unmanned vehicle, which not only ensures safety but also takes into account operational efficiency. This dynamic decision-making mechanism based on real-time data greatly improves the system's ability to cope with complex and changing mining environments. At the same time, since the decision-making process is based on the vehicle's own perception and calculation, it reduces dependence on external networks and central servers, and improves the robustness and real-time performance of the system. Through accurate collision prediction, positional relationship analysis and intelligent decision-making, the safe and efficient coordinated operation of unmanned vehicles in mines is achieved, significantly improving the overall transportation efficiency and safety level.

[0039] In a specific embodiment, according to the first position relationship and the second position relationship, an action decision analysis is performed on the first unmanned vehicle to obtain a target action decision of the first unmanned vehicle, including:

[0040] When the first position relationship indicates that the first unmanned vehicle has not yet reached the collision zone, and the second position relationship indicates that the second unmanned vehicle has not yet reached the collision zone, an action decision analysis is performed on the first unmanned vehicle according to the risk levels of the first unmanned vehicle and the second unmanned vehicle to obtain a target action decision for the first unmanned vehicle.

[0041] Specifically, in the coordinated operation of unmanned vehicles in mines, when the first unmanned vehicle and the second unmanned vehicle have not yet reached the collision zone, the action decision analysis of the first unmanned vehicle is performed. This process first requires determining the positional relationship and risk level of the two vehicles, and then formulating the optimal action strategy based on this information. The determination of the positional relationship is achieved by calculating the distance from the vehicle to the collision zone and the estimated arrival time (Time To Entry, TTE). The calculation of TTE takes into account the current speed of the vehicle and the distance from the front of the collision zone. The risk level is divided according to the value of TTE: when TTE is less than or equal to 4.0 seconds, it is dangerous (Dangerous), TTE is between 4.0 and 7.0 seconds, it needs attention (Attentive), and TTE is greater than 7.0 seconds. Safe (Safe).

[0042] When it is determined that neither vehicle has reached the collision zone, the action decision analysis will take into account multiple factors, including the risk level of the two vehicles, business priorities, vehicle performance parameters, and road environment. The decision process uses a multi-level decision tree algorithm combined with a fuzzy logic control method to achieve accurate action decisions. First, a preliminary judgment is made based on the risk level of the two vehicles. If both vehicles are at a safe level, the first unmanned vehicle may be decided to maintain the current speed. If one of the vehicles is at a level that requires attention or danger, further analysis is required. Next, business priorities are considered. For example, if the first unmanned vehicle is carrying emergency supplies or performing a high-priority task, it may be granted the right of way even if its risk level is high. At this point, the decision may be to maintain or slightly adjust the current speed while sending a deceleration or yield instruction to the second unmanned vehicle. Vehicle performance parameters, such as maximum acceleration, braking performance, etc., are also taken into account in the decision. These parameters affect the maneuverability of the vehicle, which in turn affects the calculation of the safety margin. For example, if the braking performance of the first unmanned vehicle is better than that of the second unmanned vehicle, the first unmanned vehicle may be granted a higher right of way if other conditions are the same.

[0043] Road environmental factors, such as slope and friction coefficient, also affect decision-making. In mining environments with complex terrain, these factors have a particularly significant impact on vehicle behavior. The decision-making algorithm dynamically adjusts safety thresholds and behavior strategies based on real-time collected environmental data. The output of the action decision includes but is not limited to the following options: maintain the current speed, accelerate through, decelerate through, temporarily stop and wait, change the driving route, etc. Each decision is accompanied by specific execution parameters, such as target speed, acceleration or deceleration values, etc.

[0044] For example, suppose that on a mining area transportation channel, the first unmanned vehicle A is traveling at a speed of 25 km / h, 150 meters away from the potential collision zone. At the same time, the second unmanned vehicle B is approaching the same collision zone from the side at a speed of 20 km / h, 120 meters away. By calculation, the TTE of vehicle A is about 21.6 seconds, and the TTE of vehicle B is 21.6 seconds. Both vehicles are at a safe level. However, considering that vehicle A is performing an emergency material transportation mission, it has a higher business priority. At the same time, the road information shows that there is a downhill section 150 meters ahead of vehicle A, which may increase the speed of the vehicle. Based on these factors, the decision algorithm formulates a target action decision for vehicle A to "maintain the current speed and slow down slightly before entering the downhill." The specific parameters are: maintain a speed of 25 km / h on the flat road section, and start to decelerate evenly to 22 km / h 100 meters away from the downhill to ensure safe passage through the collision zone. At the same time, a deceleration instruction is sent to vehicle B, requiring it to start decelerating to 15 km / h 100 meters away from the collision zone to give way to vehicle A. This decision not only ensured the smooth execution of high-priority tasks, but also took into account road environment factors and ensured the safety of the two vehicles. This not only improved the efficiency of mine transportation, but also greatly reduced the risk of safety accidents.

[0045] In a specific embodiment, determining a target action decision of the first unmanned vehicle according to the risk levels of the first unmanned vehicle and the second unmanned vehicle includes:

[0046] (1) determining a first risk level of the first unmanned vehicle according to a first arrival time of the first unmanned vehicle at the collision zone and a preset risk level classification rule;

[0047] (2) determining a second risk level of the second unmanned vehicle according to a second arrival time of the second unmanned vehicle at the collision zone and a preset risk level classification rule;

[0048] (3) Compare the first risk level with the second risk level, and determine a target action decision for the first unmanned vehicle based on the comparison result.

[0049] Specifically, the first risk level is compared with the second risk level, and a target action decision of the first unmanned vehicle is determined according to the comparison result, including:

[0050] (1) If the risk level of the first unmanned vehicle is lower than that of the second unmanned vehicle, control the first unmanned vehicle to maintain the current driving state; and / or

[0051] (2) if the risk level of the first unmanned vehicle is higher than that of the second unmanned vehicle, controlling the first unmanned vehicle to stop at the stop line; and / or

[0052] (3) If the risk level of the first unmanned vehicle is equal to that of the second unmanned vehicle, the target action decision of the first unmanned vehicle is determined according to the right-of-way priority of the first unmanned vehicle and the second unmanned vehicle.

[0053] Specifically, in the coordinated operation of unmanned vehicles in mines, it is a key process to determine the target action decision of the first unmanned vehicle based on the risk level of the first unmanned vehicle and the second unmanned vehicle. This process first requires calculating the risk level of the two vehicles separately, and then formulating the optimal action strategy based on the comparison results of the risk levels. Calculate the first arrival time of the first unmanned vehicle at the collision zone. This time is calculated based on the vehicle's current position, speed, acceleration, and distance to the collision zone. For example, if the first unmanned vehicle is 150 meters away from the collision zone and the current speed is 20 meters / second, assuming uniform motion, then the first arrival time is 150 / 20=7.5 seconds.

[0054] Then, the first risk level of the first unmanned vehicle is determined according to the preset risk level classification rules. The risk level classification rules are usually based on the time to enter the collision zone (Time To Entry, TTE). For example, when TTE is less than or equal to 4.0 seconds, it is dangerous (Dangerous), TTE is between 4.0 and 7.0 seconds, it is attentive (Attentive), and TTE is greater than 7.0 seconds. It is safe (Safe). In this example, the TTE of the first unmanned vehicle is 7.5 seconds, so its risk level is safe (Safe).

[0055] The same method is used to calculate the second arrival time and second risk level of the second unmanned vehicle. Assuming that the second unmanned vehicle is 100 meters away from the collision zone and its current speed is 15 meters per second, its TTE is 100 / 15≈6.67 seconds. According to the risk level classification rules, its risk level is Attentive. After determining the risk levels of the two vehicles, the next step is to compare the first risk level and the second risk level, and determine the target action decision of the first unmanned vehicle based on the comparison results. This process uses a rule-based decision-making algorithm, and the specific rules are as follows:

[0056] If the risk level of the first unmanned vehicle is lower than that of the second unmanned vehicle, the first unmanned vehicle is controlled to maintain the current driving state. In this example, the risk level (Safe) of the first unmanned vehicle is lower than the risk level (Attentive) of the second unmanned vehicle, so the decision result is to control the first unmanned vehicle to maintain the current driving state, that is, continue to drive at a speed of 20 m / s.

[0057] If the risk level of the first unmanned vehicle is higher than that of the second unmanned vehicle, the first unmanned vehicle is controlled to stop at the stop line. The stop line is usually set at a certain distance from the collision zone, such as 10 meters. This decision ensures that the higher-risk vehicle can stop in time to avoid potential collision risks.

[0058] If the risk level of the first unmanned vehicle is equal to that of the second unmanned vehicle, the right-of-way priority of the two vehicles needs to be further considered to determine the target action decision of the first unmanned vehicle. The right-of-way priority may be based on a variety of factors, such as the urgency of the vehicle's mission, vehicle load, road conditions, etc. For example, if the first unmanned vehicle is performing an emergency material transportation mission, then even if the risk level is the same, it may be given a higher right-of-way priority and thus obtain the right of way.

[0059] In actual applications, this decision-making process is dynamic and continuous. It continuously updates the vehicle's location, speed and other information, calculates the risk level in real time, and adjusts the decision based on the latest situation. This dynamic decision-making mechanism ensures that the unmanned vehicle can adapt to the complex and changing mining environment and achieve safe and efficient collaborative operation. Figure 2 The figure shows a schematic diagram of the decision-making process in an embodiment of the present application.

[0060] For example: Assume that on a transport channel in a mining area, the first unmanned vehicle A is traveling at a speed of 20 m / s, 150 meters away from the potential collision zone. At the same time, the second unmanned vehicle B is approaching the same collision zone from the side at a speed of 15 m / s, 100 meters away. First, the TTE of the two vehicles is calculated: 7.5 seconds (Safe) for vehicle A and 6.67 seconds (Attentive) for vehicle B. Comparing the risk levels of the two vehicles, vehicle A is lower than vehicle B, so vehicle A is instructed to maintain its current driving state. However, 5 seconds later, the system reassesses the situation. At this time, vehicle A is 50 meters away from the collision zone, the speed is still 20 m / s, and the TTE drops to 2.5 seconds (Dangerous); vehicle B is 25 meters away from the collision zone, the speed drops to 10 m / s (probably because it received a deceleration instruction before), and the TTE is 2.5 seconds (Dangerous). Since the two vehicles have the same risk level, the system further considers the right of way priority. Assuming that vehicle A is transporting emergency supplies and has a higher road priority, vehicle A is instructed to continue driving while an emergency stop command is sent to vehicle B. This can coordinate the behavior of multiple vehicles in a complex environment, minimize the risk of collision, and improve transportation efficiency.

[0061] In a specific embodiment, the preset risk level classification rule includes: classifying the risk levels from high to low according to the arrival time from short to long.

[0062] In a specific embodiment, the driving state information further includes: a second position coordinate and a second speed of the second unmanned vehicle;

[0063] The first arrival time and the second arrival time are determined as follows:

[0064] (1) Obtaining the first position coordinates and first speed of the current first unmanned vehicle;

[0065] (2) determining a first collision distance according to the first position coordinates and position information of the collision area;

[0066] (3) determining a first arrival time based on the first speed and the first collision distance; and

[0067] (4) determining a second collision distance according to the second position coordinates and the position information of the collision area;

[0068] (5) calculating a second arrival time according to the second speed and the second collision distance;

[0069] Optionally, when the first arrival time is greater than zero, it indicates that the first unmanned vehicle has not yet arrived at the collision zone; when the first arrival time is not greater than zero, it indicates that the first unmanned vehicle is in the collision zone; when the second arrival time is greater than zero, it indicates that the second unmanned vehicle has not yet arrived at the collision zone; when the second arrival time is not greater than zero, it indicates that the second unmanned vehicle is in the collision zone.

[0070] Specifically, in the collaborative operation system of unmanned vehicles in mines, the preset risk level division rule is a key factor in ensuring safe and efficient operation. This rule divides the risk level from high to low according to the arrival time from short to long. This division method is intuitive and effective, because the shorter the arrival time, the closer the vehicle is to the potential collision point, and the higher the risk. In order to accurately calculate the risk level, it is necessary to obtain and process the driving status information of the vehicle. For the second unmanned vehicle, in addition to the information mentioned above, it also includes its second position coordinates and second speed. These data are collected in real time through on-board sensors, including position information provided by the GPS positioning system and speed data provided by the speed sensor.

[0071] First, obtain the first position coordinates and first speed of the current first unmanned vehicle. The position coordinates are usually expressed as latitude and longitude values, while the speed is in meters per second. Next, determine the first collision distance based on the first position coordinates and the location information of the collision zone. Then, use the first speed and the first collision distance to calculate the first arrival time. This calculation assumes that the vehicle moves at a constant speed at the current speed. Although the vehicle speed may fluctuate in actual situations, this simplified calculation method is accurate enough in most cases. For the second unmanned vehicle, the same method is used to calculate the second arrival time. First, determine the second collision distance based on the second position coordinates and the location information of the collision zone, and then calculate the second arrival time using the second speed and the second collision distance.

[0072] In practical applications, the positive and negative values ​​of the arrival time have important practical significance. When the first arrival time is greater than zero, it means that the first unmanned vehicle has not yet reached the collision zone; when the first arrival time is not greater than zero, it means that the first unmanned vehicle has entered the collision zone. The same judgment criteria apply to the second unmanned vehicle. This judgment mechanism provides a clear critical point, which helps to make corresponding decisions and adjustments in a timely manner.

[0073] For example, suppose there is a potential collision zone on the transportation channel of a mining area, and its center point coordinates are (longitude: 116.3, latitude: 39.9). The current position coordinates of the first unmanned vehicle A are (116.298, 39.901), and the speed is 10 meters per second; the current position coordinates of the second unmanned vehicle B are (116.302, 39.899), and the speed is 8 meters per second.

[0074] First, calculate the distance from the two cars to the center of the collision zone. Using the geographic coordinate distance calculation formula, it is found that the distance from car A to the collision zone is about 200 meters, and the distance from car B to the collision zone is about 150 meters. Then, calculate the arrival time of the two cars respectively: the arrival time of car A is 200 / 10=20 seconds, and the arrival time of car B is 150 / 8≈18.75 seconds.

[0075] According to the preset risk level classification rules, Car B has a shorter arrival time, so its risk level is higher. Assume that the risk level is divided into three levels: high, medium, and low, and the corresponding time thresholds are 10 seconds and 20 seconds respectively. In this example, both Car A and Car B are at a medium risk level, but Car B is slightly more risky than Car A. Based on this risk assessment result, it is decided to slow down Car B slightly to increase its arrival time, thereby reducing the risk of collision. At the same time, the position and speed changes of the two vehicles are continuously monitored, and the risk assessment results are dynamically updated.

[0076] If a few seconds later, the speed of car A increases to 12 m / s, while the speed of car B decreases to 7 m / s, the arrival time is recalculated immediately: the new arrival time of car A is (200-10×a few seconds) / 12≈15 seconds, and the new arrival time of car B is (150-8×a few seconds) / 7≈19 seconds. At this point, the risk level of car A will increase, and car A may be required to slow down or change its route.

[0077] In a specific embodiment, determining the first collision distance according to the first position coordinates and the position information of the collision area includes:

[0078] (1) Based on the bounding box of the first unmanned vehicle and the first predicted trajectory, identify the collision intersection of the bounding box in the collision area to obtain the coordinates of the first collision intersection;

[0079] (2) determining a first collision distance according to the first position coordinates and the first collision intersection coordinates;

[0080] (3) Determining a second collision distance according to the second position coordinates and the position information of the collision area, including:

[0081] (4) Based on the bounding box of the second unmanned vehicle and the second predicted trajectory, identify the collision intersection of the bounding box in the collision area to obtain the coordinates of the second collision intersection;

[0082] (5) Determine a second collision distance according to the second position coordinates and the second collision intersection coordinates.

[0083] Specifically, for the first unmanned vehicle, collision analysis needs to be performed based on its bounding box and predicted trajectory. The bounding box is a virtual cube that completely surrounds the outer contour of the vehicle, including the body, wheels, and any protruding parts. This bounding box simplifies the geometry of the vehicle and facilitates rapid collision detection. The predicted trajectory is the path that the vehicle may travel in the future based on the vehicle's current speed, direction, and predetermined route.

[0084] The collision zone is identified by the bounding box collision intersection. The collision zone is usually a road intersection or a potential danger zone, which is also represented by a bounding box. By comparing the intersection of the bounding box of the first unmanned vehicle and the bounding box of the collision zone as it moves along the predicted trajectory, the point where they first contact can be determined. This point is the first collision intersection. After obtaining the coordinates of the first collision intersection, combined with the current position coordinates of the first unmanned vehicle, the first collision distance can be calculated. This distance takes into account not only the straight-line distance, but also the actual shape of the road and the vehicle's driving trajectory. For example, if the road is curved, the collision distance will be the curve distance along the predicted trajectory, rather than a simple straight-line distance.

[0085] For the second unmanned vehicle, the same method is used for calculation. Based on the bounding box of the second unmanned vehicle and the second predicted trajectory, the collision intersection of the bounding box is identified for the collision area to obtain the coordinates of the second collision intersection. Then, based on the current position coordinates of the second unmanned vehicle and the coordinates of the second collision intersection, the second collision distance is calculated. This embodiment takes into account the actual size and predicted trajectory of the vehicle, rather than simply treating the vehicle as a point. This is especially important in narrow or complex mine road environments, where the size and turning radius of the vehicle may significantly affect the assessment of collision risk.

[0086] For example: Suppose that at a T-shaped intersection in a mining area, the first unmanned vehicle A is a large mining truck, 10 meters long, 5 meters wide, and 4 meters high. Its current position coordinates are (100,200), heading east, and its speed is 10 meters per second. The second unmanned vehicle B is a smaller transport vehicle, 6 meters long, 3 meters wide, and 3 meters high. Its current position coordinates are (300,180), heading north, and its speed is 8 meters per second. The collision zone is defined as a 15 meter × 15 meter square area with a center point coordinate of (250,200). Create a 10 meter × 5 meter × 4 meter bounding box for vehicle A, with the center point at (100,200). Based on its current speed and direction, predict its trajectory for the next 5 seconds. Then, identify the collision intersection of the bounding box for the collision zone. Through calculation, it is found that the bounding box of vehicle A will intersect with the collision zone bounding box for the first time after 4.5 seconds, and the intersection coordinates are (242.5,200). This means that the front end of Car A will enter the collision zone for the first time at this point.

[0087] Furthermore, calculate the distance from the current position of car A (100,200) to the collision intersection (242.5,200). In this simplified example, the distance is 142.5 meters. This is the first collision distance of car A. Perform similar calculations for car B. Create a bounding box of 6 meters × 3 meters × 3 meters with the center point at (300,180). Predict its trajectory for the next 5 seconds and identify the collision intersection. Calculations show that car B will enter the collision zone for the first time in 2.5 seconds, with the intersection coordinates at (300,195). Calculate the distance from the current position of car B to the collision intersection and get 15 meters as the second collision distance of car B.

[0088] In a specific embodiment, determining a target action decision of the first unmanned vehicle according to the right-of-way priorities of the first unmanned vehicle and the second unmanned vehicle includes:

[0089] (1) when the first unmanned vehicle has a higher right of way priority than the second unmanned vehicle, controlling the first unmanned vehicle to maintain the current driving state; and / or

[0090] (2) When the first unmanned vehicle has a lower right-of-way priority than the second unmanned vehicle, the first unmanned vehicle is controlled to stop at the stop line.

[0091] Specifically, right-of-way priority is a comprehensive indicator that reflects the weight of a vehicle's passage under specific circumstances. The calculation of this indicator takes into account multiple factors, including the vehicle's mission urgency, load condition, vehicle type, and current traffic flow. The determination of right-of-way priority involves a multi-dimensional scoring system. Each dimension has its own specific weight, which can be dynamically adjusted according to the specific operational needs of the mine. For example, mission urgency may be given the highest weight because it is directly related to the efficiency and safety of mine production. Load condition is also an important factor, because fully loaded vehicles usually require longer braking distances and may therefore be given a higher priority. Vehicle type also affects priority, and large mining trucks usually have higher priority than small transport trucks. In addition, the current traffic flow situation is also taken into account to ensure the overall smoothness of traffic.

[0092] After determining the right-of-way priority of the first unmanned vehicle and the second unmanned vehicle, the target action decision of the first unmanned vehicle will be made according to their relative priority. If the right-of-way priority of the first unmanned vehicle is greater than that of the second unmanned vehicle, the first unmanned vehicle will be controlled to maintain its current driving state. This means that the first unmanned vehicle can continue to drive along the original route and speed, while the second unmanned vehicle needs to take measures to yield. In this case, a corresponding instruction will also be sent to the second unmanned vehicle, requiring it to slow down or stop to ensure the safe passage of the first unmanned vehicle.

[0093] On the contrary, if the first unmanned vehicle has a lower right of way priority than the second unmanned vehicle, the first unmanned vehicle will be controlled to stop at the stop line. The stop line is usually set at a certain distance from the potential collision zone. This distance needs to take into account the braking performance of the vehicle and the road conditions. When the first unmanned vehicle receives the stop command, it will calculate the optimal deceleration curve based on the current speed and the distance from the stop line to ensure that it can stop smoothly at the stop line. This decision-making mechanism based on right of way priority not only ensures the safety of vehicle operation, but also improves the efficiency of the entire mine transportation system. By giving different priorities to different vehicles, the system can better handle complex traffic situations, such as priority passage for emergency tasks and avoiding frequent stops of large vehicles.

[0094] For example, suppose that at a crossroad in a mining area, the first unmanned vehicle A is a large mining truck fully loaded with ore, performing an emergency transportation task. The second unmanned vehicle B is an empty medium-sized transport vehicle, performing a routine inspection. First, calculate the right-of-way priority of the two vehicles.

[0095] For vehicle A, the following factors are considered: mission urgency (weight 0.4, full score 10, 9 points), load condition (weight 0.3, full score 10, 10 points), vehicle type (weight 0.2, full score 10, 9 points), current driving speed (weight 0.1, full score 10, 7 points). The comprehensive priority score of vehicle A is calculated as: 9×0.4+10×0.3+9×0.2+7×0.1=9.1 points.

[0096] For vehicle B, the system also considers these factors: mission urgency (5 points), load condition (3 points), vehicle type (6 points), and current driving speed (7 points). The calculated overall priority score for vehicle B is: 5×0.4+3×0.3+6×0.2+7×0.1=4.8 points.

[0097] Comparing the priority scores of the two cars, Car A is obviously higher than Car B. Therefore, Car A is controlled to maintain its current driving state, while sending a deceleration command to Car B. Specifically, Car A will continue to maintain its current speed (assuming 20km / h) through the intersection, while Car B needs to slow down to 5km / h at the virtual stop line 30 meters away from the intersection and be prepared to stop completely if necessary to ensure the safe passage of Car A.

[0098] In a specific embodiment, according to the first position relationship and the second position relationship, an action decision analysis is performed on the first unmanned vehicle to obtain a target action decision of the first unmanned vehicle, including:

[0099] (1) when the first position relationship indicates that the first unmanned vehicle has not yet reached the collision zone, and the second position relationship indicates that the second unmanned vehicle is in the collision zone, controlling the first unmanned vehicle to stop at the stop line; and / or

[0100] (2) when the first position relationship indicates that the first unmanned vehicle is within the collision zone, and the second position relationship indicates that the second unmanned vehicle has not yet reached the collision zone, controlling the first unmanned vehicle to maintain the current driving state; and / or

[0101] (3) When the first position relationship indicates that the first unmanned vehicle is within the collision zone, and the second position relationship indicates that the second unmanned vehicle is within the collision zone, the first unmanned vehicle is controlled to stop immediately.

[0102] Specifically, the positional relationship between the first unmanned vehicle and the second unmanned vehicle relative to the collision zone is determined. The collision zone here refers to the area where two vehicles may intersect or collide, usually a road intersection or a narrow road section. The positional relationship is determined based on the real-time GPS coordinates, speed, direction and predicted trajectory of the vehicle. The system uses a complex spatial analysis algorithm to compare the current position of the vehicle with the predefined collision zone boundary to determine whether the vehicle has entered the collision zone or is still outside the area.

[0103] After determining the positional relationship between the two vehicles, make corresponding decisions based on different situations. When the first unmanned vehicle has not yet reached the collision zone, and the second unmanned vehicle is already in the collision zone, control the first unmanned vehicle to stop at the stop line. The stop line is a virtual safety boundary, usually set at a certain distance from the entrance of the collision zone. This distance needs to take into account the braking performance of the vehicle and road conditions. Calculate the best deceleration curve to ensure that the vehicle can stop smoothly and safely at the stop line.

[0104] On the contrary, if the first unmanned vehicle is already in the collision zone and the second unmanned vehicle has not yet reached the collision zone, the first unmanned vehicle is controlled to maintain the current driving state. This decision is based on the principle of "first come, first served", and the vehicle that has entered the collision zone has priority. At the same time, a deceleration or stop command is sent to the second unmanned vehicle to ensure that no collision occurs. In the most dangerous case, that is, both vehicles have entered the collision zone, the first unmanned vehicle is immediately controlled to stop urgently. This is a conservative but necessary safety measure to prevent any possible collision. At the same time, the system will also send an emergency stop command to the second unmanned vehicle to minimize the risk. This decision-making mechanism based on position relationship requires a highly accurate positioning system and fast data processing capabilities. The positioning system usually combines multiple sensors such as GPS, inertial navigation system (INS) and wheel encoders to achieve centimeter-level positioning accuracy. Data processing adopts a distributed computing architecture, which can complete the entire process from data acquisition to decision output within milliseconds.

[0105] For example: Suppose at a T-shaped intersection in a mining area, a 10m x 10m square area is defined as the collision zone. The first unmanned vehicle A is a large mining truck with a load of 60 tons. It is approaching the collision zone at a speed of 15km / h. The current position is 5 meters away from the boundary of the collision zone. The second unmanned vehicle B is a 20-ton medium-sized transport vehicle and has entered the collision zone by 2 meters. The system first processes the GPS data and speed information of the two vehicles, and combines it with the predefined collision zone map to determine that vehicle A has not yet entered the collision zone, while vehicle B is already in the collision zone.

[0106] Based on this positional relationship, a decision is made immediately to control car A to stop at the virtual stop line 3 meters away from the collision zone. The system calculates that considering the current speed and weight of car A, it needs to reduce the speed from 15km / h to 0 within 1.5 seconds, which requires about -2.8m / s 2The system then sends a command to the brake system of car A to start a precise deceleration process. At the same time, the system sends a command to car B to maintain its current speed and continue through the collision zone to clear the intersection as soon as possible. During the entire process, the position and speed changes of both cars are continuously monitored. If any abnormal situation is detected, such as the braking system of car A not responding as expected, the decision is immediately escalated, and emergency stop commands may be issued to both cars at the same time.

[0107] In a specific embodiment, after the first unmanned vehicle stops in the collision zone, the method further includes:

[0108] (1) if the right-of-way priority of the first unmanned vehicle is higher than the right-of-way priority of the second unmanned vehicle, controlling the first unmanned vehicle to re-plan its driving route and drive according to the new driving route; and / or,

[0109] (2) If the right-of-way priority of the first unmanned vehicle is lower than that of the second unmanned vehicle, the first unmanned vehicle is controlled to remain parked and the driving status of the second unmanned vehicle is monitored; and if it is detected that the second unmanned vehicle has not driven within a preset monitoring time interval, the first unmanned vehicle is controlled to re-plan the driving route and drive according to the new driving route.

[0110] Specifically, after the first unmanned vehicle stops in the collision zone, further decisions need to be made based on the right-of-way priority and real-time monitoring information. First, the right-of-way priority of the first unmanned vehicle and the second unmanned vehicle is compared. The right-of-way priority is a comprehensive indicator that takes into account multiple factors, such as the urgency of the vehicle's mission, load conditions, vehicle type, etc. Each factor is assigned a weight, and the final right-of-way priority score is obtained through weighted calculation. For example, the urgency of the mission may account for 40% of the weight, the load condition accounts for 30%, the vehicle type accounts for 20%, and the current driving speed accounts for 10%.

[0111] If the first unmanned vehicle has a higher priority than the second unmanned vehicle, the path replanning algorithm is immediately started. This algorithm first analyzes the current traffic conditions and road network, and then uses the A* algorithm or Dijkstra algorithm to calculate a new optimal path. During the calculation process, the algorithm takes into account multiple factors, such as path length, estimated travel time, road slope, turning radius, etc. The goal of the new path is to leave the current collision zone as quickly as possible and reach the destination while ensuring safety.

[0112] Once the new path is determined, it is converted into a series of navigation instructions, including turning, accelerating, decelerating, etc., and these instructions are sent to the control system of the first unmanned vehicle. At the same time, the system will also send notifications to other related vehicles to inform them of the new driving plan of the first unmanned vehicle to ensure the coordination of the overall traffic.

[0113] On the other hand, if the first unmanned vehicle has a lower right of way priority than the second unmanned vehicle, the system will control the first unmanned vehicle to remain parked and start a real-time monitoring program for the second unmanned vehicle. This monitoring program will continue to track the location, speed, direction and other information of the second unmanned vehicle. The system will set a preset monitoring time interval, such as 10 seconds. During this time interval, if the second unmanned vehicle is detected to start moving, the system will continue to keep the first unmanned vehicle parked until the second unmanned vehicle safely passes the collision zone.

[0114] However, if the second unmanned vehicle is still not moving within the preset monitoring time interval, this may indicate that the second unmanned vehicle has a malfunction or other abnormal situation. In this case, the system will re-evaluate the current traffic conditions and initiate a path replanning procedure for the first unmanned vehicle. This replanning process is similar to the process described above, but may take into account more safety factors because the status of the second unmanned vehicle is uncertain at this time.

[0115] For example, suppose that at a T-shaped intersection in a mining area, the first unmanned vehicle A is a large mining truck with a load capacity of 70 tons, which is carrying out an emergency material transportation mission. The second unmanned vehicle B is a medium-sized transport vehicle with a load capacity of 30 tons, which is carrying out regular transportation. The two vehicles meet in the collision zone, and vehicle A is forced to stop.

[0116] First, the right-of-way priority of the two vehicles is calculated. The calculation result of vehicle A is as follows: mission urgency (9 points × 40%), load condition (10 points × 30%), vehicle type (9 points × 20%), current driving speed (7 points × 10%), and the total score is 9.1 points. The calculation result of vehicle B is: mission urgency (5 points × 40%), load condition (6 points × 30%), vehicle type (7 points × 20%), current driving speed (7 points × 10%), and the total score is 5.9 points.

[0117] Since car A has a higher right of way priority than car B, re-routing is initiated immediately for car A. Assuming the original planned route is to go straight through the T-shaped intersection, the newly planned route may choose to turn right, then turn left at the next intersection 100 meters later, and finally return to the original route. Although this new route is 50 meters longer than the original route, it can avoid the current traffic conflict and is expected to save 2 minutes.

[0118] The new route is converted into a series of instructions: turn right 45 degrees, accelerate to 20 km / h, go straight for 100 meters, turn left 90 degrees, accelerate to 35 km / h, etc. These instructions are sent to the control system of car A, and car A then drives according to the new route. At the same time, the system sends a notification to car B and other nearby vehicles to inform them of car A's new driving plan.

[0119] In a specific embodiment, the right-of-way priority between the first unmanned vehicle and the second unmanned vehicle is determined in the following manner;

[0120] (1) Obtaining first right-of-way information of the first unmanned vehicle and second right-of-way information of the second unmanned vehicle;

[0121] (2) Determine the right-of-way priority of the first unmanned vehicle and the second unmanned vehicle according to a preset right-of-way priority principle and the first right-of-way information and the second right-of-way information.

[0122] Specifically, the right-of-way information of the first unmanned vehicle and the second unmanned vehicle is obtained. The right-of-way information is a set of data that comprehensively reflects the current status of the vehicle and the importance of the task, including the empty and loaded status of the vehicle, the type of task being performed, the vehicle identity (such as a large mining truck or a medium-sized transport truck), and the current driving speed. This information is collected and updated in real time through on-board sensors and the task management system.

[0123] The empty and loaded states directly affect the braking performance and energy efficiency of the vehicle. For example, a fully loaded vehicle usually requires a longer braking distance and may therefore be given a higher priority in some cases. The mission type reflects the urgency and importance of the mission currently being performed by the vehicle, such as emergency material transportation, routine ore transportation, equipment maintenance, etc. Vehicle identity information includes vehicle type, load capacity, etc. These factors affect the vehicle's position in mining traffic. The current driving speed is closely related to the continuity and safety of traffic flow.

[0124] The evaluation is carried out according to the preset right of way priority principle. This principle is a multi-dimensional scoring system, each dimension corresponds to an aspect of the right of way information and is assigned a specific weight. For example, the mission type may account for 40% of the weight, the empty or loaded state accounts for 30%, the vehicle identity accounts for 20%, and the current speed accounts for 10%.

[0125] The specific scoring process is as follows: For the mission type, emergency material transportation may get 10 points, regular transportation may get 7 points, and equipment maintenance may get 5 points. For empty and heavy load status, full load may get 9 points, half load may get 7 points, and empty load may get 5 points. For vehicle identity, a large mining truck may get 9 points, a medium-sized transport truck may get 7 points, and a small maintenance truck may get 5 points. For speed, an ideal speed range can be set, and the score is highest within this range. If it deviates, the score will be reduced.

[0126] Then, the scores of each dimension are multiplied by the corresponding weights and the results are added together to obtain the final road priority score. This score reflects the comprehensive priority of the vehicle in the current situation.

[0127] For example, suppose the first unmanned vehicle A is a large mining truck fully loaded with ore, carrying out an emergency material transportation mission, and its current speed is 25 km / h. The second unmanned vehicle B is a half-loaded medium-sized transport vehicle, carrying out a regular transportation mission, and its current speed is 30 km / h.

[0128] For vehicle A, the calculation is as follows:

[0129] Mission type (emergency material transportation): 10 points x 40% = 4 points

[0130] Empty and heavy load state (full load): 9 points × 30% = 2.7 points

[0131] Vehicle identity (large mining truck): 9 points x 20% = 1.8 points

[0132] Current speed (25 km / h, assuming it is within the ideal range): 9 minutes x 10% = 0.9 minutes

[0133] Total score of car A: 4+2.7+1.8+0.9=9.4 points

[0134] For vehicle B, the calculation is as follows:

[0135] Mission type (conventional transport): 7 points x 40% = 2.8 points

[0136] Empty and heavy load state (half load): 7 points × 30% = 2.1 points

[0137] Vehicle status (medium-sized transport vehicle): 7 points x 20% = 1.4 points

[0138] Current speed (30 km / h, assumed to be slightly higher than the ideal range): 8 minutes x 10% = 0.8 minutes

[0139] Total score of car B: 2.8+2.1+1.4+0.8=7.1 points

[0140] By comparison, the right-of-way priority score of car A (9.4 points) is higher than that of car B (7.1 points), so in this case, car A will be given a higher right-of-way priority. This means that in the event of a traffic conflict, car A will get the right of way, while car B may need to give way or slow down.

[0141] In a specific embodiment, the preset right-of-way priority principle includes: presetting a plurality of right-of-way priority levels; each right-of-way priority level is provided with a corresponding priority index; in the same right-of-way priority level, the higher the index value of the corresponding priority index is, the higher the right-of-way priority of the unmanned vehicle is; when the index values ​​of the corresponding priority indexes in the same right-of-way priority level are the same, the right-of-way priority of the unmanned vehicle is determined in turn according to the index values ​​of the corresponding priority indexes of the adjacent lower right-of-way priority levels;

[0142] Determining the right-of-way priorities of the first unmanned vehicle and the second unmanned vehicle according to the preset right-of-way priority principle and the first right-of-way information and the second right-of-way information includes:

[0143] The highest road right priority level among the multiple road right priority levels is determined as the current road right priority level, and the following steps are performed for the current road right priority level:

[0144] Step 1: Determine the index values ​​of the priority indexes corresponding to the first and second unmanned vehicles at the current right-of-way priority levels according to the first right-of-way information and the second right-of-way information;

[0145] Step 2: If the first unmanned vehicle and the second unmanned vehicle have the same index value for the priority index corresponding to the current right-of-way priority level, then the right-of-way priority level one level lower than the current right-of-way priority level is used as the new current right-of-way priority level, and step 1 is re-executed; otherwise, the right-of-way priority of the first unmanned vehicle and the second unmanned vehicle is determined.

[0146] In a specific embodiment, the priority indicators corresponding to the multiple road right priority levels are related to at least one of the following information: empty load, mission type, unmanned vehicle identity, and current driving speed.

[0147] Specifically, the road right priority system is usually divided into multiple levels, for example, it can be set to five levels: highest level, high level, medium level, low level and lowest level. Each level has its corresponding priority indicators, which reflect the importance or urgency of the vehicle in a specific aspect. For example, the highest level may correspond to the type of emergency mission, the high level corresponds to the vehicle identity, the medium level corresponds to the empty or heavy load status, the low level corresponds to the current driving speed, and the lowest level may be the vehicle number and other information used for the final distinction.

[0148] In the same right-of-way priority level, the higher the indicator value of the priority indicator, the higher the right-of-way priority of the unmanned vehicle. This design ensures that within the same level, more important or more urgent vehicles can get priority. For example, in the mission type level, emergency material transportation may be given 10 points, while daily ore transportation may only be given 7 points. When two vehicles have the same indicator value at the same level, the evaluation system automatically turns to the next lower level indicator for comparison. This level-by-level comparison method ensures that clear priority judgments can still be made in complex situations.

[0149] The specific road right priority determination process is as follows: First, set the highest road right priority level as the current evaluation level. Then, based on the road right information of the first unmanned vehicle and the second unmanned vehicle, determine the specific index values ​​of their corresponding priority indicators at the current level. If the index values ​​of the two vehicles at this level are different, the vehicle with a higher index value will be given a higher road right priority. If the index values ​​are the same, the evaluation level will be lowered by one level and compared again. This process will continue until a level that can distinguish the priorities of the two vehicles is found, or until all levels are exhausted.

[0150] For example, suppose two unmanned vehicles A and B are driving in a mining area and are about to meet at an intersection. It is necessary to determine which vehicle has the right of way.

[0151] Highest level (task type):

[0152] Vehicle A: Emergency material transportation (index value 10)

[0153] Vehicle B: Emergency material transportation (index value 10)

[0154] Result: Same, continue to compare the next level

[0155] Second level (vehicle identity):

[0156] Vehicle A: Large mining truck (index value 9)

[0157] Vehicle B: Large mining truck (index value 9)

[0158] Result: Still the same, continue to compare

[0159] The third level (empty reload state):

[0160] Car A: Fully loaded (index value 9)

[0161] Car B: half loaded (index value 7)

[0162] Result: Car A has higher priority than Car B

[0163] In this example, three rounds of comparisons ultimately determined that vehicle A had a higher right-of-way priority. This decision-making process takes multiple factors into account and conducts rigorous comparisons level by level according to the preset priority principles. This multi-level comparison method can not only effectively handle complex traffic conditions, but also flexibly adapt to the specific needs of different mining areas. For example, some mining areas may pay more attention to the type of task, while other mining areas may pay more attention to the vehicle's load status. By adjusting the order of the priority levels and the indicators corresponding to each level, optimization can be performed according to the characteristics of different mining areas.

[0164] In a specific embodiment, the stop line is determined in the following manner:

[0165] (1) extracting a first stop point at a third preset distance from the first collision intersection in the first predicted trajectory based on the third preset distance and the coordinates of the first collision intersection, wherein the coordinates of the first collision intersection are obtained by performing bounding box collision intersection recognition on the collision area based on the bounding box of the first unmanned vehicle and the first predicted trajectory;

[0166] (2) With the first stopping point as the center, a line segment of a preset length is constructed in a direction perpendicular to the first predicted trajectory as a stopping line.

[0167] Specifically, the bounding box is a virtual cube that completely surrounds the outer contour of the unmanned vehicle, including the body, wheels, and any protruding parts. This bounding box simplifies the geometry of the vehicle and facilitates fast collision detection.

[0168] The first predicted trajectory is the path that the vehicle may travel in the future based on the vehicle's current speed, direction, and planned route. This trajectory is usually expressed as a series of continuous coordinate points, each of which represents the expected position of the vehicle at a specific point in time.

[0169] The collision zone usually refers to a road intersection or other potential dangerous area, which is also represented by a bounding box. The collision intersection identification of the collision zone is to calculate the point where the boundary bounding box of the first unmanned vehicle first intersects with the collision zone bounding box when it moves along the first predicted trajectory. This point is the first collision intersection, and its coordinates are crucial for determining the position of the stop line.

[0170] The process of determining the stop line begins with setting the third preset distance. This distance is a safety buffer, usually determined based on the vehicle's braking performance, current speed, and road conditions. For example, for a large mining truck, this preset distance may be set to 30 meters, taking into account its longer braking distance.

[0171] Next, find a point on the first predicted trajectory that is a third preset distance away from the first collision intersection. This point is the first stop point. This process involves traversal of trajectory points and distance calculation. Specifically, starting from the first collision intersection, search backward along the first predicted trajectory, and calculate the distance from each trajectory point to the first collision intersection until a point is found whose distance is equal to or slightly greater than the third preset distance. After finding the first stop point, the actual stop line needs to be constructed. The stop line is a line segment perpendicular to the first predicted trajectory, and its center point is the first stop point. The length of the line segment is pre-set and is usually slightly larger than the lane width to ensure that the entire driving path is covered. For example, if the lane width is 5 meters, the length of the stop line may be set to 6 meters to provide additional safety margin.

[0172] In order to calculate the two endpoints of the stop line, it is necessary to determine the direction vector of the first predicted trajectory at the first stop point. Then, a unit vector perpendicular to this direction vector is calculated, multiplied by half of the preset length, and added to and subtracted from the coordinates of the first stop point, respectively, to obtain the two endpoints of the stop line.

[0173] For example, suppose a large mining truck A is approaching a T-shaped intersection. Through the onboard sensors and prediction algorithms, the first predicted trajectory of vehicle A is obtained, which is a series of coordinate points, such as {(100,200),(105,205),(110,210),...}. The collision zone is defined as a 15m x 15m square area with the center point coordinates (150,250).

[0174] First, identify the collision intersection of the bounding box. Assume that the bounding box of car A is a cuboid of 10 meters × 3 meters × 4 meters. Through calculation, it is found that the bounding box of car A will intersect the collision zone for the first time at the coordinates (140, 240), which is the first collision intersection. Then, set the third preset distance to 30 meters. Starting from the first collision intersection (140, 240), search backward along the first predicted trajectory until a point with a distance of about 30 meters is found. Assume that the coordinates of this point are (115, 215), which is the first stopping point.

[0175] Next, calculate the direction vector of the first predicted trajectory at point (115,215), assuming it is (1,1). The unit vector perpendicular to it is (-1 / √2,1 / √2). If the preset stop line length is 6 meters, then the coordinates of the two endpoints of the stop line are:

[0176] Endpoint 1: (115-3 / √2, 215+3 / √2)

[0177] Endpoint 2: (115+3 / √2,215-3 / √2)

[0178] In this way, the stop line is determined. This stop line is perpendicular to the predicted trajectory of vehicle A, is 6 meters long, and its center point is located 30 meters away from the potential collision point. When vehicle A approaches this stop line, if a potential collision risk is detected, the control system will instruct the vehicle to stop here and wait for safe passage conditions.

[0179] The above describes the multi-agent yielding method based on road rights in the embodiment of the present application. The following describes the multi-agent yielding device based on road rights in the embodiment of the present application. Figure 3 In the embodiment of the present application, an embodiment of the multi-agent yielding device based on road rights includes:

[0180] An acquisition module 201 is used to acquire driving state information of a second unmanned vehicle; wherein the driving state information at least includes a second predicted trajectory of the second unmanned vehicle;

[0181] A collision zone analysis module 202, configured to perform a collision zone analysis on the first unmanned vehicle and the second unmanned vehicle according to the first predicted trajectory of the first unmanned vehicle and the second predicted trajectory, and obtain a collision zone analysis result;

[0182] A position relationship determination module 203, configured to respectively determine a first position relationship between the first unmanned vehicle and the collision zone and a second position relationship between the second unmanned vehicle and the collision zone when the collision zone analysis result indicates that a collision zone exists;

[0183] The decision module 204 is used to perform action decision analysis on the first unmanned vehicle according to the first position relationship and the second position relationship, obtain a target action decision of the first unmanned vehicle, and control the first unmanned vehicle to perform an action according to the target action decision.

[0184] The decision module 204 includes: an analysis unit, which is used to perform an action decision analysis on the first unmanned vehicle according to the risk levels of the first unmanned vehicle and the second unmanned vehicle to obtain a target action decision for the first unmanned vehicle when the first position relationship indicates that the first unmanned vehicle has not yet reached the collision zone and the second position relationship indicates that the second unmanned vehicle has not yet reached the collision zone.

[0185] The analysis unit comprises:

[0186] a first determining unit, configured to determine a first risk level of the first unmanned vehicle according to a first arrival time of the first unmanned vehicle at the collision zone and a preset risk level classification rule;

[0187] a second determining unit, configured to determine a second risk level of the second unmanned vehicle according to a second arrival time of the second unmanned vehicle at the collision zone and a preset risk level classification rule;

[0188] A comparison unit is used to compare the first risk level with the second risk level, and determine a target action decision of the first unmanned vehicle according to the comparison result.

[0189] The first determination unit is specifically used to: obtain the first position coordinates and the first speed of the current first unmanned vehicle; determine the first collision distance according to the first position coordinates and the position information of the collision zone; and determine the first arrival time according to the first speed and the first collision distance.

[0190] The second determining unit is specifically configured to: determine a second collision distance according to the second position coordinates and the position information of the collision area; and calculate a second arrival time according to the second speed and the second collision distance.

[0191] The first determining unit is further specifically configured to:

[0192] Based on the boundary bounding box of the first unmanned vehicle and the first predicted trajectory, identifying the collision intersection point of the bounding box of the collision area to obtain the coordinates of the first collision intersection point;

[0193] determining a first collision distance according to the first position coordinates and the first collision intersection coordinates;

[0194] Determining a second collision distance according to the second position coordinates and the position information of the collision area includes:

[0195] Based on the boundary bounding box of the second unmanned vehicle and the second predicted trajectory, identifying the collision intersection point of the bounding box on the collision area to obtain the coordinates of the second collision intersection point;

[0196] A second collision distance is determined according to the second position coordinates and the second collision intersection coordinates.

[0197] The comparison unit is configured to control the first unmanned vehicle to maintain a current driving state when the first unmanned vehicle has a higher right of way priority than the second unmanned vehicle; and / or

[0198] When the right-of-way priority of the first unmanned vehicle is lower than that of the second unmanned vehicle, the first unmanned vehicle is controlled to stop at the stop line.

[0199] The analysis unit is further specifically configured to: when the first position relationship indicates that the first unmanned vehicle has not yet reached the collision zone, and the second position relationship indicates that the second unmanned vehicle is within the collision zone, control the first unmanned vehicle to stop at the stop line; and / or

[0200] When the first position relationship indicates that the first unmanned vehicle is within the collision zone, and the second position relationship indicates that the second unmanned vehicle has not yet reached the collision zone, controlling the first unmanned vehicle to maintain a current driving state; and / or

[0201] When the first position relationship indicates that the first unmanned vehicle is within the collision zone, and when the second position relationship indicates that the second unmanned vehicle is within the collision zone, the first unmanned vehicle is controlled to stop immediately.

[0202] The multi-agent yielding device based on road right also includes:

[0203] a planning module, configured to control the first unmanned vehicle to replan a driving route and drive along the new driving route if the right-of-way priority of the first unmanned vehicle is higher than the right-of-way priority of the second unmanned vehicle; and / or,

[0204] A monitoring module is used to control the first unmanned vehicle to remain in a parked state and monitor the driving state of the second unmanned vehicle if the right-of-way priority of the first unmanned vehicle is lower than that of the second unmanned vehicle; and to control the first unmanned vehicle to re-plan a driving route and drive along the new driving route if it is detected that the second unmanned vehicle has not driven within a preset monitoring time interval.

[0205] The multi-agent yielding device based on road right also includes:

[0206] A collection module, used to obtain first right-of-way information of the first unmanned vehicle and second right-of-way information of the second unmanned vehicle;

[0207] An information determination module is used to determine the right of way priority of the first unmanned vehicle and the second unmanned vehicle according to a preset right of way priority principle and the first right of way information and the second right of way information.

[0208] The information determination module is specifically used to determine the right of way priorities of the first unmanned vehicle and the second unmanned vehicle according to a preset right of way priority principle and the first right of way information and the second right of way information, including:

[0209] The highest road right priority level among the multiple road right priority levels is determined as the current road right priority level, and the following steps are performed for the current road right priority level:

[0210] Step 1: determining the index values ​​of the priority indexes corresponding to the first unmanned vehicle and the second unmanned vehicle at the current road right priority level according to the first road right information and the second road right information;

[0211] Step 2: If the first unmanned vehicle and the second unmanned vehicle have the same index value of the priority index corresponding to the current road right priority level, the road right priority level one level lower than the current road right priority level will be used as the new current road right priority level, and step 1 will be re-executed; otherwise, the road right priority of the first unmanned vehicle and the second unmanned vehicle will be determined.

[0212] The multi-agent yielding device based on road right also includes:

[0213] a stop line determination module, configured to extract, in the first predicted trajectory, a first stop point that is at the third preset distance from the first collision intersection based on a third preset distance and the coordinates of the first collision intersection, wherein the coordinates of the first collision intersection are obtained by performing bounding box collision intersection recognition on the collision area based on a boundary bounding box of the first unmanned vehicle and the first predicted trajectory;

[0214] Taking the first stopping point as the center, a line segment of a preset length is constructed in a direction perpendicular to the first predicted trajectory as a stopping line.

[0215] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0216] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A control method for an unmanned vehicle, characterized in that: The following steps are involved: Acquire driving state information of the second unmanned vehicle; wherein the driving state information at least includes a second predicted trajectory of the second unmanned vehicle; Performing collision zone analysis on the first unmanned vehicle and the second unmanned vehicle according to the first predicted trajectory of the first unmanned vehicle and the second predicted trajectory to obtain a collision zone analysis result; When the collision zone analysis result indicates that a collision zone exists, determining a first position relationship between the first unmanned vehicle and the collision zone and a second position relationship between the second unmanned vehicle and the collision zone respectively; According to the first position relationship and the second position relationship, an action decision analysis is performed on the first unmanned vehicle to obtain a target action decision of the first unmanned vehicle, and the first unmanned vehicle is controlled to perform an action according to the target action decision.

2. The method according to claim 1, characterized in that According to the first position relationship and the second position relationship, an action decision analysis is performed on the first unmanned vehicle to obtain a target action decision of the first unmanned vehicle, including: When the first position relationship indicates that the first unmanned vehicle has not yet reached the collision zone, and the second position relationship indicates that the second unmanned vehicle has not yet reached the collision zone, an action decision analysis is performed on the first unmanned vehicle according to the risk levels of the first unmanned vehicle and the second unmanned vehicle to obtain a target action decision for the first unmanned vehicle.

3. The method according to claim 2, characterized in that Determining a target action decision of the first unmanned vehicle according to the risk levels of the first unmanned vehicle and the second unmanned vehicle includes: Determining a first risk level of the first unmanned vehicle according to a first arrival time of the first unmanned vehicle at the collision zone and a preset risk level classification rule; Determining a second risk level of the second unmanned vehicle according to a second arrival time of the second unmanned vehicle at the collision zone and a preset risk level classification rule; The first risk level is compared with the second risk level, and a target action decision of the first unmanned vehicle is determined according to the comparison result.

4. The method according to claim 3, characterized in that: Comparing the first risk level with the second risk level, and determining a target action decision of the first unmanned vehicle according to the comparison result, includes: If the risk level of the first unmanned vehicle is lower than that of the second unmanned vehicle, controlling the first unmanned vehicle to maintain the current driving state; and / or If the risk level of the first unmanned vehicle is higher than that of the second unmanned vehicle, controlling the first unmanned vehicle to stop at the stop line; and / or If the risk level of the first unmanned vehicle is equal to that of the second unmanned vehicle, the target action decision of the first unmanned vehicle is determined according to the road right priority of the first unmanned vehicle and the second unmanned vehicle.

5. The method according to claim 3, characterized in that: The preset risk level classification rule includes: classifying the risk levels from high to low according to the arrival time from short to long.

6. The method according to claim 3, characterized in that: The driving state information also includes: a second position coordinate and a second speed of the second unmanned vehicle; The first arrival time and the second arrival time are determined in the following manner: Obtaining a first position coordinate and a first speed of the first unmanned vehicle; Determining a first collision distance according to the first position coordinates and the position information of the collision area; determining a first arrival time according to the first speed and the first collision distance; and determining a second collision distance according to the second position coordinates and the position information of the collision area; Calculate a second arrival time according to the second speed and the second collision distance; Optionally, when the first arrival time is greater than zero, it indicates that the first unmanned vehicle has not yet arrived at the collision zone; when the first arrival time is not greater than zero, it indicates that the first unmanned vehicle is in the collision zone; when the second arrival time is greater than zero, it indicates that the second unmanned vehicle has not yet arrived at the collision zone; when the second arrival time is not greater than zero, it indicates that the second unmanned vehicle is in the collision zone.

7. The method according to claim 6, characterized in that Determining a first collision distance according to the first position coordinates and the position information of the collision area includes: Based on the boundary bounding box of the first unmanned vehicle and the first predicted trajectory, identifying the collision intersection point of the bounding box of the collision area to obtain the coordinates of the first collision intersection point; determining a first collision distance according to the first position coordinates and the first collision intersection coordinates; Determining a second collision distance according to the second position coordinates and the position information of the collision area includes: Based on the boundary bounding box of the second unmanned vehicle and the second predicted trajectory, identifying the collision intersection point of the bounding box on the collision area to obtain the coordinates of the second collision intersection point; A second collision distance is determined according to the second position coordinates and the second collision intersection coordinates.

8. The method according to claim 4, characterized in that Determining a target action decision of the first unmanned vehicle according to the right-of-way priorities of the first unmanned vehicle and the second unmanned vehicle includes: When the first unmanned vehicle has a higher right of way priority than the second unmanned vehicle, controlling the first unmanned vehicle to maintain a current driving state; and / or When the right-of-way priority of the first unmanned vehicle is lower than that of the second unmanned vehicle, the first unmanned vehicle is controlled to stop at the stop line.

9. The method according to claim 1, characterized in that: According to the first position relationship and the second position relationship, an action decision analysis is performed on the first unmanned vehicle to obtain a target action decision of the first unmanned vehicle, including: When the first position relationship indicates that the first unmanned vehicle has not yet reached the collision zone, and the second position relationship indicates that the second unmanned vehicle is within the collision zone, controlling the first unmanned vehicle to stop at the stop line; and / or When the first position relationship indicates that the first unmanned vehicle is within the collision zone, and the second position relationship indicates that the second unmanned vehicle has not yet reached the collision zone, controlling the first unmanned vehicle to maintain a current driving state; and / or When the first position relationship indicates that the first unmanned vehicle is within the collision zone, and when the second position relationship indicates that the second unmanned vehicle is within the collision zone, the first unmanned vehicle is controlled to stop immediately.

10. The method according to claim 9, characterized in that After the first unmanned vehicle stops in the collision zone, the method further includes: If the right-of-way priority of the first unmanned vehicle is higher than the right-of-way priority of the second unmanned vehicle, control the first unmanned vehicle to re-plan a driving route and drive according to the new driving route; and / or, If the right-of-way priority of the first unmanned vehicle is lower than that of the second unmanned vehicle, the first unmanned vehicle is controlled to remain parked and the driving status of the second unmanned vehicle is monitored; and if it is detected that the second unmanned vehicle has not driven within a preset monitoring time interval, the first unmanned vehicle is controlled to re-plan the driving route and drive according to the new driving route.

11. The method according to claim 10, characterized in that The right-of-way priority of the first unmanned vehicle and the second unmanned vehicle is determined in the following manner; Obtaining first right-of-way information of the first unmanned vehicle and second right-of-way information of the second unmanned vehicle; The right-of-way priorities of the first unmanned vehicle and the second unmanned vehicle are determined according to a preset right-of-way priority principle and the first right-of-way information and the second right-of-way information.

12. The method according to claim 11, characterized in that The preset right-of-way priority principle includes: presetting multiple right-of-way priority levels; each right-of-way priority level is set with a corresponding priority index; in the same right-of-way priority level, the higher the index value of the corresponding priority index, the higher the right-of-way priority of the unmanned vehicle; when the index values ​​of the corresponding priority indexes of the same right-of-way priority level are the same, the right-of-way priority of the unmanned vehicle is determined according to the index values ​​of the corresponding priority indexes of the adjacent lower right-of-way priority levels in turn; Determining the right of way priorities of the first unmanned vehicle and the second unmanned vehicle according to a preset right of way priority principle and the first right of way information and the second right of way information includes: The highest road right priority level among the multiple road right priority levels is determined as the current road right priority level, and the following steps are performed for the current road right priority level: Step 1: determining the index values ​​of the priority indexes corresponding to the first unmanned vehicle and the second unmanned vehicle at the current road right priority level according to the first road right information and the second road right information; Step 2: If the first unmanned vehicle and the second unmanned vehicle have the same index value of the priority index corresponding to the current road right priority level, the road right priority level one level lower than the current road right priority level will be used as the new current road right priority level, and step 1 will be re-executed; otherwise, the road right priority of the first unmanned vehicle and the second unmanned vehicle will be determined.

13. The method according to claim 12, characterized in that The priority indicators corresponding to the multiple road right priority levels are related to at least one of the following information: empty or heavy load, mission type, unmanned vehicle identity, and current driving speed.

14. The method according to claim 4, wherein the stop line is determined in the following manner: Based on the third preset distance and the coordinates of the first collision intersection, a first stop point that is the third preset distance away from the first collision intersection is extracted in the first predicted trajectory, wherein: The first collision intersection coordinates are obtained by performing bounding box collision intersection identification on the collision area based on the boundary bounding box of the first unmanned vehicle and the first predicted trajectory; Taking the first stopping point as the center, a line segment of a preset length is constructed in a direction perpendicular to the first predicted trajectory as a stopping line.

15. An unmanned vehicle control device, characterized in that: include: An acquisition module, configured to acquire driving state information of the second unmanned vehicle; wherein the driving state information at least includes a second predicted trajectory of the second unmanned vehicle; A collision zone analysis module, configured to perform a collision zone analysis on the first unmanned vehicle and the second unmanned vehicle according to the first predicted trajectory of the first unmanned vehicle and the second predicted trajectory, and obtain a collision zone analysis result; a position relationship determination module, configured to respectively determine a first position relationship between the first unmanned vehicle and the collision zone and a second position relationship between the second unmanned vehicle and the collision zone when the collision zone analysis result indicates that the collision zone exists; A decision-making module is used to perform an action decision analysis on the first unmanned vehicle according to the first position relationship and the second position relationship, obtain a target action decision of the first unmanned vehicle, and control the first unmanned vehicle to perform an action according to the target action decision.

16. An unmanned vehicle, characterized in that: include: A control device for an unmanned vehicle as claimed in claim 15.

Images