Intelligent scheduling method and system of unmanned monorail hoist in well

Abstract

The present application relates to the technical field of data processing, and more particularly to an intelligent scheduling method and system for an unmanned single-track hoist in a mine. The method comprises the following steps: scanning the whole mine roadway by a patrol vehicle to construct a three-dimensional topological map; when the vehicle enters a blind area, a ground scheduling server generates a shadow vehicle according to the running state data of the vehicle, and generates an offline strategy package in combination with the historical running data of the vehicle under standard working conditions and the map attributes of the blind area section; the vehicle obtains a real trajectory after resuming communication when it drives out of the blind area; a vehicle aging factor and a corresponding position slope angle of the shadow vehicle are introduced to calculate a net acceleration to iteratively deduce the position of the shadow vehicle, and the total length of the safety envelope is synchronously calculated; the aging factor is corrected based on the residual error between the real trajectory and the deduced position, which is used to iteratively optimize the deduction of the shadow vehicle and safety warning when the vehicle enters the blind area next time. The present method can easily grasp the dynamic position of the vehicle during network interruption by constructing a shadow vehicle, and greatly improves the safety under complex terrain.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an intelligent scheduling method and system for an unmanned underground monorail crane. Background Technology

[0002] In modern coal mine auxiliary transportation systems, explosion-proof diesel-powered monorail cranes, with their strong load capacity, excellent climbing performance, and wide environmental adaptability, have become core equipment for underground material and personnel transportation. With the deepening of intelligent coal mine construction, achieving unmanned operation of monorail cranes and intelligent scheduling across the entire mine has become an inevitable trend for industry transformation and upgrading. Traditional monorail crane scheduling systems mostly rely on driver voice reports or fixed-point location monitoring based on RFID technology. This mode suffers from information lag and discontinuous coverage, completely failing to meet the real-time and continuous requirements of unmanned operations.

[0003] Currently, advanced monorail dispatching systems generally rely on underground industrial ring networks (such as WiFi 6 and 5G technologies) to achieve real-time transmission of vehicle location and operating status. However, due to the varied geological conditions of the tunnels, the signal shielding effect of complex metal support structures, and the high cost of building underground base stations, underground networks can never achieve complete coverage. For network blind spots, existing dispatching strategies typically adopt a conservative mechanism of area blocking. That is, once a vehicle enters a blind spot, the system loses its ability to detect the vehicle and forcibly blocks that section of the road, strictly prohibiting other vehicles from entering. This one-size-fits-all control method severely restricts transportation efficiency, easily leads to multi-vehicle congestion, and wastes underground transportation resources.

[0004] Furthermore, when the network connection is lost, existing technologies often use uniform linear extrapolation to predict vehicle positions. However, the movement of explosion-proof diesel engine monorails is greatly affected by factors such as roadway slope, exhibiting strong nonlinear characteristics. Especially on steep sections, simple linear extrapolation can lead to huge deviations between the predicted and actual positions (potentially tens of meters), which can easily cause rear-end collisions or scheduling path conflicts. At the same time, mines often contain vehicles of different service years. Due to wear and aging of the hydraulic systems and oil lines, the power response of older vehicles is completely different from that of new vehicles. The general prediction model cannot be adapted to every older vehicle, resulting in extremely poor control accuracy of the scheduling system for older vehicles. Summary of the Invention

[0005] To address the technical issues of incomplete network coverage in tunnels, which leads to blind spot blocking strategies restricting transportation efficiency, and the uniform linear extrapolation method used when the network is down being susceptible to large positional deviations due to tunnel slope, resulting in poor scheduling and control accuracy and high safety risks, this invention provides an intelligent scheduling method and system for underground unmanned monorail cranes.

[0006] In a first aspect, the present invention provides an intelligent scheduling method for an unmanned underground monorail crane, employing the following technical solution:

[0007] An intelligent scheduling method for an unmanned monorail crane in an underground mine includes the following steps:

[0008] A three-dimensional topological map is constructed by scanning all mine roadways using inspection vehicles;

[0009] Combining the historical operating data of the current vehicle under standard operating conditions with the map attributes of the blind spot road section, an offline strategy package is generated and sent to the current vehicle. When the current vehicle enters the blind spot, the offline strategy package is invoked to drive in the blind spot until the current vehicle leaves the blind spot and communication is restored, at which point the actual trajectory of the current vehicle in the blind spot is obtained.

[0010] When a vehicle enters a blind spot, the ground dispatch server instantiates a shadow vehicle object based on the vehicle's operating status data at that time. While the vehicle is in the blind spot, the net acceleration of the shadow vehicle is obtained based on the slope angle of its corresponding position in the 3D topology map. The shadow vehicle's position in the blind spot is then iteratively deduced based on the net acceleration and a preset time step. The total length of the shadow vehicle's safety envelope is obtained based on the slope angle of its corresponding position in the 3D topology map.

[0011] Based on the residual between the actual trajectory of the current vehicle in the blind spot and the projected position of the shadow vehicle in the blind spot, the aging factor of the vehicle is corrected. When the current vehicle enters the blind spot again, the dispatching system will use the corrected aging factor to iteratively project the real-time position of the shadow vehicle and implement risk warning based on the total length of the shadow vehicle's safety envelope.

[0012] The ingenuity of this invention lies in obtaining the net acceleration of the shadow vehicle based on the slope angle of its corresponding position in the 3D topology map when the vehicle is in a blind spot. Then, based on this net acceleration and a preset time step, the position of the shadow vehicle in the blind spot is nonlinearly iterated, restoring the physical characteristics of vehicle deceleration uphill or acceleration downhill. This greatly improves the accuracy of position prediction within the blind spot and ensures that the dispatch system can maintain continuous tracking of vehicle status even during network outages. Next, based on the slope angle of the shadow vehicle's corresponding position in the 3D topology map, the total length of the shadow vehicle's safety envelope is obtained, reserving more safety redundancy for steeper road sections with higher uncertainty, effectively preventing collisions. Finally, based on the residual between the current vehicle's actual trajectory in the blind spot and the shadow vehicle's projected position in the blind spot, the vehicle's aging factor is corrected. This allows for adaptive learning of the vehicle's performance degradation patterns, ensuring that the power model always matches the vehicle's current actual state and improving the accuracy of subsequent vehicle position projections in the blind spot.

[0013] Preferably, the step of combining the historical operating data of the current vehicle under standard operating conditions with the map attributes of blind spot road sections to generate an offline strategy package and send it to the current vehicle includes:

[0014] Historical operating data of the current vehicle under standard operating conditions is collected, including throttle opening command, hydraulic pump pressure, and actual acceleration. By performing multinomial regression analysis on the historical operating data, the basic dynamic response function of the current vehicle is fitted. Based on the map attributes of the blind spot road segment, including road slope and curve curvature, and combined with the basic dynamic response function of the current vehicle, the optimal speed planning curve and engine throttle opening curve of the current vehicle in the blind spot are obtained through dynamic simulation and optimization algorithms. The optimal speed planning curve and engine throttle opening curve are sent as offline strategy packages to the vehicle's on-board terminal and stored.

[0015] By using pre-generated offline strategy packages, vehicles can continue to operate at pre-planned speeds and throttles even when the network connection is lost, ensuring the predictability of blind spot operation.

[0016] Preferably, obtaining the actual trajectory of the current vehicle in the blind spot includes:

[0017] While the vehicle is running in the mine, the network signal strength and positioning coordinates of the on-board terminal are collected at each moment. If the network signal strength of the on-board terminal is less than the preset safety signal threshold at any moment, or if the positioning coordinates of the on-board terminal are in the blind spot warning buffer zone marked on the high-precision map at that moment, it is determined that the vehicle has entered the blind spot. The offline autonomous operation mode is immediately triggered. The on-board controller calls the pre-stored offline strategy package and drives in the blind spot according to the optimal speed planning curve and the engine throttle opening curve until the network signal strength of the on-board terminal is greater than or equal to the preset safety signal threshold. Then the vehicle drives out of the blind spot, exits the offline autonomous operation mode, and obtains the real trajectory of the vehicle in the blind spot.

[0018] Preferably, when the current vehicle enters the blind spot, the ground dispatch server instantiates a shadow vehicle object based on the current vehicle's operating status data when it enters the blind spot, including:

[0019] When the vehicle enters the blind spot, the on-board terminal immediately collects the vehicle's operating status data and sends the operating status data to the ground dispatch server with the highest priority. After receiving the operating status data, the ground dispatch server instantiates a shadow vehicle object in the digital twin system and uses the operating status data as the shadow vehicle's operating status data at the first moment.

[0020] The operational status data includes: vehicle position, speed, load capacity, and engine throttle opening.

[0021] Instantiating a shadow car object makes it easier to deduce the shadow car's position within the blind spot later.

[0022] Preferably, obtaining the net acceleration of the shadow vehicle based on the slope angle of its corresponding position in the 3D topological map includes:

[0023] By inputting the engine throttle opening of the shadow car at time k into the basic dynamic response function of the current vehicle, the theoretical thrust generated by the engine of the shadow car at time k can be obtained.

[0024] ;

[0025] In the formula, This represents the net acceleration of the shadow car at time k. This represents the basic frictional resistance of the current vehicle; The slope angle represents the position of the shadow car at time k in the 3D topological map. Represents gravitational acceleration; This represents the load mass of the shadow car at the k-th moment; This represents the aging factors of the current vehicle; This represents the theoretical thrust generated by the engine of the shadow car at time k.

[0026] Preferably, the iterative deduction of the shadow vehicle's position in the blind spot based on the net acceleration and a preset time step includes:

[0027] , This represents the predicted position of the shadow car at time k+1. This represents the predicted position of the shadow car at the k-th moment; This represents the speed of the shadow car at the k-th moment; This represents the net acceleration of the shadow car at time k. This represents the preset time step.

[0028] by The projected position of the shadow vehicle is calculated iteratively at intervals until the network signal strength of the vehicle's on-board terminal is greater than or equal to a preset safety signal threshold. At this point, the vehicle leaves the blind spot, and the projected position of the shadow vehicle at each moment in the blind spot is obtained.

[0029] This ensures that the dispatch system can maintain continuous tracking of vehicle status even during a moment of network outage.

[0030] Preferably, obtaining the total length of the safety envelope of the shadow vehicle based on the slope angle of its corresponding position in the 3D topology map includes:

[0031] ;

[0032] In the formula, This represents the total length of the safety envelope of the shadow car at time k. Represents the physical length of the monorail locomotive; The slope angle represents the position of the shadow car at time t in the 3D topological map. This represents the preset time step. This represents the number of moments after the shadow car enters the blind spot; This represents the preset slope sensitivity coefficient; This represents the pre-set reliability coefficient; || represents the positional error per second when a vehicle is traveling on a level, straight road; || represents the absolute value symbol.

[0033] The safety envelope that dynamically changes with the slope can reserve more safety redundancy for steep slope sections with higher uncertainty, effectively preventing collisions.

[0034] Preferably, the step of correcting the vehicle's aging factor based on the residual between the actual trajectory of the current vehicle in the blind spot and the projected position of the shadow vehicle in the blind spot includes:

[0035] , This represents the current vehicle's corrected aging factor; This represents the aging factors of the current vehicle; This represents the preset learning rate; Represents the preset hyperparameters; This represents the projected position of the shadow car at time j in the blind spot; This represents the actual position of the vehicle in the blind spot at the j-th moment. N represents the square of the throttle opening of the current vehicle at the j-th moment in the blind spot; N represents the total number of moments when the current vehicle and the shadow vehicle are traveling in the blind spot.

[0036] Preferably, when the current vehicle enters the blind spot again, the dispatching system will use a modified aging factor to iteratively extrapolate the real-time position of the shadow vehicle, and implement a risk warning based on the total length of the shadow vehicle's safety envelope, including:

[0037] When the current vehicle enters the blind spot again, the dispatch system will use the current vehicle's corrected aging factor to iteratively extrapolate the real-time position of the shadow vehicle. During the extrapolation, the system will implement risk warnings based on the total length of the shadow vehicle's safety envelope at each moment. If it is determined that the safety envelope overlaps with the safety envelopes of other vehicles in the blind spot, the system will immediately trigger a vehicle control command to force the vehicle to stop and avoid a collision.

[0038] Secondly, the present invention provides an intelligent scheduling system for an unmanned underground monorail crane, which adopts the following technical solution:

[0039] An intelligent scheduling system for an unmanned underground monorail crane includes a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned intelligent scheduling method for an unmanned underground monorail crane is implemented.

[0040] By adopting the above technical solution, a computer program is generated from the above-mentioned intelligent scheduling method for an unmanned underground monorail crane, and stored in a memory for loading and execution by a processor. Terminal equipment is then created based on the memory and processor for convenient use.

[0041] This invention has the following technical effects: First, when a vehicle is in a blind spot, the invention acquires the net acceleration of the shadow vehicle and performs nonlinear iteration on the shadow vehicle's position in the blind spot based on the net acceleration and a preset time step, greatly improving the accuracy of position prediction in the blind spot and ensuring that the scheduling system can maintain continuous tracking of the vehicle's status even during network outages. Second, based on the slope angle of the shadow vehicle's corresponding position in the 3D topology map, the total length of the shadow vehicle's safety envelope is acquired, which can reserve more safety redundancy for steep slope sections with higher uncertainty and effectively prevent collisions. Finally, based on the residual between the current vehicle's actual trajectory in the blind spot and the shadow vehicle's projected position in the blind spot, the vehicle's aging factor is corrected, which can adaptively learn the vehicle's performance degradation law, ensuring that the power model always fits the vehicle's current actual state and improving the accuracy of subsequent vehicle position projections in the blind spot. Attached Figure Description

[0042] Figure 1 This is a flowchart of an intelligent scheduling method for an unmanned underground monorail crane according to an embodiment of the present invention;

[0043] Figure 2 A comparative diagram showing the trajectory of a monorail crane as it enters a network blind spot and traverses a steep slope.

[0044] Figure 3 A diagram illustrating the real-time comparison of location prediction errors. Detailed Implementation

[0045] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0046] This invention discloses an intelligent scheduling method for an unmanned underground monorail crane, referring to... Figure 1 This includes steps S1-S4:

[0047] S1: Based on the inspection vehicle, scan all mine roadways to construct a three-dimensional topological map.

[0048] It should be noted that the system first uses an inspection vehicle equipped with a high-precision lidar and an inertial measurement unit (IMU) to scan the entire mine roadway, construct a three-dimensional topology map, and discretize the continuous roadway trajectory into a series of control node sequences with unique indices. For each node, the system acquires positioning coordinate data and slope angle, and the control coordinate data facilitates positioning.

[0049] S2: Combining the vehicle's historical operating data under standard operating conditions with the map attributes of the blind spot road section, an offline strategy package adapted to the blind spot road conditions is generated in advance. After the vehicle enters the blind spot, the offline strategy package is called to drive in the blind spot until the vehicle leaves the blind spot and communication is restored. The real trajectory of the vehicle in the blind spot is obtained. When the vehicle enters the blind spot, the ground dispatch server instantiates a shadow vehicle object based on the vehicle's operating status data before entering the blind spot.

[0050] It should be noted that for each in-service explosion-proof diesel engine monorail crane, a basic dynamic response function is obtained. Based on the basic dynamic response function and the map attributes of the blind spot road section, a dual offline strategy for speed and throttle that adapts to the blind spot road conditions and vehicle characteristics is generated in advance, so that the vehicle can drive autonomously in the blind spot according to the strategy. After obtaining the vehicle's real trajectory in the blind spot, it is convenient for subsequent analysis.

[0051] In this embodiment of the invention, historical operating data (including throttle opening command, hydraulic pump pressure, and actual acceleration) of each vehicle (each in-service explosion-proof diesel engine monorail gantry crane) under standard operating conditions are collected. Through polynomial regression analysis, the basic dynamic response function of each vehicle is fitted. This function describes the state of the throttle opening at a certain value. At that time, the thrust generated by the engine under ideal conditions;

[0052] The ground dispatch server retrieves the map attributes of the blind spot road segment in advance (including key terrain parameters such as road segment slope and curve curvature), and combines them with the current vehicle's basic dynamic response function. Through dynamic simulation and optimization algorithms, it obtains the optimal speed planning curve of the current vehicle in the blind spot, and derives the engine throttle opening curve. The optimal speed planning route and engine throttle opening curve are then sent as an offline strategy package to the vehicle's on-board terminal and stored.

[0053] The preset safety signal threshold is -85dBm. In other embodiments, the implementer can preset the safety signal threshold according to the specific implementation situation.

[0054] While the vehicle is running in the mine, the network signal strength (RSSI) and positioning coordinates of the on-board terminal are collected in real time at each moment (once per second). If the network signal strength of the on-board terminal is less than the preset safety signal threshold at any moment, or if the positioning coordinates of the on-board terminal are in the blind zone warning buffer zone marked on the high-precision map at that moment, the vehicle enters the blind zone and immediately triggers the offline autonomous operation mode. The on-board controller calls the pre-stored offline strategy package and drives in the blind zone according to the optimal speed planning curve and its corresponding engine throttle opening curve until the network signal strength of the on-board terminal is greater than or equal to the preset safety signal threshold. Then the vehicle drives out of the blind zone, exits the offline autonomous operation mode, resumes real-time communication with the ground dispatch center, and obtains the actual trajectory of the vehicle in the blind zone.

[0055] It should be noted that the optimal speed planning curve clearly defines the recommended operating speed of the vehicle at various positions within the blind spot. For example, for steep slopes, it plans a speed strategy of accelerating uphill before the steep slope; for downhill sections, it plans a speed strategy of decelerating and controlling speed before the downhill section, ensuring that the vehicle operates efficiently and safely in complex terrain within the blind spot. The engine throttle opening curve corresponding to the optimal speed planning curve specifies the engine throttle opening required for the vehicle to reach the corresponding recommended operating speed at various positions within the blind spot.

[0056] It should be noted that the accurate initialization of the shadow vehicle is completed before the vehicle enters the blind spot, providing a reliable initial state benchmark for subsequent blind spot trajectory simulation.

[0057] In this embodiment of the invention, when the current vehicle enters the blind spot, the vehicle terminal immediately collects the current vehicle's operating status data and sends the operating status data to the ground dispatch server with the highest priority. After receiving the instantaneous operating data, the ground dispatch server instantiates a shadow vehicle object in the digital twin system and uses the operating status data as the shadow vehicle's operating status data at the first moment.

[0058] The operational status data includes: the current vehicle position (the current vehicle's location coordinates in the three-dimensional topology map), speed, load capacity, and engine throttle opening.

[0059] S3: When the vehicle is in the blind spot, the ground dispatch server introduces the vehicle's aging factor and the slope angle corresponding to the position of the shadow vehicle in the three-dimensional topology map, calculates the net acceleration of the shadow vehicle, and iteratively infers the shadow vehicle's position in the blind spot based on the net acceleration. At the same time, it calculates the total length of the shadow vehicle's safety envelope based on the slope angle corresponding to the position of the shadow vehicle in the three-dimensional topology map.

[0060] It should be noted that during the period when the vehicle is in the blind spot, the ground dispatch uses the adaptive extended Kalman filter algorithm to iteratively extrapolate the state of the shadow vehicle at a preset time step. The traditional Kalman filter algorithm assumes that the vehicle is moving at a constant speed or with a fixed acceleration for a short period of time. This linear or constant model has a small error on straight and constant speed road sections, but it will have a large error on road sections with large gradient changes (such as steep slopes and downhill slopes). For example, when going uphill, the vehicle will actually decelerate, but the linear model will extrapolate at a constant speed, resulting in the extrapolated position being ahead of the actual position. When going downhill, the vehicle will actually accelerate, but the linear model will still extrapolate at a constant speed, resulting in the extrapolated position being behind the actual position. Therefore, linear extrapolation cannot adapt to deceleration on steep slopes.

[0061] To address the issue that linear extrapolation cannot adapt to deceleration on steep slopes, this invention introduces the slope angle corresponding to the location of the shadow vehicle on a three-dimensional topological map and the current vehicle's aging factor. By fusing these two parameters, the net acceleration of the shadow vehicle is obtained. This net acceleration can accurately reflect the physical phenomena of deceleration on uphill slopes or acceleration on downhill slopes. Then, based on the displacement increment term calculated from the net acceleration, the extrapolated position of the shadow vehicle under constant speed conditions is dynamically corrected, thereby significantly improving the extrapolation accuracy of the shadow vehicle's trajectory in blind spot environments.

[0062] In this embodiment of the invention, the net acceleration of the shadow car at the k-th time moment is obtained:

[0063] ;

[0064] In the formula, This represents the net acceleration of the shadow car at time k. This represents the basic frictional resistance of the current vehicle; The slope angle represents the position of the shadow car at time k in the 3D topological map. Represents gravitational acceleration; This represents the load mass of the shadow car at the k-th moment; This represents the aging factors of the current vehicle; The thrust theoretically generated by the engine of the shadow car at time k; gravitational acceleration. It is a physical constant, with the preset gravitational acceleration. Preset the basic frictional resistance of the current vehicle The preset aging factor for the current vehicle is: ;

[0065] By inputting the engine throttle opening of the shadow car at time k into the basic dynamic response function of the current vehicle, the theoretical thrust generated by the engine of the shadow car at time k can be obtained.

[0066] The effective thrust of the vehicle forward is represented by the aging factor. The greater the theoretical thrust of the engine and the aging factor, the greater the effective thrust of the vehicle forward and the greater the net acceleration of the shadow car. It is known that the engine will experience performance degradation after long-term operation. The aging factor of the current vehicle is used to correct the theoretical thrust of the engine to avoid the distortion of thrust calculation caused by engine performance degradation, thereby improving the accuracy of shadow car state inference.

[0067] This represents the term of gravitational drag, which occurs when the shadow car is driven uphill. At that time, the gravitational drag term is positive, resulting in Decreasing or even turning negative (deceleration) corresponds to the deceleration of a real vehicle going uphill; when the shadow car is simulated to go downhill, that is... At that time, the gravitational drag term is negative, resulting in Increased, corresponding to the actual acceleration of a vehicle going downhill;

[0068] The friction resistance term reflects the fixed resistance such as tire-ground friction and mechanical transmission friction when the vehicle is moving. The larger the value, the stronger the offsetting effect on the effective thrust, resulting in a smaller net acceleration. The shadow car will simulate the state of a real vehicle decelerating due to increased friction.

[0069] In this embodiment of the invention, the predicted position of the shadow car at the (k+1)th time is obtained:

[0070] ;

[0071] In the formula, This represents the predicted position of the shadow car at time k+1. This represents the predicted position of the shadow car at the k-th moment; This represents the speed of the shadow car at the k-th moment; This represents the net acceleration of the shadow car at time k. Representing a preset time step, in this embodiment of the invention, the preset time step... Second;

[0072] The position represents the deduced position under the condition of uniform velocity; The displacement increment term represents the net acceleration of the shadow car at time k.

[0073] when When the value is less than 0, the vehicle is in a situation of climbing and decelerating, which means that the estimated position under constant speed conditions is greater than the actual position. If the value is less than 0, the displacement increment term generated by the net acceleration of the shadow car at the kth moment is subtracted from the calculated position under the constant speed condition, and the calculated position of the shadow car at the (k+1)th moment is taken as the calculated position of the shadow car.

[0074] when When the value is greater than 0, the vehicle is accelerating downhill, meaning the projected position under constant speed is shorter than the actual position. If the value is greater than 0, the displacement increment term generated by the net acceleration of the shadow car at the kth moment is added to the predicted position under the constant speed condition, and this is used as the predicted position of the shadow car at the (k+1)th moment.

[0075] by The projected position of the shadow vehicle is calculated iteratively at intervals until the network signal strength of the vehicle's on-board terminal is greater than or equal to a preset safety signal threshold. At this point, the vehicle leaves the blind zone, resumes real-time communication with the ground dispatch center, and obtains the projected position of the shadow vehicle in the blind zone at each moment.

[0076] It should be noted that when extrapolating the position of the shadow vehicle in the blind spot, the extrapolated position of the shadow vehicle at the first moment is the position of the vehicle in the running status data of the shadow vehicle at the first moment. When extrapolating the position of the shadow vehicle in the blind spot, the instantaneous speed and engine throttle opening parameters of the shadow vehicle are executed synchronously with the current vehicle according to the optimal speed planning curve in the offline strategy package. The optimal speed planning curve predefines the recommended speed corresponding to each position in the blind spot. During the extrapolation process, the system will match the corresponding speed and the corresponding engine throttle opening from the optimal speed planning curve according to the current position of the shadow vehicle, to ensure that the extrapolation state is consistent with the offline running state of the real vehicle.

[0077] It should be noted that cumulative errors inevitably occur during the state simulation of the shadow vehicle in the blind spot, and there may be unknown risks such as other operating vehicles and obstacles in the blind spot. Therefore, the dispatching system cannot simplify the vehicle into a single-point model and rely solely on a single simulated position for dispatching decisions, as this may lead to collisions with other operating vehicles. Therefore, this invention defines the vehicle's occupancy state in the 3D topology map as a dynamically expanding and contracting interval that varies with the environment, and uses this interval to construct the total length of the shadow vehicle's safety envelope at each moment. For example, when the simulation shows that the shadow vehicle's safety envelope overlaps with the protection area of ​​other operating vehicles, the system can determine the collision risk in advance and immediately trigger warning commands such as deceleration and stopping. For special road sections with blind spots such as steep slopes and curves, sufficient braking distance can be reserved to avoid accidents caused by braking delays due to terrain.

[0078] It should be further explained that, as is known, the divergence rate of shadow vehicle inference error varies significantly under different road conditions. On straight road sections, the vehicle's motion is stable and the position prediction error is small, while on steep road sections, the position uncertainty increases sharply. Therefore, based on this characteristic, this invention proposes a method for obtaining the total length of the safety envelope, so that the growth rate of the safety envelope is positively correlated with the current absolute value of the slope, thereby matching the error divergence pattern under different road conditions.

[0079] In this embodiment of the invention, the total length of the safety envelope of the shadow car at the k-th moment is obtained:

[0080] ;

[0081] In the formula, This represents the total length of the safety envelope of the shadow car at time k. Represents the physical length of the monorail locomotive; The slope angle represents the position of the shadow car at time t in the 3D topological map. || represents the preset time step; || represents the absolute value sign. This represents the number of moments after the shadow car enters the blind spot; This represents the preset slope sensitivity coefficient, used to adjust the weight of the slope's influence on the error. This represents the preset confidence coefficient. The higher the confidence coefficient, the higher the safety redundancy but the lower the scheduling efficiency. The lower the confidence coefficient, the higher the scheduling efficiency but the increased risk. This represents the positional error per second that a vehicle produces while traveling on a level, straight road.

[0082] In this embodiment of the invention, a slope sensitivity coefficient is preset. Preset reliability coefficient , preset In other embodiments, the implementer can preset the speed (meters per second) according to the specific implementation situation. , as well as The value of .

[0083] The larger the value, the more likely the vehicle is going uphill or downhill. A value greater than 1 indicates that the error in location prediction should be increased, and the safety envelope length needs to be increased. When the value approaches 0, it indicates that the vehicle is on a level road. The closer the value is to 1, the more the error accumulates according to the position error generated per second when the vehicle is traveling on a level straight road; This represents the cumulative process of positional error over time.

[0084] S4: Based on the vehicle's actual trajectory in the blind spot and the residual between the shadow vehicle's estimated position at each moment in the blind spot, the vehicle's aging factor is corrected; when the vehicle enters the blind spot again, the scheduling system will use the corrected aging factor to iteratively estimate the shadow vehicle's real-time position and implement risk warning based on the total length of the shadow vehicle's safety envelope.

[0085] It should be noted that the aging factor is used to characterize the degree of performance degradation of the engine due to long-term operation, which directly determines the effective output of engine thrust. However, the performance degradation of the engine is irreversible. As the running time increases, the aging factor of the current vehicle will gradually decrease. Therefore, when the vehicle enters the blind spot, if the dispatch system only relies on the preset aging factor to extrapolate the trajectory of the shadow vehicle each time, it will eventually lead to the power model gradually becoming disconnected from the actual power characteristics of the vehicle.

[0086] Therefore, when a vehicle leaves the network blind spot, the system first obtains the vehicle's real position at each moment in the blind spot and compares it with the estimated position of the shadow vehicle at the same moment. The residual generated by the comparison is used to correct the vehicle's aging factor, so that the corrected aging factor can be used when the vehicle enters the next blind spot.

[0087] In this embodiment of the invention, the real position of the current vehicle in the blind spot at each moment is obtained based on the real trajectory of the current vehicle in the blind spot, and the time interval of the real position is consistent with the time step of the shadow car simulation.

[0088] Get the corrected aging factor for the current vehicle:

[0089] ;

[0090] In the formula, This represents the current vehicle's corrected aging factor; This represents the aging factors of the current vehicle; This represents the preset learning rate, used to control the step size of parameter updates and prevent numerical oscillations. This represents the preset hyperparameters used to prevent the denominator from being zero; This represents the projected position of the shadow car at time j in the blind spot; This represents the actual position of the vehicle in the blind spot at the j-th moment. represents the square of the throttle opening of the current vehicle at the j-th moment in the blind spot; N represents the total number of moments when the current vehicle and the shadow vehicle are traveling in the blind spot; in this embodiment of the invention, the preset learning rate is... In other embodiments, implementers may pre-set according to specific implementation conditions. The value;

[0091] This represents the position prediction error at time j within the blind zone; This represents the sum of position prediction errors at all times within the blind zone. A positive value indicates that the vehicle's power performance was overestimated when the position was predicted, and the aging factor of the vehicle needs to be reduced. A negative value indicates that the vehicle's power performance was underestimated when the position was predicted, and the aging factor of the vehicle needs to be increased.

[0092] A larger square of the throttle opening at time j in the blind spot indicates a more stable powertrain performance and relatively smaller impact of external disturbances (such as wind resistance) on position. However, the resulting error may contain more noise. The larger denominator in the formula results in a smaller correction amount, preventing the system from drastically modifying parameters due to accidental errors caused by a single aggressive driving maneuver.

[0093] If the square of the throttle opening at time j in the blind spot is smaller, and a large positional error occurs at this time, it indicates that the vehicle's power performance has deteriorated. The smaller denominator in the equation results in a larger correction amount, causing the system to significantly adjust the aging factor.

[0094] In this embodiment of the invention, when the current vehicle enters the blind spot again, the scheduling system will use the current vehicle's corrected aging factor to iteratively extrapolate the real-time position of the shadow vehicle. During the extrapolation process, the system will implement risk warning based on the total length of the shadow vehicle's safety envelope at each moment. If it is determined that the safety envelope overlaps spatially with the safety envelopes of other vehicles in the blind spot, the system will immediately trigger a vehicle control command to force the vehicle to stop running and avoid a collision.

[0095] Figure 2 This diagram illustrates the trajectory comparison of a monorail crane entering a network blind zone and traversing a steep slope. The actual vehicle trajectory shows that, upon entering the steep slope, the slope decreases due to gravity, indicating a genuine physical deceleration of the vehicle. Linear extrapolation based on the speed before entering the blind zone leads to a straight, continuously rising trajectory, significantly deviating from the actual position. The shadow trajectory of this invention, thanks to the AEKF algorithm's fusion of slope data, accurately curves downwards on the steep slope. Furthermore, the dynamic safety envelope region encompasses the extrapolated trajectory, and its width increases significantly and non-linearly when traversing steep slopes, visually demonstrating the invention's scheduling strategy of automatically expanding safety margins on complex road sections.

[0096] Figure 3The diagram showing the real-time comparison of position prediction errors reflects the changes in the absolute position error of the prior art and the present invention within the blind zone over time. The dashed line representing the error of the prior art exhibits a large-scale wave-like divergence trend with a huge peak error. In contrast, the solid line representing the error of the present invention consistently runs at a low level at the bottom, demonstrating that the algorithm continuously combats minor model mismatches and possesses extremely high robustness.

[0097] This invention also discloses an intelligent scheduling system for an unmanned underground monorail crane, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent scheduling method for an unmanned underground monorail crane provided by this invention is implemented.

[0098] The system also includes other components well-known to those skilled in the art, such as communication buses and communication interfaces, the setup and functions of which are known in the art and will not be described in detail here. In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0099] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An intelligent scheduling method for an unmanned underground monorail crane, characterized in that, include: A three-dimensional topological map is constructed by scanning all mine roadways using inspection vehicles; Combining the historical operating data of the current vehicle under standard operating conditions with the map attributes of the blind spot road section, an offline strategy package is generated and sent to the current vehicle. When the current vehicle enters the blind spot, the offline strategy package is invoked to drive in the blind spot until the current vehicle leaves the blind spot and communication is restored, at which point the actual trajectory of the current vehicle in the blind spot is obtained. When a vehicle enters a blind spot, the ground dispatch server instantiates a shadow vehicle object based on the vehicle's operating status data at that time. While the vehicle is in the blind spot, the net acceleration of the shadow vehicle is obtained based on the slope angle of its corresponding position in the 3D topology map. The shadow vehicle's position in the blind spot is then iteratively deduced based on the net acceleration and a preset time step. The total length of the shadow vehicle's safety envelope is obtained based on the slope angle of its corresponding position in the 3D topology map. Based on the residual between the actual trajectory of the current vehicle in the blind spot and the projected position of the shadow vehicle in the blind spot, the aging factor of the vehicle is corrected. When the current vehicle enters the blind spot again, the dispatching system will use the corrected aging factor to iteratively project the real-time position of the shadow vehicle and implement risk warning based on the total length of the shadow vehicle's safety envelope.

2. The intelligent scheduling method for an unmanned monorail crane in an underground mine according to claim 1, characterized in that, The process of combining the vehicle's historical operating data under standard conditions with the map attributes of blind spot road sections to generate an offline strategy package and send it to the current vehicle includes: Historical operating data of the current vehicle under standard operating conditions is collected, including throttle opening command, hydraulic pump pressure, and actual acceleration. By performing multinomial regression analysis on the historical operating data, the basic dynamic response function of the current vehicle is fitted. Based on the map attributes of the blind spot road segment, including road slope and curve curvature, and combined with the basic dynamic response function of the current vehicle, the optimal speed planning curve and engine throttle opening curve of the current vehicle in the blind spot are obtained through dynamic simulation and optimization algorithms. The optimal speed planning curve and engine throttle opening curve are sent as offline strategy packages to the vehicle's on-board terminal and stored.

3. The intelligent scheduling method for an unmanned underground monorail crane according to claim 1 or 2, characterized in that, The process of obtaining the current vehicle's true trajectory in the blind spot includes: While the vehicle is running in the mine, the network signal strength and positioning coordinates of the on-board terminal are collected at each moment. If the network signal strength of the on-board terminal is less than the preset safety signal threshold at any moment, or if the positioning coordinates of the on-board terminal are in the blind spot warning buffer zone marked on the high-precision map at that moment, it is determined that the vehicle has entered the blind spot. The offline autonomous operation mode is immediately triggered. The on-board controller calls the pre-stored offline strategy package and drives in the blind spot according to the optimal speed planning curve and the engine throttle opening curve until the network signal strength of the on-board terminal is greater than or equal to the preset safety signal threshold. Then the vehicle drives out of the blind spot, exits the offline autonomous operation mode, and obtains the real trajectory of the vehicle in the blind spot.

4. The intelligent scheduling method for an unmanned underground monorail crane according to claim 1, characterized in that, When the current vehicle enters the blind spot, the ground dispatch server instantiates a shadow vehicle object based on the current vehicle's operating status data when it enters the blind spot, including: When the vehicle enters the blind spot, the on-board terminal immediately collects the vehicle's operating status data and sends the operating status data to the ground dispatch server with the highest priority. After receiving the operating status data, the ground dispatch server instantiates a shadow vehicle object in the digital twin system and uses the operating status data as the shadow vehicle's operating status data at the first moment. The operational status data includes: vehicle position, speed, load capacity, and engine throttle opening.

5. The intelligent scheduling method for an unmanned underground monorail crane according to claim 1, characterized in that, The process of obtaining the net acceleration of the shadow vehicle based on the slope angle of its corresponding position in the 3D topological map includes: By inputting the engine throttle opening of the shadow car at time k into the basic dynamic response function of the current vehicle, the theoretical thrust generated by the engine of the shadow car at time k can be obtained. ； In the formula, This represents the net acceleration of the shadow car at time k. This represents the basic frictional resistance of the current vehicle; The slope angle represents the position of the shadow car at time k in the 3D topological map. Represents gravitational acceleration; This represents the load mass of the shadow car at the k-th moment; This represents the aging factor of the current vehicle; This represents the theoretical thrust generated by the engine of the shadow car at time k.

6. The intelligent scheduling method for an unmanned underground monorail crane according to claim 1, characterized in that, The iterative deduction of the shadow vehicle's position in the blind spot based on the net acceleration and a preset time step includes: , This represents the predicted position of the shadow car at time k+1. This represents the predicted position of the shadow car at the k-th moment; This represents the speed of the shadow car at the k-th moment; This represents the net acceleration of the shadow car at time k. This represents the preset time step. by The projected position of the shadow vehicle is calculated iteratively at intervals until the network signal strength of the vehicle's on-board terminal is greater than or equal to a preset safety signal threshold. At this point, the vehicle leaves the blind spot, and the projected position of the shadow vehicle at each moment in the blind spot is obtained.

7. The intelligent scheduling method for an unmanned underground monorail crane according to claim 1, characterized in that, The step of obtaining the total length of the safety envelope of the shadow vehicle based on the slope angle of its corresponding position in the 3D topology map includes: ； In the formula, This represents the total length of the safety envelope of the shadow car at time k. Represents the physical length of the monorail locomotive; The slope angle represents the position of the shadow car at time t in the 3D topological map. This represents the preset time step. This represents the number of moments after the shadow car enters the blind spot; This represents the preset slope sensitivity coefficient; This represents the pre-set reliability coefficient; || represents the positional error per second when a vehicle is traveling on a level, straight road; || represents the absolute value symbol.

8. The intelligent scheduling method for an unmanned underground monorail crane according to claim 1, characterized in that, The method of correcting the vehicle's aging factor based on the residual between the actual trajectory of the current vehicle in the blind spot and the projected position of the shadow vehicle in the blind spot includes: , This represents the current vehicle's corrected aging factor; This represents the aging factor of the current vehicle; This represents the preset learning rate; Represents the preset hyperparameters; This represents the projected position of the shadow car at time j in the blind spot; This represents the actual position of the vehicle in the blind spot at the j-th moment. N represents the square of the throttle opening of the current vehicle at the j-th moment in the blind spot; N represents the total number of moments when the current vehicle and the shadow vehicle are traveling in the blind spot.

9. The intelligent scheduling method for an unmanned underground monorail crane according to claim 1, characterized in that, When the current vehicle enters the blind spot again, the dispatch system will use a modified aging factor to iteratively extrapolate the real-time position of the shadow vehicle and implement risk warnings based on the total length of the shadow vehicle's safety envelope, including: When the current vehicle enters the blind spot again, the dispatch system will use the current vehicle's corrected aging factor to iteratively extrapolate the real-time position of the shadow vehicle. During the extrapolation, the system will implement risk warnings based on the total length of the shadow vehicle's safety envelope at each moment. If it is determined that the safety envelope overlaps with the safety envelopes of other vehicles in the blind spot, the system will immediately trigger a vehicle control command to force the vehicle to stop and avoid a collision.

10. An intelligent dispatching system for an unmanned underground monorail crane, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement an intelligent scheduling method for an unmanned underground monorail crane according to any one of claims 1-9.

Images