A control method, device, system and medium of an unmanned mine car

Abstract

This invention discloses a control method, device, system, and medium for an unmanned mining truck, relating to the technical field of mining truck control methods. The method includes determining the bump pattern of the unmanned mining truck during movement based on multiple sub-movement items, the overall shape of the unmanned mining truck, and multiple attitude data; determining the transport control event of the unmanned mining truck based on the bump pattern, the storage distribution map of ore relative to the unmanned mining truck, and the driving speed of the unmanned mining truck; determining the dynamic control mode of the unmanned mining truck based on the stable transport coefficient, the current attitude and overall shape of the unmanned mining truck; identifying unfinished transport tasks based on the identification of the unmanned mining truck's task list; and determining the cooperative control mode of the unmanned mining truck based on the unfinished transport tasks, the ore load of the unmanned mining truck, and the current state of the remaining unmanned mining trucks, thereby improving the cooperativeness of the unmanned mining truck control mode.

Description

Technical Field

[0001] This invention relates to the technical field of mine car control methods, and more particularly to a control method, device, system, and medium for an unmanned mine car. Background Technology

[0002] With the development of technology, driverless mining trucks are being applied to ore transportation sites and are gradually becoming one of the main ore transportation tools. In existing technologies, collecting multiple attitude data points of the mining truck and determining its current attitude based on this data does not take into account the bumpy ride conditions during its journey. Ignoring the impact of these bumps reduces the accuracy of the truck's transport control events. Furthermore, the lack of a corresponding collaborative relationship between one mining truck and the remaining trucks affects the coordination of the driverless mining truck control mode. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a control method, device, system, and medium for unmanned mining trucks.

[0004] In a first aspect, the present invention provides a control method for an unmanned mining truck, comprising:

[0005] Collect the driving status of the unmanned mining truck, and determine the driving mode of the unmanned mining truck based on the driving status and the terrain corresponding to the driving path.

[0006] The driverless mining truck travels along this driving mode, and multiple sub-driving segments are determined based on the identification of the driving path; the load movement events of the driverless mining truck in each sub-driving segment are determined based on the current driving speed of the driverless mining truck and the multiple sub-driving segments.

[0007] Based on the multiple sub-movement items corresponding to the load movement event, the overall shape of the unmanned mining truck and multiple attitude data, the bump pattern of the unmanned mining truck during the movement process is determined. Based on the bump pattern, the storage distribution map of the ore relative to the unmanned mining truck and the driving speed of the unmanned mining truck, the transportation control event of the unmanned mining truck is determined.

[0008] The smooth transport coefficient of the unmanned mining truck is determined based on the identification of transport control events. The dynamic control mode of the unmanned mining truck is determined based on the smooth transport coefficient, the current attitude and overall shape of the unmanned mining truck, and multiple driving data of the unmanned mining truck are dynamically controlled.

[0009] Based on the identification of the task list of the unmanned mining truck, an incomplete transportation task is determined. The cooperative control mode of the unmanned mining truck is determined according to the incomplete transportation task, the ore load of the unmanned mining truck, and the current status of the remaining unmanned mining trucks.

[0010] Secondly, the present invention also provides a control device for an unmanned mining truck, wherein the control device is applied to the above-mentioned control method for an unmanned mining truck, and the control device for the unmanned mining truck includes:

[0011] The driving mode module is used to collect the driving status of the unmanned mining truck and determine the driving mode of the unmanned mining truck based on the driving status and the terrain corresponding to the driving path.

[0012] The load movement event module is used to determine multiple sub-driving segments based on the identification of the driving path when the unmanned mining truck travels along the driving mode; and to determine the load movement event of the unmanned mining truck in each sub-driving segment based on the current driving speed of the unmanned mining truck and the multiple sub-driving segments.

[0013] The transport control event module is used to determine the bump pattern of the unmanned mining truck during the movement process based on multiple sub-movement items corresponding to the load movement event, the overall shape of the unmanned mining truck and multiple attitude data, and to determine the transport control event of the unmanned mining truck based on the bump pattern, the storage distribution map of the ore relative to the unmanned mining truck and the driving speed of the unmanned mining truck.

[0014] The dynamic control module is used to determine the smooth transport coefficient of the unmanned mining truck based on the identification of transport control events, determine the dynamic control mode of the unmanned mining truck based on the smooth transport coefficient, the current attitude and overall shape of the unmanned mining truck, and dynamically control multiple driving data of the unmanned mining truck.

[0015] The collaborative control module is used to identify unfinished transport tasks based on the task list of the unmanned mining truck, and to determine the collaborative control mode of the unmanned mining truck based on the unfinished transport task, the ore load of the unmanned mining truck, and the current status of the remaining unmanned mining trucks.

[0016] Thirdly, the present invention also provides a control system for an unmanned mining truck, including the control device as described in the second aspect.

[0017] Fourthly, the present invention also provides a computer-readable storage medium storing computer-executable instructions for causing a computer to perform the control method for an unmanned mining vehicle as described in the first aspect.

[0018] Compared with the prior art, the beneficial effects of the present invention are:

[0019] 1. The method described in this invention determines the bump pattern of the unmanned mining truck during its movement based on multiple sub-movement items corresponding to the load movement events of the unmanned mining truck in each sub-driving section, the overall shape of the unmanned mining truck, and multiple attitude data. Based on this bump pattern, the storage distribution map of the ore relative to the unmanned mining truck, and the driving speed of the unmanned mining truck, the transportation control events of the unmanned mining truck are determined. Multiple sub-driving sections are introduced to further control the load movement events. This method incorporates a holistic consideration of the bump pattern, the storage distribution map of the ore relative to the unmanned mining truck, and the driving speed of the unmanned mining truck, thereby improving the accuracy of the transportation control events of the unmanned mining truck.

[0020] 2. Based on the identification of transport control events, the stable transport coefficient of the unmanned mining truck is determined. Based on this stable transport coefficient, the current attitude and overall shape of the unmanned mining truck, the dynamic control mode of the unmanned mining truck is determined, and multiple driving data of the unmanned mining truck are dynamically controlled. Based on the identification of the unmanned mining truck's task list, incomplete transport tasks are identified. Based on these incomplete transport tasks, the ore load of the unmanned mining truck, and the current state of the remaining unmanned mining trucks, the collaborative control mode of the unmanned mining truck is determined. The introduction of a dynamic control mode achieves a holistic consideration of incomplete transport tasks, the ore load of the unmanned mining truck, and the current state of the remaining unmanned mining trucks, improving the synergy of the unmanned mining truck control mode. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the control method for an unmanned mining truck in an embodiment of the present invention;

[0022] Figure 2 This is a flowchart illustrating step S11 in the control method for the unmanned mining truck of the present invention.

[0023] Figure 3 This is a flowchart illustrating step S12 in the control method for the unmanned mining truck of the present invention.

[0024] Figure 4 This is a flowchart illustrating step S13 in the control method for the unmanned mining truck of the present invention.

[0025] Figure 5 This is a flowchart illustrating step S14 in the control method for the unmanned mining truck of the present invention.

[0026] Figure 6 This is a flowchart illustrating step S15 in the control method for the unmanned mining truck of the present invention.

[0027] Figure 7 This is a schematic diagram of the control device for the unmanned mining truck in an embodiment of the present invention.

[0028] The attached diagram is labeled as follows:

[0029] 21. Driving Mode Module; 22. Load Movement Event Module; 23. Transport Control Event Module; 24. Dynamic Regulation Module; 25. Cooperative Control Module. Detailed Implementation

[0030] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0031] Please see Figures 1 to 6 A control method for an unmanned mining truck, applied to mining truck control scenarios; the control method for the unmanned mining truck includes:

[0032] Step S11: Collect the driving status of the unmanned mining truck, and determine the driving mode of the unmanned mining truck based on the driving status and the terrain corresponding to the driving path.

[0033] Step S12: The unmanned mining truck travels along the driving mode, and multiple sub-driving segments are determined based on the identification of the driving path; the load movement events of the unmanned mining truck in each sub-driving segment are determined based on the current driving speed of the unmanned mining truck and the multiple sub-driving segments.

[0034] Step S13: Based on the multiple sub-movement items corresponding to the load movement event, the overall shape of the unmanned mining truck and multiple attitude data, determine the bump pattern of the unmanned mining truck during the movement process, and based on the bump pattern, the storage distribution map of the ore relative to the unmanned mining truck and the driving speed of the unmanned mining truck, determine the transportation control event of the unmanned mining truck.

[0035] Step S14: Determine the smooth transport coefficient of the unmanned mining truck based on the identification of transport control events, determine the dynamic control mode of the unmanned mining truck based on the smooth transport coefficient, the current attitude and overall shape of the unmanned mining truck, and dynamically control multiple driving data of the unmanned mining truck.

[0036] Step S15: Based on the identification of the task list of the unmanned mining truck, determine the unfinished transportation task, and determine the cooperative control mode of the unmanned mining truck according to the unfinished transportation task, the ore load of the unmanned mining truck and the current status of the remaining unmanned mining trucks.

[0037] refer to Figure 2 In step S11, the specific steps are as follows:

[0038] S111: Real-time monitoring of the autonomous driving process of the unmanned mining truck, while the unmanned mining truck is carrying ore; collecting multiple driving data and ore load data of the unmanned mining truck; determining the driving status of the unmanned mining truck based on the multiple driving data and ore load data of the unmanned mining truck.

[0039] S112: Determine the corresponding driving path based on the identification of the unmanned mining truck, determine the terrain shape corresponding to the driving path based on the driving path, the location of the unmanned mining truck and the site distribution map of the ore transportation site, and determine the driving mode of the unmanned mining truck based on the driving status of the unmanned mining truck, the corresponding driving speed and the terrain shape corresponding to the driving path.

[0040] In the embodiments of this application, during the data acquisition phase, the system acquires two types of key information in parallel. The first type of key information is the vehicle's own driving data, which forms the basis for describing the vehicle's motion state. Inertial measurement units installed near the center of mass acquire three-axis acceleration and angular velocity in real time. Wheel speed encoders on each wheel provide high-precision rotational speed information for calculating vehicle speed and monitoring slip. A high-precision GNSS / RTK system provides centimeter-level absolute position and velocity as a global positioning reference. At the same time, steering angle sensors and suspension displacement sensors capture changes in vehicle attitude caused by driving operation inputs and road surface excitations, respectively. By combining these sensor data, the system can construct a set of six-degree-of-freedom motion parameters, including position, velocity, attitude angle, angular velocity, and acceleration.

[0041] The second category of key information focuses on the ore load, which has a significant impact on dynamic characteristics. The system mainly uses indirect measurement methods, such as using air suspension pressure sensors combined with suspension stiffness models, to accurately estimate the total mass and the lateral position of the center of gravity. At the same time, it uses a 3D lidar installed above the carriage for direct scanning to generate a three-dimensional point cloud model of the ore pile. This not only calculates the three-dimensional coordinates of the total mass and the center of gravity, but more importantly, it generates an "ore storage distribution map" to intuitively quantify the ore accumulation pattern and provide a basis for assessing the risk of cargo displacement during transportation.

[0042] All raw, noisy sensor data is integrated into a unified, smooth, and highly reliable vehicle state representation, which can be achieved using extended Kalman filtering or unscented Kalman filtering algorithms. This algorithm establishes a state-space model that includes vehicle kinematics and dynamics, and uses observation data from various sensors to recursively correct the state predicted by the model.

[0043] The absolute position and velocity of GNSS / RTK can continuously correct the drift caused by the integration of the inertial measurement unit (IMU); wheel speed data can help improve the estimation accuracy of longitudinal velocity; suspension displacement data can be used to verify and calibrate the roll angle calculated by the IMU; the filter outputs a high-frequency updated "driving state vector", which integrates all key information such as the vehicle's position, velocity, attitude, and the mass and center of gravity of the load, providing comprehensive and accurate input for subsequent decision-making modules.

[0044] Furthermore, it receives high-precision vehicle location coordinates from S111 and compares them with a pre-built high-precision map of the ore transportation site; this map is not a simple road network, but a database of "driving routes" containing multiple topological connections, each route having a unique ID.

[0045] By employing particle filtering or Mahalanobis distance matching algorithms, the system comprehensively considers positioning errors and vehicle heading to "attach" the vehicle to a specific driving path with the highest probability, ultimately outputting a unique path ID and the vehicle's relative position on that path. After determining the driving path, the system will perform forward-looking terrain feature extraction on the road segment that the vehicle is about to enter based on this path.

[0046] Starting from the current location, the system will query a preset "look-ahead distance" and extract key attributes of all points along the path from the attribute layer of the high-precision map, such as slope, curvature, road surface smoothness index, road width, and special area markers. The system will generate a terrain morphology sequence describing the road conditions ahead, such as transitioning from "straight road surface" to "sharp turn + slight downhill" and then to "crossroads".

[0047] After completing the path and terrain analysis, the final driving mode decision-making stage begins. This is a process that comprehensively evaluates the vehicle's own "capabilities" and the "challenges" ahead, and is usually executed by a rule-based or fuzzy logic-based decision engine. The engine receives the vehicle status (especially the total mass and current speed) from S111 as input, as well as the terrain pattern sequence generated in the previous step.

[0048] The decision engine has an embedded rule base generated by expert experience or data. For example, when the vehicle is heavily loaded and there is a long uphill ahead, it will switch to "Power Mode"; when entering a curve or loading area, it will switch to "Precision Mode"; when the road surface is bumpy, it will switch to "Stable Mode". These modes correspond to different control strategies: in Power Mode, the power system responds more actively; in Precision Mode, the path tracking accuracy is prioritized and the vehicle speed is limited; in Stable Mode, the vehicle speed is actively reduced and the suspension is adjusted to minimize vibration. Through this logic, the system can select the most suitable driving mode for the upcoming road conditions.

[0049] refer to Figure 3 In step S12, the specific steps are as follows:

[0050] S121: Collect the driving mode, determine multiple sub-driving modes based on the recognition of the driving mode, determine the autonomous driving of the unmanned mining truck based on the multiple sub-driving modes, the current position of the unmanned mining truck and the corresponding mode trigger node, and the unmanned mining truck drives along the driving mode.

[0051] S122: Collect driving paths, identify multiple path nodes based on the identification of driving paths, identify multiple sub-driving road segments based on the node positions and shapes of multiple path nodes and the terrain shapes corresponding to the driving paths, monitor the current driving speed of unmanned mining trucks in real time, and determine the load movement events of unmanned mining trucks in each sub-driving road segment based on the current driving speed of unmanned mining trucks and multiple sub-driving road segments.

[0052] In the embodiments of this application, a high-level driving mode (such as "precision mode") is decomposed into multiple low-level sub-modes that are bound to specific driving scenarios. This process is usually implemented through a hierarchical state machine, in which the driving mode determined by S11 serves as the top-level state, and encapsulates a set of refined sub-states for specific scenarios. For example, in "precision mode", the system defines sub-modes such as "approaching a curve", "steady-state cornering", "exiting a curve", and "approaching a stopping point".

[0053] Based on the results of S112 forward-looking terrain recognition, the system pre-selects a sub-mode sequence to be activated, preparing for subsequent execution. After completing the mode decomposition, the system enters the mode triggering and execution phase, the core of which is an event-driven mechanism based on geofencing. The system defines virtual "mode trigger nodes" on the high-precision map at a certain distance before key terrain features (such as curve start points and intersection stop lines). By monitoring the vehicle's position in real time, once it is determined that the vehicle has crossed a trigger node, the system will immediately trigger a state transition event, switching the currently active sub-mode to the target sub-mode bound to that node. This automated triggering process forms the basis of "autonomous driving".

[0054] When switching sub-modes, the motion controller receives a new set of control commands and adjusts the control target accordingly. For example, when switching from cruise to approaching a curve, the controller will begin to execute a smooth deceleration curve. During the closed-loop control and state maintenance phase, the system continuously acquires the real-time state of the vehicle through S111 and compares it with the target state set in the current sub-mode, thereby generating an error signal. The motion controller continuously adjusts its output based on this error, forming a closed-loop feedback system to minimize tracking errors and ensure that the actual behavior of the vehicle is precisely consistent with the target set in the sub-mode, thus ensuring the continuity and predictability of the entire driving process.

[0055] For the scenario of an unmanned mining truck in a mining site, assuming the truck is carrying 88 tons and is currently in "precision mode," it is traveling in a straight line at 25 km / h on "Main Transportation Road - 03." There is a sharp right turn with a radius of 30 meters 300 meters ahead. After the system collects the data and finds that the top-level mode is "precision mode," it has predicted the sharp turn ahead through terrain recognition of S112 and has prepared two sub-modes internally: "approaching the curve" and "steady-state cornering." On the high-precision map, a virtual trigger node named "Trigger_Node_P03_Curve_Approach" is preset 80 meters away from the start of the curve.

[0056] When the real-time position of the mining truck shows that it has just passed this node, the system immediately triggers a state transition, switching the sub-mode from "cruising" to "approaching the curve". After the transition, the mining truck's model predictive controller (MPC) receives new instructions, and its optimization objective becomes to smoothly reduce the speed from 25 km / h to 12 km / h in the next 80 meters, while smoothly transitioning the vehicle to the "inner tangent" of the curve.

[0057] During the subsequent journey, the system uses closed-loop control to compare the deviation between the actual vehicle speed and the planned deceleration curve in real time, and continuously adjusts the braking force to ensure a smooth and shock-free deceleration process. When the vehicle reaches the starting point of the curve, another node is triggered, which switches the sub-mode from "approaching the curve" to "steady-state cornering". The controller then adjusts the target and begins to accurately track the curve at a constant speed of 12 km / h. The mining truck does not react hastily when it sees the curve, but begins to execute a carefully choreographed, multi-stage sequence of actions for "safe cornering" from a distance. Its behavior is highly structured and predictable.

[0058] Furthermore, the system discretizes and characterizes the path, transforming a smooth curve into a series of feature points with clear physical meaning, i.e., path nodes. This process is achieved through feature extraction algorithms based on curvature, slope, and topological relationships, identifying key points from the driving path on the high-precision map, such as curvature extrema (curve apex), curvature zero points (curve start and end points), slope inflection points, and intersection center points. The system outputs an ordered list of path nodes, with each node containing its location, shape, and key attributes.

[0059] After identifying these nodes, the system proceeds to the segment division and attribute assignment stage. It defines the path segment between two adjacent path nodes as a "sub-driving segment." Since the nodes are defined by features, the geometric characteristics within each segment are relatively uniform. The system calculates or queries a set of standardized attributes for each sub-driving segment, including segment length, average curvature, average slope, road surface smoothness index, and the recommended maximum safe speed calculated based on these physical characteristics. The system generates an ordered list of sub-driving segments, which constitutes a "digital twin" fragment of the road ahead of the vehicle. Each fragment has a clear "identity label" and "physical characteristic description."

[0060] In the core prediction phase, the system uses the current state of the vehicle and load to perform a "virtual dynamic simulation" of each road segment ahead. It runs a simplified vehicle-load coupled dynamic model and evaluates the stability of the load by calculating the longitudinal, lateral, and vertical inertial forces experienced by the load during vehicle movement. When the predicted combination of inertial forces and gravity causes the load's stability indicators (such as the dynamic stability angle or the internal friction angle of the cargo) to exceed a preset threshold, a "load movement event" is defined and marked as Low, Medium, or High according to the degree of exceedance. These events specifically refer to potential roll risks or cargo displacement risks.

[0061] Specifically, assuming the mining truck is heavily loaded with 88 tons and its center of gravity is 2.6 meters high, it is traveling at a speed of 20 km / h and is about to enter a complex section of road with a long downhill followed by a sharp turn; the system identifies four key nodes on the path: the start of the downhill, the start of the curve, the apex of the curve, and the end of the curve; these nodes divide the path into three sub-driving segments: the segment from the current position to the start of the curve (including straight sections and downhill sections), the complete downhill sharp turn segment, and the straight section after the curve.

[0062] When determining load movement events, the system makes predictions for each road segment. For the first segment, the model predicts that the vehicle will accelerate slightly under gravity, causing the ore to shift slightly backward, but the force is very small, so a low-level longitudinal displacement event is predicted, with no stability risk. However, for the second downhill sharp turn segment, assuming the vehicle enters at a speed of 22 km / h, the model calculates significant lateral acceleration and continuous longitudinal acceleration. Combined with the vehicle's relatively high center of gravity, the dynamic stability angle calculated by the system will be close to the critical value. Therefore, the system predicts a high-level roll risk event on this segment, accompanied by a medium-level lateral displacement event.

[0063] The output of S122 is a "warning map" of the road ahead with risk markings; the system clearly knows that the first road segment is not a big problem, but the second road segment is a high-risk road segment; this prediction result about the "load movement event" will be directly passed to subsequent steps, triggering the system to initiate the highest level of "cargo control event", such as applying stronger braking before entering a high-risk curve, strictly limiting the vehicle speed in the curve, and actively adjusting the suspension system to maximize anti-roll capability, thereby ensuring transportation safety.

[0064] refer to Figure 4 In step S13, the specific steps are as follows:

[0065] S131: Based on the identification of the movement event, multiple sub-movement items are identified. At the same time, multiple posture data of the unmanned mining truck are collected. A first bump coefficient is determined based on the multiple sub-movement items and the multiple posture data of the unmanned mining truck. A second bump coefficient is determined based on the overall shape of the multiple sub-movement items and the unmanned mining truck.

[0066] S132: Determine the bump pattern of the unmanned mining truck during its movement based on the mapping relationship between the first bump coefficient, the second bump coefficient, and the bump pattern, and determine multiple sub-bump parts of the unmanned mining truck based on the identification of the bump pattern.

[0067] S133: Collect the storage distribution map of ore relative to the unmanned mining truck, determine the ore transportation event based on multiple sub-bumping parts of the unmanned mining truck and the storage distribution map of ore relative to the unmanned mining truck, determine the dynamic driving event based on multiple sub-bumping parts of the unmanned mining truck and the driving speed of the unmanned mining truck, and determine the transportation control event of the unmanned mining truck based on the ore transportation event, the dynamic driving event and the overall shape of the unmanned mining truck.

[0068] In the embodiments of this application, multiple sub-movement items are determined based on the identification of load movement events of the unmanned mining truck in each sub-driving section. At the same time, multiple attitude data of the unmanned mining truck are collected. A first bump coefficient is determined based on the multiple sub-movement items and the multiple attitude data of the unmanned mining truck. A second bump coefficient is determined based on the overall shape of the multiple sub-movement items and the unmanned mining truck. This approach takes into account the overall shape of the multiple sub-movement items and the unmanned mining truck, ensuring the accuracy of the second bump coefficient.

[0069] At this point, when a predicted event is received by S12, such as "Load Shift_Lateral", S131 will trace its physical root cause; the system will identify which specific vehicle motion stimuli caused this result and define these stimuli as "sub-movement items".

[0070] The sub-movement items include: Lateral_Acceleration_Surge: This is a direct manifestation of the centrifugal force generated by the vehicle in a curve, and is the main excitation that causes the load to slide laterally or overturn; Longitudinal_Acceleration_Jerk: This is the rate of change of acceleration; a sharp negative value of Jerk represents sudden braking, which will cause the load to lurch forward violently; while a positive value represents rapid acceleration, which will cause the load to recoil; Vertical_Vibration_Impact: Vertical acceleration of the wheels caused by road potholes, speed bumps, etc., which will cause the load to bounce or even collide with the passenger compartment; Roll_Angle_Rate: The rate at which the vehicle body rolls; when entering or exiting a curve, an excessively large roll rate will instantly exacerbate the lateral instability of the load, even if the final roll angle is not large.

[0071] For each incoming "load shift event", the system outputs a list of one or more associated "sub-shift items" with predicted peak or root mean square (RMS) values; for example, a "load lateral shift" event is specifically broken down into:

[0072] {Lateral_Acceleration_Surge:1.5m / s²,Roll_Angle_Rate:8deg / s}.

[0073] A weighted evaluation algorithm based on multi-sensor data fusion is typically used. Its input consists of two parts: first, the aforementioned multiple "sub-movement items" (as evaluation targets), and second, real-time sensor data streams from S111, mainly including triaxial acceleration, triaxial angular velocity, and data from suspension displacement sensors measured by the IMU.

[0074] The system precisely matches real-time attitude data with "sub-movement items"; for example, it uses the lateral acceleration a_y measured by the IMU to respond to the Lateral_Acceleration_Surge excitation, and synthesizes these dynamic response quantities into a scalar value through a preset weighted summation function (or a more complex nonlinear model); the formula can be simplified to:

[0075] C1= ×|a_y_measured|+ ×|r_measured|+ ×|a_z_measured|+…;where the weights , and The C1 value is determined by the vehicle's dynamic characteristics. For example, for a mining truck with a high center of gravity, roll stability is crucial, so roll-related sensor data (such as r_measured) will be given higher weight. C1 reflects the overall performance of the vehicle's suspension system, tire damping, etc., under real road conditions. A high C1 value means that the vehicle is experiencing violent dynamic motion that exceeds the comfort or safety range, and is direct evidence that "bumps are happening violently."

[0076] The second bump coefficient is determined based on the overall shape of multiple sub-movement projects and the unmanned mining truck. This coefficient measures the vehicle's inherent physical sensitivity to specific stimuli; it is a "structural characteristic" index based on the vehicle's static parameters, answering the question of "is this vehicle inherently prone to 'bumps'?" Simultaneously, a quasi-static calculation based on the vehicle's static parameters is performed. Its inputs are also "sub-movement projects" (as "virtual stimuli" applied to the vehicle model) and an "overall shape" database containing key physical parameters of the vehicle. This database includes: geometric parameters: wheelbase, track width, center of gravity height (CoG_z); mass parameters: vehicle curb weight, current gross weight (m_total); stiffness and damping parameters: suspension system stiffness and damping coefficients.

[0077] The system runs a simplified vehicle model. When a "sub-motion item" (such as a Lateral_Acceleration_Surge of 1.5 m / s²) is input, the model calculates the vehicle's theoretical response based on the overall morphological parameters. For example, based on m_total and CoG_z, it can calculate how much load will be transferred to the outer wheels by this lateral acceleration, thus generating a theoretical roll angle. This calculated theoretical response value constitutes the core of C2. C2 is essentially a reflection of the vehicle's transfer function. For the same excitation, a vehicle with a high center of gravity and soft suspension will produce a larger theoretical response than a vehicle with a low center of gravity and hard suspension, and therefore its C2 value will be higher. It reveals the vehicle's structural fragility or sensitivity. The formula for calculating C2 can be simplified as follows:

[0078] C2 =α×Norm(L T ) + β ×Norm(I xx ) + γ ×Norm(VSP), where, L T For Load_Transfer (load transfer amount); I xx For Inertia (roll inertia); VSP stands for Vehicle_Structure_Parameters (vehicle structural parameters, which can be simplified to suspension stiffness K). roll α, β, and γ are weighting coefficients, set by engineers based on experience or experimental data, used to adjust the contribution of each factor to C2. For example, α can be set to 0.6, β to 0.3, and γ to 0.1. Norm(.) is the normalization function, which maps the original values ​​of each physical quantity to the interval [0,1] for weighted summation.

[0079] Specifically, the mining truck (88 tons heavy load, center of gravity height 2.6m) is entering the Segment_B (downhill sharp bend) section of the ore transportation site at a speed of 22km / h; S12 has predicted the High_Risk_Roll (high-risk tilt) event; the system decomposes the High_Risk_Roll event into the following specific physical excitations: Lateral_Acceleration_Surge: predicted peak value of 1.5m / s²; Roll_Angle_Rate: predicted peak value of 10deg / s.

[0080] The mine truck's IMU real-time data stream showed that at the moment of entering the curve, the measured lateral acceleration a_y rapidly increased to 1.4 m / s², and the yaw rate r reached 9.5 deg / s. The system substituted these measured values ​​into the weighting function (simplified example): C1 = 0.6 × |1.4| + 0.4 × |9.5|. The calculated C1 was a high value, indicating that the vehicle was experiencing severe dynamic turbulence, which was highly consistent with the prediction of S12, confirming that "the situation was indeed very bad".

[0081] The system retrieves the overall morphological parameters of the mine car: m_total=88t, CoG_z=2.6m (this is a very high value); the system uses Lateral_Acceleration_Surge=1.5m / s² as input and applies it to a simplified tilt model; the model is based on the formula:

[0082] L T =(m_total×a_y×CoG_z) / Track_Width=88000×1.5×2.6) / 4.0=85800N;

[0083] I xx ≈m_total×(CoG_z) 2 =88000 ×(2.6) 2 =594880kg·m 2 ;

[0084] VSP considers suspension stiffness K here. roll The value is 1200000 Nm / rad.

[0085] Normalization is performed, and the L values ​​of all mining trucks in the database are preset. T The range is [20000, 100000]N, I xx The range is [100000, 700000] kg·m², K roll The range is [800000, 1, 500000] Nm / rad, calculated according to the formula:

[0086] Norm(L T ) = (85800 - 20000) / (100000 - 20000) = 0.8225;

[0087] Norm(I xx ) = (594880 - 100000) / (700000 - 100000) = 0.8248;

[0088] Norm(VSP)=(1200000-800000) / (1500000-800000)=0.5714.

[0089] Substituting α, β, and γ into the weights, we get C2 = 0.6 × 0.8225 + 0.3 × 0.8248 + 0.1 × 0.5714 = 0.79808 ≈ 0.8. The massive mass of the mine car at 88 tons means that any motion is accompanied by enormous inertia, further amplifying the theoretical response. After comprehensive calculation, C2 is also rated as a high value (usually in the [0, 1] interval, C2 > 0.7 is considered a high value), indicating that the mine car, due to its high center of mass and heavy load, is extremely sensitive to lateral tilting excitation, confirming that "this car itself is very afraid of this situation."

[0090] S131 outputs two mutually corroborating high bump coefficients: a high C1 indicates a severe dynamic response, and a high C2 indicates a sensitive vehicle structure. These two solid quantitative indicators provide an indispensable data foundation for the next step, S132, to accurately determine the "bump pattern," and S133, to formulate a robust "carrier control event."

[0091] Furthermore, the bump pattern of the unmanned mining truck during its movement is determined based on the mapping relationship between the first bump coefficient, the second bump coefficient, and the bump pattern. Multiple sub-bump parts of the unmanned mining truck are determined based on the identification of the bump pattern. This approach takes into account the overall consideration of the mapping relationship between the first bump coefficient, the second bump coefficient, and the bump pattern, ensuring the accuracy of the bump pattern of the unmanned mining truck during its movement.

[0092] At this point, the system treats the first turbulence coefficient (C1, dynamic response) and the second turbulence coefficient (C2, structural sensitivity) as coordinate points (C1, C2) in a two-dimensional feature space. In this space, different regions are pre-labeled through offline training or expert experience, corresponding to different turbulence patterns. There are two main ways to establish this mapping relationship:

[0093] Data-driven approach: Collect (C1, C2) data under various typical working conditions through a large number of real vehicle tests or high-fidelity simulations, and have domain experts label the data (e.g., "This is a side-tilt bump"); use classifiers such as support vector machines (SVM), decision trees or neural networks to learn the mapping boundary from the feature space to the morphological label.

[0094] Rule-based approach: Design a fuzzy logic system or a series of IF-THEN rules; for example:

[0095] IF(C1_Roll_Component>Threshold1)AND(C2_Roll_Component>Threshold2)THEN Bump Pattern = Roll_Dominant. The system predefines several bump pattern categories to describe the dominant characteristics of vehicle motion, such as: Roll_Dominant (tilt-dominant): mainly characterized by the swaying of the vehicle body around the longitudinal axis, typically seen when driving on a curve; Pitch_Dominant (pitch-dominant): mainly characterized by the forward and backward nodding of the vehicle body around the lateral axis, typically seen during acceleration / braking; Heave_Dominant (heave-dominant): mainly characterized by the up and down jumping of the vehicle body along the vertical axis, typically seen when driving over bumpy roads; Complex_Coupled (complex coupling): two or more patterns are very significant, usually occurring in complex road conditions; this step will output one or more bump pattern labels, such as {Roll_Dominant,Heave_Dominant}, and the confidence level corresponding to each pattern.

[0096] The system maintains a simplified multibody dynamics model of the vehicle, which defines the connection relationships and force / torque transmission paths between various components (such as suspension, frame, and axle). The positioning logic is that once the "bump mode" is determined, the system will use it as the main excitation input into the vehicle model and analyze the transmission path and response distribution of the excitation in the vehicle structure through simulation.

[0097] The system predefines a series of key "sub-bumps," which are the ultimate bearers of dynamic loads, including: suspension system: such as Left_Front_Suspension_Strut (left front suspension strut) and Right_Rear_Suspension_Arm (right rear suspension control arm); frame and body: such as Chassis_Longitudinal_Rail (frame longitudinal beam) and Cargo_Bed_Mounting_Points (cargo box mounting points); axle and tire: such as Front_Axle_Assembly (front axle assembly) and Rear_Tire_Bead (rear tire bead).

[0098] Examples of location analysis are as follows: For Roll_Dominant configuration: the model shows that centrifugal force causes body roll, one side of the suspension is extremely compressed, the other side is stretched, and the frame bears huge torsional stress; therefore, the sub-bumps are identified as the suspension on the compression side, the suspension on the stretching side, and the anti-torsional components of the frame; For Pitch_Dominant configuration: the model shows that braking force causes the center of gravity to shift forward, the front suspension is compressed, and the rear suspension is stretched; therefore, the sub-bumps are identified as the front suspension system and the rear suspension system; For Heave_Dominant configuration: the model shows that road impacts are transmitted to the frame through the tires and suspension; therefore, the sub-bumps are mainly the suspension assemblies of the four wheels and the floor of the vehicle.

[0099] Specifically, in Segment_B (downhill sharp bend) of the ore transport site, the system has calculated that both C1 and C2 are high values. For the bumpy morphology, the inputs are: C1 (high value, where the tilt-related components a_y and r are dominant), C2 (high value, mainly contributed by the high tilt sensitivity caused by the high center of mass). The mapping process is: the point (C1, C2) falls within the pre-labeled Roll_Dominant region in the feature space. The classifier output is: bumpy morphology = Roll_Dominant, with a confidence level of 95%. The conclusion is: the system clearly determines that the ore truck is currently experiencing mainly tilt-dominated bumpy morphology.

[0100] For the sub-bump locations, input: Bump morphology Roll_Dominant; Location analysis: The system applies the Roll_Dominant excitation to the vehicle model of the mining truck; Since the vehicle is turning right, the model simulation shows: Centrifugal force points to the left, causing the left suspension to be stretched and the right suspension to be strongly compressed; The frame undergoes counterclockwise torsion, with the right frame longitudinal beam bearing compressive stress and the left bearing tensile stress; The right rear wheel, as the outermost and drive wheel, bears the largest vertical load and lateral force; Location results: Based on the model analysis, the system identifies the following sub-bump locations: Right_Rear_Suspension (right rear suspension): identified as the main pressure and stress concentration point; Left_Front_Suspension (left front suspension): identified as the main tension point; Chassis_Frame_Twist_Section (frame torsional section): identified as the structural component bearing the maximum torsional stress; Right_Rear_Tire_Sidewall (right rear tire sidewall): identified as the component bearing a huge lateral thrust.

[0101] The output of S132 is a precise "diagnostic report": the mining truck is experiencing a tilt-dominant type of jolt, with the most severe physical effects occurring in specific areas such as the right rear suspension, left front suspension, torsional section of the frame, and right rear tire. This precise identification from "type" to "location" provides irreplaceable accuracy for the next step, S133, to directly link "vehicle jolt" with "cargo status" and "control strategy." The system no longer simply knows that "the truck is shaking," but precisely knows "how the truck is shaking" and "where the shaking is most severe."

[0102] Therefore, by collecting the storage distribution map of ore relative to the unmanned mining truck, determining ore transport events based on multiple sub-bump locations of the unmanned mining truck and the ore storage distribution map relative to the unmanned mining truck, determining dynamic driving events based on multiple sub-bump locations of the unmanned mining truck and the driving speed of the unmanned mining truck, and determining the transport control events of the unmanned mining truck based on the ore transport events, dynamic driving events, and the overall shape of the unmanned mining truck, this approach takes into account the overall consideration of ore transport events, dynamic driving events, and the overall shape of the unmanned mining truck, ensuring the accuracy of the transport control events of the unmanned mining truck. At the same time, multiple sub-driving sections are introduced to further control the load movement events, taking into account the overall consideration of the bump pattern, the ore storage distribution map relative to the unmanned mining truck, and the driving speed of the unmanned mining truck, thus improving the accuracy of the transport control events of the unmanned mining truck.

[0103] At this point, the inputs include "multiple sub-bump locations" (the locations on the vehicle experiencing the most severe stress) from S132 and "storage distribution map of ore relative to the unmanned mining truck" from S111 (a digital model describing the three-dimensional spatial distribution of ore within the truck). The system performs spatial overlay analysis on the positions of the "sub-bump locations" in the vehicle coordinate system and the "storage distribution map." The underlying principle is that the cargo above the most severely bumpy location experiences the greatest inertial excitation. If the cargo at that location is inherently unstable (e.g., steep slopes or suspended structures), the risk increases exponentially.

[0104] Key assessment metrics include: Cargo stability margin: calculating the probability of cargo collapse or slippage by combining the direction of inertial forces generated by bumps and the natural angle of repose of the ore pile; Impact energy transfer: assessing how much mass of ore will be transferred through the impact force at the bumpy location, and the energy attenuation along the transfer path; this step outputs a structured Ore_Transport_Event, which clearly describes the cargo risk, for example:

[0105] {Type:'Lateral_Collapse',Severity:'High',Location:'Right_Rear_Cargo_Area'}.

[0106] In addition, the inputs include "multiple sub-bump locations" from S132 (reflecting the intensity and type of bumps) and "driving speed of the unmanned mining truck" from S111 (a key variable that determines the vehicle's kinetic energy and inertial force); the system uses the information from the "sub-bump locations" to deduce the dynamic excitations (such as lateral acceleration and vertical impact force) that the vehicle is currently experiencing; and combines these excitations with the current "driving speed" to calculate the dynamic stability boundary of the vehicle in real time.

[0107] Key evaluation metrics include: Dynamic rollover threshold: Calculates the load transfer amount under current lateral acceleration and speed to determine if it is approaching or exceeding the physical limit of the vehicle lifting the inner wheel; Tire grip margin: Calculates whether the lateral force demand of the tires is approaching their friction limit with the ground to determine if there is a risk of sideslip; Suspension travel margin: Assesses whether the compression or extension of the suspension has approached the end of its physical travel to determine if there is a risk of "bottoming out" or "loosening"; This step outputs a Dynamic_Driving_Event, which describes the overall dangerous state of the vehicle, for example:

[0108] {Type:'Approaching_Roll_Over_Limit',Severity:'Critical'}.

[0109] The transportation control events of the unmanned mining truck are determined based on ore transportation events, dynamic driving events, and the overall shape of the unmanned mining truck. The inputs include the aforementioned "ore transportation events", "dynamic driving events", and "overall shape of the unmanned mining truck" (the control capability boundary of the vehicle, such as whether there is active suspension, independent wheel control braking, etc.).

[0110] The decision-making logic is as follows: determine the severity level of two events; events with a Critical level have the highest priority; match one or more basic control actions from the policy library based on the risk type and level; verify whether the vehicle is capable of executing the policy based on the overall situation; for example, if the vehicle does not have active suspension, the "adjust suspension stiffness" command cannot be generated; and merge multiple basic control actions into a unified, conflict-free Transport_Control_Event.

[0111] Typical examples of vehicle control events include: Execute_Emergency_Roll_Stabilization: a compound command that includes "reducing vehicle speed", "applying independent braking force to the outer wheels", and "activating active anti-side bars"; Switch_to_Low_Speed_High_Stability_Mode: a mode switching command that changes the parameters of the entire motion controller, sacrificing speed for ultimate stability.

[0112] Specifically, the mining car (heavy load of 88 tons, high center of gravity, ore piled up on the right rear) is in Segment_B (downhill sharp bend) of the ore transportation site. S132 has identified the bump pattern as Roll_Dominant and the sub-bump location as Right_Rear_Suspension, etc.

[0113] For the ore transport event, input: Sub-bump location = Right_Rear_Suspension (right rear suspension); the stored distribution map shows that the ore has accumulated into a small hill on the right rear side of the car; the system spatially superimposes "severe compression of the right rear suspension" and "ore accumulation on the right rear side"; the conclusion is: a strong lateral inertial force will directly act on the already unstable ore pile; output:

[0114] Ore_Transport_Event={Type:'Imminent_Lateral_Collapse',Severity:'High',Location:'Right_Rear_Cargo_Area'} (The right rear cargo area is about to experience a lateral collapse.)

[0115] For dynamic driving events, the inputs are: the sub-bump area reflects severe roll; the driving speed is 22 km / h; based on the severity of the roll and the speed of 22 km / h, the system calculates that the load transfer has reduced the load on the inner wheel to less than 10% of the total weight, which is very close to the physical limit of dynamic rollover; the output is: Dynamic_Driving_Event={Type:'Approaching_Roll_Over_Limit',Severity:'Critical'} (close to the rollover limit).

[0116] For vehicle control events, input:

[0117] Ore_Transport_Event(High) + Dynamic_Driving_Event(Critical) + overall shape of the mining truck (assuming it has brake-by-wire and independent wheel control capabilities); Decision: The Critical level of Dynamic_Driving_Event triggers the highest priority response; the system determines that the most effective measures must be taken immediately to reduce lateral forces and increase anti-roll moment; the strategy library matches the "Emergency Roll Stabilization" strategy, which is supported by the mining truck's hardware capabilities; Output:

[0118] Transport_Control_Event=Execute_Emergency_Roll_Stabilization.

[0119] This final control event is passed to the execution layer of S14; it is not a simple "brake" command, but a carefully choreographed complex control action: 1) The motion controller immediately plans a smooth but forced deceleration curve, aiming to reduce the vehicle speed to below 15 km / h within 2 seconds; 2) The brake controller applies a small, independent braking force to the left front and left rear wheels, generating a corrective yaw moment opposite to the roll direction, helping the vehicle to "pull back" its posture; 3) If the mining truck is equipped with active suspension, the system will instantly adjust the right suspension damping to the maximum to resist compression; This control event directly addresses the two specific risks of "cargo about to collapse" and "vehicle about to overturn," achieving unprecedentedly refined control.

[0120] refer to Figure 5 In step S14, the specific steps are as follows:

[0121] S141: Collect transport control events, determine multiple sub-transport control contents based on the detection of transport control events, determine key transport control features based on the identification of each sub-transport control content, and collect multiple key transport control features; determine the smooth transport coefficient of the unmanned mining truck based on the feature position, corresponding feature shape and moving noise data of the unmanned mining truck based on the multiple key transport control features.

[0122] S142: Collect the current attitude of the unmanned mining truck, determine the first control parameter based on the stable transport coefficient of the unmanned mining truck and the current attitude of the unmanned mining truck, and determine the second control parameter based on the stable transport coefficient of the unmanned mining truck and the overall shape of the unmanned mining truck.

[0123] S143: Determine the dynamic control mode of the unmanned mining truck based on the mapping relationship between the first control parameter, the second control parameter and the dynamic control mode. Determine multiple dynamic control items based on the identification of the dynamic control mode. Determine multiple dynamic control events of driving data based on multiple dynamic control items, corresponding project nodes and unmanned mining truck, so as to dynamically control multiple driving data of unmanned mining truck.

[0124] In the embodiments of this application, the Transport_Control_Event (such as Execute_Emergency_Roll_Stabilization) in S13 is a target-oriented instruction; S141 needs to assign it to different subsystems of the vehicle to complete it collaboratively; each "sub-transport control content" corresponds to a specific action of a vehicle subsystem; for the Execute_Emergency_Roll_Stabilization event, it can be decomposed into: Differential_Braking_Control: executed by the braking subsystem; Driveline_Torque_Limiting: executed by the powertrain subsystem; Active_Suspension_Stiffening: executed by the suspension subsystem (if equipped); for the Initiate_Gentle_Speed_Control event, it can be decomposed into: Progressive_Brake_Application; Engine_Retardation_Engagement.

[0125] Each "sub-vehicle control content" requires one or more physical quantities to accurately describe its control objective; these physical quantities are the "key vehicle control features"; for Differential_Braking_Control, the key features are: Target_Brake_Pressure_Delta (target braking pressure difference), in bar; Target_Brake_Wheel_Selection (target brake wheel selection), such as Left_Rear_Wheel; for Driveline_Torque_Limiting, the key feature is: Max_Allowable_Torque (maximum allowable torque), in Nm; for Active_Suspension_Stiffening, the key feature is: Target_Damping_Coefficient_Factor (target damping coefficient factor), a dimensionless multiple (e.g., 1.5 times).

[0126] Input: Multiple key vehicle control features: These features together constitute a "desired system response"; for example, Target_Deceleration = -0.5 m / s²; Feature location: Indicates the location of the control action, such as Left_Rear_Wheel, which helps to locate the noise source; Feature morphology: Describes the time-domain characteristics of the control command, whether it is a step signal or a ramp signal, which determines the dynamic process of the desired response; Moving noise data: This is the "actual system response", from high-frequency sensor data from S111, such as IMU outputs a_x, a_y, a_z, and wheel speed sensor fluctuation data; Here, "noise" does not refer to useless signals, but rather to all high-frequency dynamic components of the system during the response command process.

[0127] The system has an internal vehicle reference model; when "key features" are input, this model generates an ideal, noise-free response curve; the system compares the "moving noise data" (actual response) with this ideal response curve in real time; the smoothness coefficient S is usually calculated based on the following indicators: tracking error: the deviation between the actual response and the ideal response; the smaller the deviation, the higher S; response smoothness: evaluated by calculating the jerk (acceleration) or high-frequency vibration energy of the actual response; the smaller the vibration, the higher S; overshoot: the extent to which the actual response exceeds the target value; the smaller the overshoot, the higher S; S is a normalized value (e.g., 0 to 1), S=1 represents perfectly smooth control, and S=0 represents severe loss of control.

[0128] Furthermore, the current attitude of the unmanned mining truck is collected, and the first control parameter is determined based on the stable transport coefficient and the current attitude of the unmanned mining truck. The second control parameter is determined based on the stable transport coefficient and the overall shape of the unmanned mining truck. This comprehensive consideration of the stable transport coefficient and the overall shape of the unmanned mining truck ensures the accuracy of the second control parameter.

[0129] At this point, the inputs are: the stationary load factor (S): from S141, which is a scalar feedback signal that measures the control quality; the current attitude: the real-time state vector from S111, such as [a_x, a_y, roll_angle, pitch_angle, yaw_rate]; the stationary load factor (S) plays the role of "gain scheduling" here; if the S value is low (control is not stationary), the system will automatically reduce the controller gain to make the response smoother and slower in order to suppress oscillations; if the S value is high (control is stationary), the system can appropriately increase the gain to obtain a faster response speed.

[0130] The error between the current attitude and the target attitude is the core that the controller needs to correct; for example, roll_angle_error = target_roll_angle - current_roll_angle; the first control parameter is the direct control command calculated by the adaptive controller based on the gain adjusted by S and the corrected error; it is a dynamic, high-frequency correction quantity.

[0131] Parameter examples: Target_Steering_Torque_Adjustment: Used for fine-tuning steering, correcting path deviations, or suppressing body roll; Individual_Wheel_Brake_Torque_Vector: A vector containing the target braking torque for all four wheels, used to achieve differential braking or precise brake force distribution; Instantaneous_Torque_Request: Torque commands sent to the engine / motor for precise control of acceleration and deceleration.

[0132] The second control parameter is determined based on the smooth transport coefficient and the overall shape of the unmanned mining truck. Input: Smooth transport coefficient (S): as part of the optimization objective function; the optimization objective is to maximize S under the premise of satisfying all constraints; Overall shape: from the vehicle static parameter database of S111, such as m_total (total mass), CoG_z (center of gravity height), suspension_stiffness, etc.

[0133] The overall shape parameter defines the physical boundaries of the control command; for example, the total mass m_total determines that the maximum braking force cannot exceed μ×m_total×g (μ is the coefficient of friction); the second control parameter is usually not a direct action command, but sets an operating range or reference value for the first control parameter; it tells the controller: "You can move within this range, but do not go beyond the boundary."

[0134] The system will calculate an "ideal" control tone based on the overall shape; for example, for heavy-duty vehicles, the system will preset a more conservative upper limit of acceleration; it will fine-tune this tone based on the smooth load factor (S); if S continues to be low, the system will further tighten this operating range.

[0135] Parameter examples: Dynamic_Roll_Stability_Threshold: A safety boundary calculated based on CoG_z and m_total. The roll angle generated by the first control parameter cannot exceed this threshold. Adaptive_Speed_Limit_Profile: A speed limit curve dynamically generated based on the curvature of the road ahead and the vehicle's current m_total. The torque request from the first control parameter cannot cause the vehicle to exceed this speed. Suspension_Damping_Map_Selection: Selects a "heavy load" or "light load" damping map based on m_total, providing a benchmark for suspension adjustments by the first control parameter.

[0136] Therefore, the dynamic control mode of the unmanned mining truck is determined based on the mapping relationship between the first control parameter, the second control parameter, and the dynamic control mode. Multiple dynamic control items are determined based on the identification of the dynamic control mode. Dynamic control events for multiple driving data are determined based on the multiple dynamic control items, the corresponding project nodes, and the unmanned mining truck, so as to dynamically control multiple driving data of the unmanned mining truck. This approach takes into account the overall consideration of multiple dynamic control items, the corresponding project nodes, and the unmanned mining truck, ensuring the accuracy of the dynamic control events for multiple driving data.

[0137] At this point, the dynamic control mode of the unmanned mining truck is determined based on the mapping relationship between the first control parameter, the second control parameter, and the dynamic control mode. The first control parameter reflects the dynamic corrections that need to be made (such as Target_Brake_Torque_Vector), representing "what to do" and "how much to do"; the second control parameter reflects the operating boundary of the vehicle (such as Dynamic_Roll_Stability_Threshold), representing "how much can be done" and "where the safety limit is".

[0138] The system internally predefines a mapping relationship from the parameter space to the "dynamic control mode"; this relationship can be established through an expert rule base or a machine learning classifier.

[0139] Each mode represents a complete and coordinated control system; High_Precision_Stabilization: Activated when the first parameter requires strong intervention and the second parameter provides a relatively wide boundary; this mode pursues the fastest attitude convergence speed, sacrificing some comfort; Gentle_Correction_Mode: Activated when the first parameter requires correction, but the second parameter boundary is tight (or the smoothness coefficient is low); this mode prioritizes smoothness, and the control action is very gentle.

[0140] Boundary_Maintenance_Mode: Activated when the correction of the first parameter is small, but the vehicle state is close to the safe boundary defined by the second parameter; the goal of this mode is to "stay within the boundary", that is, to precisely maintain operation within the safe boundary and avoid any actions that cross the boundary. Emergency_Avoidance_Maneuver: Activated when the correction requirement of the first parameter is huge and the second parameter indicates that the boundary is about to be crossed; this mode avoids instability at all costs and will trigger the most violent control actions.

[0141] Based on the identification of dynamic control modes, multiple dynamic control items are determined. Each "dynamic control mode" is associated with a preset task list, and these tasks are assigned to different subsystems of the vehicle. For the High_Precision_Stabilization mode, its dynamic control items include: Modulate_Individual_Wheel_Braking (modulate individual wheel braking); Adjust_Driveline_Torque_Distribution (adjust powerline torque distribution); Control_Active_Roll_Stabilizer (control active anti-side bars). For Gentle_Correction_Mode, its items are simpler: Apply_Progressive_Brake_Pressure (apply progressive braking pressure); Slightly_Reduce_Engine_Torque (slightly reduce engine torque).

[0142] Input: Multiple dynamic control items: a list of tasks to be executed; corresponding item nodes: specific conditions that trigger each task, which can be time, location, or state thresholds; for example, Node:'Roll_Angle_Exceeds_2.0_deg' or Node:'Distance_To_Curve_Apex<5m'; driverless mining truck: the real-time status of the vehicle, used to determine whether the conditions of the item node are met; the system continuously monitors the vehicle status, and when the conditions of a certain "item node" are met, a "dynamic control event" is triggered; this event is a structured instruction that directly acts on a certain "driving data" (i.e., the underlying control variable).

[0143] Event structure: {Target_Data, Control_Action, Value, Duration}; Target_Data: Driving data to be controlled, such as wheel_speed_rr (right rear wheel speed), steering_angle (steering angle), engine_torque_request (engine torque request); Control_Action: Control action, such as Set, Increase, Decrease; Value: Target value or change; Duration: Duration of the instruction; Outputs a series of high-frequency, low-level control instructions directly sent to actuators such as the vehicle ECU, brake controller, and motor controller, thereby completing the dynamic control of the unmanned mining truck.

[0144] refer to Figure 6 In step S15, the specific steps are as follows:

[0145] S151: Collect the task list of the unmanned mining truck, identify multiple transportation tasks based on the identification of the task list of the unmanned mining truck, and identify unfinished transportation tasks based on the multiple transportation tasks, the completed events of the unmanned mining truck and the current position of the unmanned mining truck.

[0146] S152: Determine the corresponding transport location based on the identification of unfinished transport tasks, mark the ore transport volume at the transport location, and determine the ore control tasks to be completed based on each transport location, the corresponding ore transport volume, and the ore load of the unmanned mining truck.

[0147] S153: Collect the location of the remaining unmanned mining trucks and mark their current status. Based on the current status of the remaining unmanned mining trucks, the corresponding ore capacity, and the ore control tasks to be completed, determine the collaborative transportation content of the remaining unmanned mining trucks. Based on the collaborative transportation content of the remaining unmanned mining trucks and their current transportation content, determine the collaborative control mode of the unmanned mining truck.

[0148] In the embodiments of this application, the input is a task list of the unmanned mining truck, which is typically a structured data file (such as XML, JSON, or database records) issued by a central scheduling system or stored in the onboard task manager. The system has a built-in task syntax definer that can identify and parse each task item in the list. A standard "transportation task" typically includes the following key attributes: Task_ID: a globally unique task identifier; Task_Type: the task type, such as LOAD, HAUL, DUMP, RETURN; Waypoint_Sequence: an ordered list of waypoints that defines the path of the task; Associated_Assets: the physical resources associated with the task, such as the specified loader, crushing station, etc.; Priority_Level: the priority of the task; Temporal_Constraints: time constraints, such as the start / end window of the task. Through this process, a vague list of instructions is transformed into standardized task units that are machine-readable and manageable.

[0149] The input includes the previously parsed "multiple transport tasks", the vehicle's "completed event log" (which records the IDs of all successfully completed tasks and their completion timestamps), and the "current location of the unmanned mining truck" from S111; the system iterates through all transport tasks, comparing the Task_ID of each task with the ID in the completed event log; all tasks that can be matched in the log are marked as COMPLETED and removed from the candidate set; the remaining tasks form a candidate incomplete task set; the system further confirms the tasks in the candidate incomplete task set to determine their precise status.

[0150] Current task: Check the highest priority task in the candidate set; if the vehicle's current position is near the task's starting waypoint, the task is marked as IN_PROGRESS (in progress); Next task: If the vehicle is not at the starting waypoint, but on the path to the starting waypoint, the task is marked as NEXT_TO_EXECUTE (pending execution); Future tasks: Tasks further down the sequence are marked as SCHEDULED (scheduled); This step outputs an ordered list of one or more incomplete payload tasks, each with its precise status label.

[0151] Furthermore, based on the identification of incomplete transport tasks, the corresponding transport locations are determined, and the ore transport volume at each transport location is marked. Based on each transport location, the corresponding ore transport volume, and the ore load of the unmanned mining truck, the ore control tasks to be completed are determined. This approach takes into account the overall consideration of each transport location, the corresponding ore transport volume, and the ore load of the unmanned mining truck, ensuring the accuracy of the ore control tasks to be completed.

[0152] At this point, the corresponding transport location is determined based on the identification of the unfinished transport task, and the ore transport volume of the transport location is marked. The input includes the "unfinished transport task" from S151 and the "transport location" (waypoint) defined in the task.

[0153] The system extracts the key "Source" and "Destination" locations from the task definition. For example, for a RETURN task, the source is the unloading point, and the destination is the loading point. The system queries the real-time material data associated with each location through GIS and MMS interfaces. For the source location (such as a crushing station), the system queries its "output" or "consumption" capacity, such as Current_Hopper_Level (current hopper level). For the destination location (such as a loading point), the system queries its "input" or "inventory" status, such as Available_Ore_Stockpile (available ore stockpile). This step outputs a structured description that associates the task, location, and material status, for example:

[0154] {Task:T004,Source:'Crusher_B',Source_Status:'Adequate_Receiving_Capacity',Destination:'Loader_A',Destination_Status:'Sufficient_Stockpile_Available'}.

[0155] The inputs include the previously associated "each transport location and its corresponding ore transport volume" and the "ore load of the unmanned mining truck" from S111; the system determines the physical operation type of the ore based on the Type of the incomplete transport task; LOAD task → PICKUP operation; DUMP task → DEPOSIT operation; HAUL / RETURN task → TRANSPORT operation, but essentially it is an intermediate state connecting PICKUP and DEPOSIT; the amount of ore in the control task is directly determined by the ore load (or maximum load) of the unmanned mining truck; This value represents the basic unit of material that the vehicle can control in one task cycle; the system combines "operation type", "material quantity", "source" and "target" into a complete "ore control task"; this step will output an ore control task to be completed that describes the specific physical operation, and its standard format is: {Action:[PICKUP / DEPOSIT],Material:'Ore',Quantity:[Current_Load],Source:[Location_A],Destination:[Location_B]}.

[0156] Therefore, the location of the remaining unmanned mining trucks is collected, and their current status is marked. Based on the current status of the remaining unmanned mining trucks, their corresponding ore capacity, and the ore control tasks to be completed, the collaborative transport content of the remaining unmanned mining trucks is determined. Based on the collaborative transport content of the remaining unmanned mining trucks and the current transport content of the unmanned mining trucks, the collaborative control mode of the unmanned mining trucks is determined. This approach takes into account both the collaborative transport content of the remaining unmanned mining trucks and the current transport content of the unmanned mining trucks, ensuring the accuracy of the collaborative control mode. At the same time, a dynamic control mode is introduced, which realizes the overall consideration of the unfinished transport tasks, the ore load of the unmanned mining trucks, and the current status of the remaining unmanned mining trucks, thereby improving the synergy of the unmanned mining truck control mode.

[0157] At this point, the input is obtained through V2V broadcasting or the scheduling system API, which provides the "location of the remaining unmanned mining trucks" and the "current status of the remaining unmanned mining trucks." This status vector typically includes: Operational_Status (running status): such as IDLE (idle), EN_ROUTE_TO_LOAD (en route to the loading point), etc.; Current_Load (current load): in tons; Estimated_Time_of_Arrival (ETA) (estimated arrival time); Health_Status (health status): such as HEALTHY, DEGRADED_PERFORMANCE, etc. The system will process the received raw data according to a predefined tagging system to form a clear global fleet status diagram, for example:

[0158] Vehicle_B:{Position:(x,y),Status:'EN_ROUTE_TO_DUMP',Load:0t,ETA_to_Loader_A:5min}.

[0159] Based on the current state of the remaining unmanned mining trucks, their corresponding ore capacity, and the ore control tasks to be completed, the collaborative transportation content of the remaining unmanned mining trucks is determined, using task allocation based on optimization algorithms (such as the Hungarian algorithm or genetic algorithm) or heuristic rules. Its inputs include the "global state graph" from the previous step, the "corresponding ore capacity" of each truck, and the "ore control tasks to be completed" pool from S152. The determination logic is as follows: the algorithm uses maximizing the total transportation volume and minimizing the total waiting time as global optimization objectives, evaluates the "cost" or "benefit" of each truck performing each task, and generates a suggested "collaborative transportation content" for each truck. This content can be Assign_Task(T006), Reroute_To(Stockpile_C), Form_Platoon_With(Vehicle_A), or Hold_Position, etc.

[0160] The input includes "the collaborative transport content of the remaining unmanned mining trucks" (i.e., global decision suggestions) and "the current transport content of the unmanned mining truck" (i.e., the truck's original plan). The system analyzes the relationship between "its own task" and "others' tasks" and selects the best matching mode from a predefined mode library based on the analysis results. Typical modes include: Independent_Execution_Mode: selected when the collaborative analysis shows that executing the original plan is the globally optimal solution; Coordinated_Handshake_Mode: selected when task or information handover with another truck is required; Platooning_Mode: selected when the path and speed of another truck highly overlap, and platooning brings significant benefits; Dynamic_Rerouting_Mode: selected when the collaborative analysis finds congestion on the original planned path or a new, more efficient global path; Yield_Priority_Mode: selected when yielding to higher priority vehicles.

[0161] Please see Figure 7 , Figure 7 This is a schematic diagram of the control device for the unmanned mining truck in an embodiment of the present invention; the control device for the unmanned mining truck includes:

[0162] The driving mode module 21 is used to collect the driving status of the unmanned mining truck and determine the driving mode of the unmanned mining truck based on the driving status of the unmanned mining truck and the terrain corresponding to the driving path.

[0163] The load movement event module 22 is used to determine multiple sub-driving segments based on the identification of the driving path when the unmanned mining truck travels along the driving mode; and to determine the load movement event of the unmanned mining truck in each sub-driving segment based on the current driving speed of the unmanned mining truck and the multiple sub-driving segments.

[0164] The transport control event module 23 is used to determine the bump pattern of the unmanned mining truck during the movement process based on multiple sub-movement items corresponding to the load movement event, the overall shape of the unmanned mining truck and multiple attitude data, and to determine the transport control event of the unmanned mining truck based on the bump pattern, the storage distribution map of the ore relative to the unmanned mining truck and the driving speed of the unmanned mining truck.

[0165] The dynamic control module 24 is used to determine the smooth transport coefficient of the unmanned mining truck based on the identification of transport control events, determine the dynamic control mode of the unmanned mining truck based on the smooth transport coefficient, the current attitude and overall shape of the unmanned mining truck, and dynamically control multiple driving data of the unmanned mining truck.

[0166] The collaborative control module 25 is used to identify unfinished transport tasks based on the task list of the unmanned mining truck, and to determine the collaborative control mode of the unmanned mining truck based on the unfinished transport task, the ore load of the unmanned mining truck, and the current status of the remaining unmanned mining trucks.

[0167] This application provides a control system for an unmanned mining truck, including the control device for the unmanned mining truck as described above.

[0168] This application provides a computer-readable storage medium with computer-executable instructions for causing a computer to perform the control method for an unmanned mining vehicle as described above. Those skilled in the art will understand that all or some steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, or suitable combinations thereof. Some or all physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which may include a computer storage medium or a non-transitory medium and a communication medium or a transient medium.

[0169] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A control method for an unmanned mining truck, characterized in that, include: Collect the driving status of the unmanned mining truck, and determine the driving mode of the unmanned mining truck based on the driving status and the terrain corresponding to the driving path. The driverless mining truck travels along this driving mode, and multiple sub-driving segments are determined based on the recognition of the driving path; The load movement events of the unmanned mining truck in each sub-driving segment are determined based on the current driving speed of the unmanned mining truck and multiple sub-driving segments. Based on the multiple sub-movement items corresponding to the load movement event, the overall shape of the unmanned mining truck and multiple attitude data, the bump pattern of the unmanned mining truck during the movement process is determined. Based on the bump pattern, the storage distribution map of the ore relative to the unmanned mining truck and the driving speed of the unmanned mining truck, the transportation control event of the unmanned mining truck is determined. The smooth transport coefficient of the unmanned mining truck is determined based on the identification of transport control events. The dynamic control mode of the unmanned mining truck is determined based on the smooth transport coefficient, the current attitude and overall shape of the unmanned mining truck, and multiple driving data of the unmanned mining truck are dynamically controlled. Based on the identification of the task list of the unmanned mining truck, an incomplete transportation task is determined, and the cooperative control mode of the unmanned mining truck is determined according to the incomplete transportation task, the ore load of the unmanned mining truck, and the current status of the remaining unmanned mining trucks. The process of determining the bump pattern of the unmanned mining truck during its movement based on multiple sub-movement items corresponding to the load movement event, the overall shape of the unmanned mining truck, and multiple attitude data, and determining the transport control event of the unmanned mining truck based on the bump pattern, the storage distribution map of the ore relative to the unmanned mining truck, and the driving speed of the unmanned mining truck, includes: Based on the identification of the movement event, multiple sub-movement items are identified. At the same time, multiple attitude data of the unmanned mining truck are collected. The first bump coefficient is determined based on the multiple sub-movement items and the multiple attitude data of the unmanned mining truck. The second bump coefficient is determined based on the overall shape of the multiple sub-movement items and the unmanned mining truck. The bump pattern of the unmanned mining truck during its movement is determined based on the mapping relationship between the first bump coefficient, the second bump coefficient, and the bump pattern. Multiple sub-bump parts of the unmanned mining truck are then identified based on the identification of the bump pattern. Collect the storage distribution map of ore relative to the unmanned mining truck, determine the ore transportation event based on multiple sub-bumps of the unmanned mining truck and the storage distribution map of ore relative to the unmanned mining truck, determine the dynamic driving event based on multiple sub-bumps of the unmanned mining truck and the driving speed of the unmanned mining truck, and determine the transportation control event of the unmanned mining truck based on the ore transportation event, the dynamic driving event and the overall shape of the unmanned mining truck.

2. The control method for an unmanned mining truck according to claim 1, characterized in that, The process of collecting the driving status of the unmanned mining truck and determining its driving mode based on the driving status and the terrain corresponding to the driving path includes: The autonomous driving process of the unmanned mining truck is monitored in real time, while the unmanned mining truck is carrying ore; multiple driving data and ore load data of the unmanned mining truck are collected; the driving status of the unmanned mining truck is determined based on the multiple driving data and ore load data of the unmanned mining truck. The corresponding driving path is determined based on the identification of the unmanned mining truck. The terrain corresponding to the driving path is determined based on the driving path, the location of the unmanned mining truck and the site distribution map of the ore transportation site. The driving mode of the unmanned mining truck is determined based on the driving status of the unmanned mining truck, the corresponding driving speed and the terrain corresponding to the driving path.

3. The control method for unmanned mining trucks according to claim 1, characterized in that, The driverless mining truck travels along this driving mode and determines multiple sub-driving segments based on the identification of the driving path; Based on the current speed of the unmanned mining truck and multiple sub-driving segments, the load movement events of the unmanned mining truck in each sub-driving segment are determined, including: The driving mode is collected, and multiple sub-driving modes are determined based on the recognition of the driving mode. The autonomous driving of the unmanned mining truck is determined based on the multiple sub-driving modes, the current position of the unmanned mining truck, and the corresponding mode trigger node. The unmanned mining truck drives along the driving mode. The system collects driving paths, identifies multiple path nodes based on the identified driving paths, determines multiple sub-driving road segments based on the node positions and shapes of the multiple path nodes and the terrain shapes corresponding to the driving paths, and monitors the current driving speed of the unmanned mining truck in real time. Based on the current driving speed of the unmanned mining truck and the multiple sub-driving road segments, the system determines the load movement events of the unmanned mining truck in each sub-driving road segment.

4. The control method for unmanned mining trucks according to claim 1, characterized in that, The process involves determining the stable transport coefficient of the unmanned mining truck based on the identification of transport control events, determining the dynamic control mode of the unmanned mining truck based on this stable transport coefficient, the current attitude and overall shape of the unmanned mining truck, and dynamically controlling multiple driving data of the unmanned mining truck, including: Collect transport control events, determine multiple sub-transport control contents based on the detection of transport control events, determine key transport control features based on the identification of each sub-transport control content, and collect multiple key transport control features; determine the smooth transport coefficient of the unmanned mining truck based on the feature location, corresponding feature shape and moving noise data of the unmanned mining truck based on the multiple key transport control features. The current attitude of the unmanned mining truck is collected. The first control parameter is determined based on the stable transport coefficient and the current attitude of the unmanned mining truck. The second control parameter is determined based on the stable transport coefficient and the overall shape of the unmanned mining truck. The dynamic control mode of the unmanned mining truck is determined based on the mapping relationship between the first control parameter, the second control parameter, and the dynamic control mode. Multiple dynamic control items are determined based on the identification of the dynamic control mode. Dynamic control events of multiple driving data are determined based on the multiple dynamic control items, the corresponding project nodes, and the unmanned mining truck, so as to dynamically control the multiple driving data of the unmanned mining truck.

5. The control method for an unmanned mining truck according to claim 1, characterized in that, The process involves identifying incomplete transport tasks based on the task list of the unmanned mining truck, and determining the cooperative control mode of the unmanned mining truck based on the incomplete transport task, the ore load of the unmanned mining truck, and the current status of the remaining unmanned mining trucks. This includes: Collect the task list of the unmanned mining truck, identify multiple transportation tasks based on the identification of the task list of the unmanned mining truck, and identify unfinished transportation tasks based on the multiple transportation tasks, the completed events of the unmanned mining truck and the current position of the unmanned mining truck. Based on the identification of incomplete transport tasks, the corresponding transport location is determined and the ore transport volume at that location is marked. Based on each transport location, the corresponding ore transport volume, and the ore load of the unmanned mining truck, the ore control task to be completed is determined. The location of the remaining unmanned mining trucks is collected, and their current status is marked. Based on the current status of the remaining unmanned mining trucks, their corresponding ore capacity, and the ore control tasks to be completed, the collaborative transportation content of the remaining unmanned mining trucks is determined. Based on the collaborative transportation content of the remaining unmanned mining trucks and their current transportation content, the collaborative control mode of the unmanned mining truck is determined.

6. A control device for an unmanned mining truck, characterized in that, The control device for the unmanned mining truck is applied to the control method for the unmanned mining truck as described in any one of claims 1-5, and the control device for the unmanned mining truck includes: The driving mode module is used to collect the driving status of the unmanned mining truck and determine the driving mode of the unmanned mining truck based on the driving status and the terrain corresponding to the driving path. The load movement event module is used to determine multiple sub-driving segments based on the identification of the driving path when the unmanned mining truck travels along the driving mode; and to determine the load movement event of the unmanned mining truck in each sub-driving segment based on the current driving speed of the unmanned mining truck and the multiple sub-driving segments. The transport control event module is used to determine the bump pattern of the unmanned mining truck during the movement process based on multiple sub-movement items corresponding to the load movement event, the overall shape of the unmanned mining truck and multiple attitude data, and to determine the transport control event of the unmanned mining truck based on the bump pattern, the storage distribution map of the ore relative to the unmanned mining truck and the driving speed of the unmanned mining truck. The dynamic control module is used to determine the smooth transport coefficient of the unmanned mining truck based on the identification of transport control events, determine the dynamic control mode of the unmanned mining truck based on the smooth transport coefficient, the current attitude and overall shape of the unmanned mining truck, and dynamically control multiple driving data of the unmanned mining truck. The collaborative control module is used to identify unfinished transport tasks based on the task list of the unmanned mining truck, and to determine the collaborative control mode of the unmanned mining truck based on the unfinished transport task, the ore load of the unmanned mining truck, and the current status of the remaining unmanned mining trucks.

7. A control system for an unmanned mining truck, characterized in that, Includes the control device for the unmanned mining truck as described in claim 6.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions for causing a computer to perform the control method for an unmanned mining vehicle as described in any one of claims 1 to 5.

Images