Intelligent path planning method and system for unmanned loader

Abstract

The invention discloses an intelligent path planning method and system for an unmanned loader, and the method comprises the steps: obtaining environment data through a sensor, carrying out the fusion processing of the environment data, and carrying out the rasterization processing of the fused environment data, and generating an occupied raster map; the method comprises the following steps: initializing a driving path of an unmanned loader, acquiring an initial global path, setting a global path planning objective function, and adjusting the initial global path by minimizing the global path planning objective function to generate a new global path; and track verification is carried out on the new global path, so that the path planning of the unmanned loader is completed.

Description

Technical Field

[0001] The present invention belongs to the technical field of driving path planning, and more specifically, relates to an intelligent path planning method and system for an unmanned loader. Background Art

[0002] As the core carrier of the intelligent transformation of construction machinery, the intelligent path planning system of an unmanned loader is a key technology to achieve autonomous operation and improve construction efficiency and safety. With the surge in demand for unmanned equipment in scenarios such as mines, ports, and mixing plants, the traditional manual driving mode faces efficiency bottlenecks and safety hazards. Therefore, it is urgent to achieve dynamic obstacle avoidance, multi-objective optimization, and complex terrain adaptability of unmanned loaders through intelligent path planning. Summary of the Invention

[0003] To solve the above technical problems, the present invention proposes an intelligent path planning method for an unmanned loader, which is characterized by including: Obtain environmental data through sensors, perform fusion processing on the environmental data, and generate an occupancy grid map by rasterizing the fused environmental data; Perform an initialization operation on the driving path of the unmanned loader, obtain an initial global path, set a global path planning objective function, and adjust the initial global path by minimizing the global path planning objective function to generate a new global path; Verify the trajectory of the new global path, thereby completing the path planning of the unmanned loader.

[0004] Furthermore, the global path planning objective function includes: , wherein, is the global path planning objective function of the initial global path , is the weight of the path length optimization function, is the path length optimization function of the initial global path , is the weight of the energy consumption optimization function, is the energy consumption optimization function of the unmanned loader of the initial global path , is the weight of the safety optimization function, is the safety optimization function of the initial global path , is the weight of the operation efficiency optimization function, is the operation efficiency optimization function of the initial global path .

[0005] Further, the initial global path path length optimization function includes: , wherein, is the number of path points, is the initial global path on the th path point, is the initial global path on the th path point, is the weight of the curvature, is the th path point's curvature, is the weight of the slope angle, is the th path point's slope angle.

[0006] Further, the energy consumption optimization function of the driverless loader for the initial global path includes: , wherein, is the speed of the driverless loader at the th path point, is the acceleration of the driverless loader at the th path point, is the weight of the traction force, is the traction force of the driverless loader at the th path point, is the weight of the resistance, is the resistance of the driverless loader at the th path point.

[0007]

[0007] Further, the safety optimization function of the initial global path includes: , wherein, is the distance from the th path point to the obstacle, is the weight of the turning radius, is the th path point's turning radius, is the weight of the slope angle, is the th path point's slope angle, is the weight of the speed difference between the dynamic obstacle and the driverless loader, is the speed of the dynamic obstacle at the th path point, is the speed of the driverless loader at the th path point.

[0008] Furthermore, the operation efficiency optimization function of the initial global path includes: , wherein, is the driving time of the driverless loader from the th path point to the th path point , is the adjustment factor of the operation efficiency optimization function, is the driving load of the driverless loader from the th path point to the th path point , is the mass of the driverless loader at the th path point .

[0009] Furthermore, generating an occupancy grid map by rasterizing the fused environmental data includes: decomposing the fused environmental data into uniform or non-uniform grid cells, and updating the grid state in real time according to the sensor data to adapt to environmental changes.

[0010] Furthermore, trajectory verification of the new global path is performed through static logic verification, dynamic simulation verification, and / or hardware-in-the-loop testing.

[0011] Furthermore, it also includes optimizing the local path of the driverless loader through a multi-objective optimization algorithm and assisting in screening the local path through a reinforcement learning model.

[0012] The present invention also proposes an intelligent path planning system for a driverless loader, including: a map generation module, configured to obtain environmental data through sensors, perform fusion processing on the environmental data, and generate an occupancy grid map by rasterizing the fused environmental data; a global path generation module, configured to perform an initialization operation on the driving path of the driverless loader, obtain an initial global path, set a global path planning objective function, and adjust the initial global path by minimizing the global path planning objective function to generate a new global path; A trajectory verification module is used to verify the trajectory of the new global path, thereby completing the path planning of the driverless loader.

[0013] Generally speaking, compared with the prior art, the above technical solutions conceived by the present invention have the following beneficial effects: 1. Double improvement in production efficiency and operation accuracy: Through the global path planning objective function and local path optimization, the optimal path is selected, effectively shortening the operation cycle and reducing invalid paths, and comprehensively improving the production efficiency.

[0014] 2. Significant improvement in safety: Through real-time obstacle detection and dynamic obstacle avoidance design, the collision risk can be avoided, and the operation safety of the equipment can be improved. Brief Description of the Drawings

[0015] Figure 1 is the flowchart of the method according to Embodiment 1 of the present invention; Figure 2 is the system structure diagram according to Embodiment 2 of the present invention; Detailed Embodiments

[0016] In order to better understand the above technical solutions, the above technical solutions will be described in detail below in conjunction with the accompanying drawings of the specification and specific embodiments.

[0017] The method provided by the present invention can be implemented in the following terminal environment. The terminal may include one or more of the following components: a processor, a storage medium, and a display screen. Among them, at least one instruction is stored in the storage medium, and the instruction is loaded and executed by the processor to implement the method described in the following embodiments.

[0018] The processor may include one or more processing cores. The processor connects various parts inside the entire terminal through various interfaces and lines, and executes various functions of the terminal and processes data by running or executing instructions, programs, code sets or instruction sets stored in the storage medium, and calling data stored in the storage medium.

[0019] The storage medium may include a random access memory (RAM), and may also include a read-only memory (ROM). The storage medium can be used to store instructions, programs, codes, code sets or instructions.

[0020] The display screen is used to display the user interfaces of various application programs.

[0021] In addition, those skilled in the art can understand that the structure of the above terminal does not constitute a limitation on the terminal. The terminal may include more or fewer components, or combine certain components, or have different component arrangements. For example, the terminal also includes components such as a radio frequency circuit, an input unit, a sensor, an audio circuit, and a power supply, which will not be elaborated here.

[0022] Embodiment 1 As Figure 1 , this embodiment proposes an intelligent path planning method for an unmanned loader, including: Step 101, obtain environmental data through sensors, perform fusion processing on the environmental data, and generate an occupancy grid map by rasterizing the fused environmental data; Regarding Step 101, specifically, it includes: 1. Environmental data collection, sensor fusion Use sensors such as lidar (LiDAR), cameras, millimeter-wave radars, and IMUs (inertial measurement units) to obtain raw environmental data (terrain elevation model: slope, gully, unevenness; 3D point cloud coordinates of fixed obstacles (stockpiles, buildings, equipment); ground roughness and passability assessment (such as gravel ground, muddy areas), etc.). For example, lidar provides high-precision point cloud data, and cameras are used to identify obstacles, etc.

[0023] 2. Real-time dynamic data acquisition In the mixing plant scenario, it is necessary to combine radar positioning and SLAM (simultaneous localization and mapping) technology to update the map information in real time to deal with moving obstacles (such as vehicles, personnel). The data content includes: the real-time positions, speeds, and accelerations of dynamic obstacles (personnel, vehicles, robotic arms).

[0024] 3. Map modeling Map modeling preparation: Occupancy grid map.

[0025] Grid division: Decompose the environmental data into uniform or non-uniform two-dimensional / three-dimensional grid cells, and label them as "free" or "obstacle" areas.

[0026] Dynamic update mechanism: Fusion of multi-sensor data through a probability model to update the grid state in real time to adapt to environmental changes.

[0027] Topological map: Node connection relationship: Represent the feasible path with key nodes (such as road intersections, loading and unloading points) and edges, which is suitable for rapid planning in large-scale scenarios.

[0028] Lightweight storage: Reduce redundant information and improve the planning efficiency by combining geometric features (such as the boundary of the loader operation area).

[0029] Data preprocessing: Noise filtering: Remove sensor noise (such as flying points in the point cloud) and invalid data (such as the reflective area of the camera).

[0030] Coordinate alignment: Unify the multi-sensor coordinate system to achieve data fusion.

[0031] Obstacle segmentation and classification: Use clustering algorithms to segment obstacles in the point cloud and distinguish static obstacles (buildings) from dynamic targets (moving vehicles).

[0032] Map construction and optimization: Grid map generation: Map the processed data into an occupancy probability grid to support collision detection for path search algorithms.

[0033] Topological structure extraction: Extract key path nodes from the grid map to construct a lightweight topological network.

[0034] Dynamic map update: Adopt an incremental update strategy, combined with Kalman filtering or particle filtering to correct the map in real time to adapt to environmental changes during the operation of the loader.

[0035] Step 102, perform an initialization operation on the driving path of the driverless loader, obtain the initial global path, set the global path planning objective function, and adjust the initial global path by minimizing the global path planning objective function to generate a new global path; Specifically, the global path planning objective function includes: , where, is the global path planning objective function of the initial global path , is the weight of the path length optimization function, is the initial global path is the path length optimization function of the initial global path is the weight of the energy consumption optimization function, is the initial global path is the energy consumption optimization function of the driverless loader for the initial global path is the weight of the safety optimization function, is the initial global path is the safety optimization function of the initial global path is the weight of the operation efficiency optimization function, is the initial global path is the operation efficiency optimization function of the initial global path

[0036] Specifically, the path length optimization function of the initial global path includes: , Among them, is the number of path points, is the initial global path the th path point, is the initial global path the th path point, is the weight of curvature, is the th path point's curvature, is the weight of slope angle, is the th path point's slope angle.

[0037] Specifically, the energy consumption optimization function of the driverless loader for the initial global path includes: , where, is the speed of the driverless loader at the th path point, is the acceleration of the driverless loader at the th path point, is the weight of traction force, is the traction force of the driverless loader at the th path point, is the weight of resistance, is the resistance of the driverless loader at the th path point.

[0038] Specifically, the safety optimization function of the initial global path includes: , where, is the th path point's distance to the obstacle, is the weight of turning radius, is the th path point's turning radius, is the weight of slope angle, is the th path point's slope angle, is the weight of the speed difference between the dynamic obstacle and the driverless loader, is the speed of the dynamic obstacle at the th path point, is the speed of the driverless loader at the th path point.

[0039] Specifically, the initial global path operation efficiency optimization function includes: , wherein, is the driving time of the driverless loader from the th path point to the th path point , is the adjustment factor of the operation efficiency optimization function, is the driving load of the driverless loader from the th path point to the th path point , is the mass of the driverless loader at the th path point .

[0040] Another global path planning function is also set in this embodiment, as follows: Objective: Generate a global path that meets the operation requirements such as loading, unloading, transportation, and obstacle avoidance.

[0041] Multi-objective optimization: Balance objectives such as path length, energy consumption, safety (such as staying away from dangerous areas), and operation efficiency (such as reducing the number of turns).

[0042] Dynamic adaptability: Support path replanning when the local map is updated.

[0043] First, perform map initialization, as follows: Input: Task objectives (loading and unloading points, transportation paths).

[0044] Processing: Load a high-precision static map and parse semantic layer information (passable area attributes, task-related areas).

[0045] Output: A rasterized or graph-structured global environment model.

[0046] Select the path search algorithm, that is, the global path planning function, as follows: Multi-objective optimization and constraint embedding, including: 1. Basic parameters (parameters to be considered): path length, steering angle change, path driving speed, slope (affecting energy consumption).

[0047] 2. Dynamic weights: Adjust the weights according to the task stage. For example, when fully loaded, prioritize minimizing the slope, and when unloaded, focus on the shortest path.

[0048] Constraint conditions, including: 1. Kinematic constraints: The minimum turning radius of the articulated loader (path curvature continuity is required).

[0049] 2. Stability constraints: Avoid exceeding the lateral slope limit (such as areas with a slope > 15°).

[0050] Path smoothing and semantic annotation: 1. Curve fitting: Use Bezier curves or spline curves to smooth the original path to ensure the feasibility of motion control; insert transition arcs in the turning area to meet the motion continuity of the articulated steering mechanism.

[0051] 2. Semantic enhancement: Annotate the attributes of path key points (such as deceleration sections in loading and unloading areas, power mode switching points on ramps); bind task instructions (such as triggering the lifting action at the material stacking point).

[0052] Dynamic update and replanning, including: 1. Trigger conditions: Updates to the dynamic layer of the map (such as new obstacles, changes in task objectives), and path following deviation exceeding the limit (such as positioning drift > 30 cm).

[0053] 2. Incremental replanning: Only locally adjust the affected path segments instead of re-searching the entire path, and combine historical path information to ensure the coherence of the replanned path.

[0054] Step 103: Perform trajectory verification on the new global path, thus completing the path planning of the driverless loader.

[0055] Specifically, perform trajectory verification on the new global path through static logic verification, dynamic simulation verification, and / or hardware-in-the-loop testing.

[0056] This embodiment gives the following examples to describe the trajectory verification, as follows: Trajectory verification: Trajectory verification is the core link to ensure the safety, feasibility, and reliability of the generated trajectory. It is necessary to verify whether the trajectory meets the dynamic environment adaptation, kinematic constraints, and task requirements through multi-dimensional testing and quantitative evaluation.

[0057] 1. Static logic verification Input: Trajectory data generated by planning (sequence of path points, speed curve, steering instructions).

[0058] Kinematic check: Verify the continuity of trajectory curvature and whether the minimum turning radius meets the limitations of the articulated vehicle body (such as the minimum turning radius ≥ 5 m) through geometric calculations.

[0059] Dynamics check: Calculate whether the trajectory acceleration and lateral force exceed the limit (such as the lateral acceleration ≤ 0.3g) based on the dynamics model.

[0060] Collision pre-check: Detect the distance between the trajectory and fixed obstacles (stockpiles, buildings) in the static map (safety distance ≥ 0.5 m).

[0061] Output: Logic pass / fail report, marking violation points (such as curvature mutation, collision risk).

[0062] 2. Dynamic simulation verification Input: Trajectory data; dynamic scenario library (moving obstacles, sudden terrain changes, sensor noise model).

[0063] Scenario injection: Simulate the following scenarios in the simulation platform: Normal scenario: The loader travels along the path and avoids vehicles moving at a constant speed.

[0064] Extreme scenario: Obstacles suddenly cut in, and partial sensor failure (such as loss of lidar point cloud).

[0065] Closed-loop simulation: Integrate the control model to simulate the impact of trajectory tracking error on actual motion.

[0066] Output: Simulation result video and data log.

[0067] Key metrics: Number of collisions, root mean square error of trajectory deviation, emergency braking success rate.

[0068] 3. Hardware-in-the-loop test Input: Real vehicle-mounted controller. Control instructions (steering angle, throttle / brake signal) output by the trajectory planning module.

[0069] Real-time signal simulation: Simulate sensor inputs (such as virtual lidar point cloud) through dSPACE or NI platform to verify whether the controller response meets expectations.

[0070] Fault injection: Artificially create communication delays (such as 100 ms), loss of control signals, and test the system's fault tolerance mechanism.

[0071] Output: Controller response time, fault recovery success rate (such as recovery time ≤ 200 ms after signal interruption).

[0072] 4. Real-field test Input: Planning system installed on the real vehicle. Preset test cases (narrow channel S-curve, heavy-load ramp, multi-vehicle cooperation scenario).

[0073] Performance acquisition: Record the deviation between the actual trajectory and the planned trajectory through in-vehicle IMU and lidar positioning equipment.

[0074] Extreme condition test: Heavy load stability: Verify the anti-rollover ability of the ramp trajectory when fully loaded (lateral tilt angle ≤ 10°); Dynamic obstacle avoidance: Test the emergency avoidance effect when an obstacle suddenly appears (simulating a person breaking in).

[0075] Output: Actual trajectory data, control instruction execution log.

[0076] Quantitative indicators: Lateral tracking error (≤ 15 cm), maximum steering angle overshoot (≤ 5%).

[0077] 5. Data analysis and iterative optimization Input: Raw data from simulation and real machine tests.

[0078] Abnormality attribution: Locate the root cause of the problem through data mining (such as trajectory jitter caused by sensor noise).

[0079] Parameter tuning: Adjust the weights of the planning algorithm (such as safety vs. smoothness), and regenerate the trajectory for verification.

[0080] Regression test: Ensure that the optimized trajectory passes the retest of historical failure scenarios.

[0081] Output: Algorithm iteration version, verification report (including pass rate, improvement points).

[0082] Verification link: Kinematic check: Whether the joint rotation angle exceeds the limit (such as the stroke of the bucket hydraulic cylinder).

[0083] Dynamic check: Whether the driving torque meets the requirements of ramp driving (torque compensation is triggered when the slope > 15°).

[0084] Control execution: Model predictive control: Optimize energy consumption while tracking the trajectory (the power saving rate of the electric loader > 18%).

[0085] Coordinated control: Linkage between the steering system and the bucket hydraulics (such as automatically adjusting the bucket inclination angle when steering).

[0086] Specifically, this embodiment further includes optimizing the local path of the driverless loader through multi-objective optimization algorithms (such as Non-dominated Sorting Genetic Algorithm II (NSGA-II), Pareto Simulated Annealing Algorithm (PSA), etc.), and assisting in screening local paths through reinforcement learning models (such as Q-Learning, Deep Q-Network (DQN), Deep Deterministic Policy Gradient (DDPG), etc.). Specific examples are as follows: Local trajectory generation and optimization: Local trajectory generation and optimization are key links for real-time dynamic obstacle avoidance and fine control. Based on the global path, combined with real-time perception data, safe, smooth and kinematically constrained local trajectories need to be generated.

[0087] Input: Global path, real-time obstacle information (such as temporary stockpiles) Trajectory segmentation: Decompose the movement of the bucket end into "lifting - translation - discharging" phases, and use a fifth-order polynomial to fit each segment of the trajectory.

[0088] Optimization objectives: Minimize the operation time and avoid bucket collisions (distance from the ground > 0.2m).

[0089] The specific steps for local trajectory generation and optimization include: 1. Real-time perception and dynamic environment modeling 2. Initial trajectory generation Candidate trajectory sampling: Based on the kinematic model: According to the current state of the loader (position, speed, articulation angle), generate multiple candidate trajectories through kinematic equations.

[0090] Example: Sample in the speed - steering space to generate a trajectory cluster with different speeds and steering angles.

[0091] Path - speed decoupling: Horizontally sample multiple candidate paths near the global path, and assign different speed curves to each path.

[0092] Trajectory parameterization: Use polynomials or spline curves to parameterize the trajectory to ensure the continuity of position, speed, and acceleration.

[0093] 3. Multi-objective trajectory optimization 3.1. Cost function design: Safety: The minimum distance from obstacles.

[0094] Smoothness: The change rate of trajectory curvature, acceleration / jerk constraints.

[0095] Fitting the global path: Penalty for the lateral deviation from the global reference line.

[0096] Energy consumption optimization: Power loss caused by slope changes and frequent steering.

[0097] 3.2. Constraint conditions: Kinematic constraints: Maximum steering angle, steering rate limit of the articulation mechanism.

[0098] Dynamic constraints: Maximum acceleration / deceleration, threshold of lateral acceleration for anti-rollover.

[0099] 3.3. Optimization algorithm: Numerical optimization method: Use QP (Quadratic Programming), SQP (Sequential Quadratic Programming) to solve the minimum value of the cost function.

[0100] 3.4. Convex optimization relaxation: Convert non-convex obstacle constraints into a convex space.

[0101] Reinforcement learning assistance: Rapidly screen high-quality candidate trajectories through pre-trained DRL models.

[0102] 4. Trajectory evaluation and selection Multi-index scoring: Perform weighted scoring on each candidate trajectory in dimensions such as safety, smoothness, and energy consumption, and introduce a risk probability model.

[0103] Real-time guarantee: Use parallel computing to shorten the evaluation time; set a trajectory scoring threshold to quickly eliminate high-risk trajectories.

[0104] 5. Trajectory execution and feedback correction Control instruction generation: Decompose the optimized trajectory into control instructions such as articulated steering angles, vehicle speeds, and bucket actions, and perform rolling optimization through model predictive control to compensate for execution errors.

[0105] Closed-loop feedback: Monitor the trajectory tracking deviation in real time and trigger local replanning or global path adjustment.

[0106] This embodiment also includes control execution logic, which is specifically as follows: Control execution: Control execution is a key link in converting the planned trajectory into actual mechanical actions. It is necessary to ensure that the loader safely and stably tracks the target trajectory through precise instruction issuance, dynamic adjustment, and closed-loop feedback.

[0107] 1. Trajectory decomposition and instruction conversion Input: Optimized local trajectory (including sequence of path points, speed curve, steering angle sequence).

[0108] Kinematics calculation: According to the kinematic model of the articulated loader, decompose the trajectory into front vehicle body steering angle, rear vehicle body following angle, and vehicle speed instructions.

[0109] Example: Calculate the change rate of the articulated angle through geometric relationships to ensure smooth steering.

[0110] Multi-mechanism coordination: Synchronize the bucket lifting and tilting actions with the driving trajectory (e.g., trigger the lifting instruction when reaching the loading and unloading point).

[0111] Output: Discrete control instruction sequence (steering angle, throttle / brake, hydraulic action instruction).

[0112] 2. Control algorithm drive Model predictive control: Rolling optimization: Use the current state as the initial condition to predict the control quantities for multiple future steps (e.g., 3 steps) and solve for the optimal instructions; Constraint embedding: Directly add kinematic (maximum steering rate) and dynamic (acceleration limit) constraints to the optimization problem.

[0113] PID and Adaptive Control: Lateral Control: Adjust the front wheel steering angle through PID to minimize the lateral path deviation; Longitudinal Control: Adjust the throttle / brake based on the speed curve to match the target acceleration; Parameter Adaptation: Dynamically adjust the control gain according to the ground adhesion coefficient (such as muddy road surface).

[0114] 3. Actuator Drive Steering System: The electro-hydraulic servo system converts the steering angle command into the displacement of the hydraulic cylinder to control the steering of the articulated vehicle body. Feedforward compensation (such as the steering system delay model) is introduced to improve the response speed.

[0115] Power System: Motor / Engine Torque Control: Send the throttle opening command through the CAN bus and dynamically adjust the output torque in combination with the slope information.

[0116] Hydraulic Actuator: Bucket Lifting / Tilting Control: Trigger the preset action sequence according to the trajectory semantic label (such as loading and unloading points).

[0117] 4. Closed-loop Feedback and State Estimation Multi-source Data Fusion: Fusion of IMU, wheel speed meter, and lidar positioning data, and real-time estimation of the vehicle pose through Kalman filtering.

[0118] Lidar / Visual Matching: Correct the positioning drift through point cloud registration.

[0119] Deviation Monitoring: Calculate the lateral / longitudinal error between the actual pose and the trajectory, and trigger replanning when the limit is exceeded (such as deviation > 30 cm).

[0120] Execution Health Detection: Monitor the actuator response delay (such as the difference between the steering angle command and the actual angle > 5°) and trigger fault handling.

[0121] 5. Safety Monitoring and Emergency Response Safety Boundary Calculation: Real-time calculate the minimum braking distance (based on the current speed and ground friction coefficient) and dynamically adjust the following distance.

[0122] Multi-level Fault Tolerance Strategy: First-level Response: Local trajectory fine-tuning (such as DWA resampling); Second-level Response: Global path replanning; Third-level Response: Emergency braking + system alarm (manual takeover).

[0123] Embodiment 2 As Figure 2 shown, this embodiment proposes an intelligent path planning system for an unmanned loader, including: A map generation module, configured to obtain environmental data through sensors, perform fusion processing on the environmental data, and generate an occupancy grid map by rasterizing the fused environmental data; Specifically, generating an occupancy grid map from the fused environmental data through rasterization processing includes: decomposing the fused environmental data into uniform or non-uniform grid cells and updating the grid status in real time according to the sensor data to adapt to environmental changes.

[0124] Generate a global path module for initializing the driving path of the driverless loader, obtaining an initial global path, setting a global path planning objective function, and adjusting the initial global path by minimizing the global path planning objective function to generate a new global path; Specifically, the global path planning objective function includes: , where is the global path planning objective function of the initial global path , is the weight of the path length optimization function, is the initial global path 's path length optimization function, is the weight of the energy consumption optimization function, is the initial global path 's energy consumption optimization function of the driverless loader, is the weight of the safety optimization function, is the initial global path 's safety optimization function, is the weight of the operation efficiency optimization function, is the initial global path 's operation efficiency optimization function.

[0125] Specifically, the path length optimization function of the initial global path includes: where , where is the number of path points, is the th path point on the initial global path , is the th path point on the initial global path , is the weight of the curvature, is the curvature of the th path point, is the weight of the slope angle, is the th slope angle of the path point.

[0126] Specifically, the initial global path Energy consumption optimization function of driverless loaders Includes: , Wherein, is the speed of the driverless loader at the th path point, is the acceleration of the driverless loader at the th path point, is the weight of the traction force, is the traction force of the driverless loader at the th path point, is the weight of the resistance, is the resistance of the driverless loader at the th path point.

[0127] Specifically, the safety optimization function of the initial global path Includes: Includes: , Wherein, is the distance from the th path point to the obstacle, is the weight of the turning radius, is the turning radius of the th path point, is the weight of the slope angle, is the slope angle of the th path point, is the weight of the speed difference between the dynamic obstacle and the driverless loader, is the speed of the dynamic obstacle at the th path point, is the speed of the driverless loader at the th path point.

[0128] Specifically, the operation efficiency optimization function of the initial global path Includes: Includes: , Wherein, is the travel time of the driverless loader from the th path point to the th path point , is the adjustment factor of the operation efficiency optimization function, is the travel time of the driverless loader from the th path point to the a path point traveling load For the driverless loader at the a path point mass on it.

[0129] A trajectory verification module is used to verify the trajectory of the new global path, thereby completing the path planning of the driverless loader.

[0130] Specifically, the trajectory of the new global path is verified through static logic verification, dynamic simulation verification, and / or hardware-in-the-loop testing.

[0131] Specifically, this embodiment further includes optimizing the local path of the driverless loader through a multi-objective optimization algorithm (such as Non-dominated Sorting Genetic Algorithm II (NSGA-II), Pareto Simulated Annealing Algorithm (PSA), etc.), and assisting in screening the local path through a reinforcement learning model (such as Q-Learning, Deep Q-Network (DQN), Deep Deterministic Policy Gradient (DDPG), etc.).

[0132] Embodiment 3 This embodiment of the present invention also proposes a storage medium storing multiple instructions for implementing the intelligent path planning method of a driverless loader.

[0133] Optionally, in this embodiment, the above storage medium can be located in any computer terminal in a computer terminal group in a computer network or in any mobile terminal in a mobile terminal group.

[0134] Optionally, in this embodiment, the storage medium is set to store program codes for executing the method steps of Embodiment 1.

[0135] Embodiment 4 This embodiment of the present invention also proposes an electronic device including a processor and a storage medium connected to the processor. The storage medium stores multiple instructions that can be loaded and executed by the processor so that the processor can execute the intelligent path planning of a driverless loader.

[0136] Specifically, the electronic equipment in this embodiment can be a computer terminal, and the computer terminal can include: one or more processors and a storage medium.

[0137] Among them, the storage medium can be used to store software programs and modules, such as an intelligent path planning for an unmanned loader in an embodiment of the present invention, and the corresponding program instructions / modules. The processor executes various functional applications and data processing by running the software programs and modules stored in the storage medium, that is, to implement the above-mentioned intelligent path planning for an unmanned loader. The storage medium can include a high-speed random access storage medium, and can also include a non-volatile storage medium, such as one or more magnetic storage systems, flash memory, or other non-volatile solid-state storage media. In some instances, the storage medium can further include a storage medium remotely disposed relative to the processor, and these remote storage media can be connected to the terminal through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0138] The processor can call the information and application programs stored in the storage medium through the transmission system to execute the method steps of Embodiment 1.

[0139] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments.

[0140] In the above embodiments of the present invention, the descriptions of each embodiment have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0141] In several embodiments provided by the present invention, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the system embodiments described above are only illustrative. For example, the division of the units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of units or modules can be in electrical or other forms.

[0142] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0143] In addition, the functional units in each embodiment of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0144] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, read-only storage media (ROM, Read-Only Memory), random access storage media (RAM, Random Access Memory), mobile hard disks, magnetic disks, or optical discs.

[0145] Obviously, the above embodiments are merely examples for clear illustration and not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or variations can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. And the obvious changes or variations derived therefrom still fall within the protection scope of the present invention.

Claims

1. An intelligent path planning method for an unmanned loader, characterized in that, Including: Obtain environmental data through sensors, perform fusion processing on the environmental data, and generate an occupancy grid map by rasterizing the fused environmental data; Perform an initialization operation on the driving path of the driverless loader, obtain an initial global path, set a global path planning objective function, and adjust the initial global path by minimizing the global path planning objective function to generate a new global path; Perform trajectory verification on the new global path, thereby completing the path planning of the driverless loader.

2. The intelligent path planning method for an unmanned loader according to claim 1, wherein The global path planning objective function includes: ， Among them, is the initial global path of the global path planning objective function, is the weight of the path length optimization function, is the initial global path of the path length optimization function, is the weight of the energy consumption optimization function, is the initial global path of the energy consumption optimization function of the driverless loader, is the weight of the safety optimization function, is the initial global path of the safety optimization function, is the weight of the operation efficiency optimization function, is the initial global path of the operation efficiency optimization function.

3. The intelligent path planning method for an unmanned loader according to claim 1, characterized in that, Initial global path Path length optimization function including: ， wherein, is the number of path points, is the initial global path the th path point, is the initial global path the th path point, is the weight of curvature, the th curvature of the path point, is the weight of slope angle, the th slope angle of the path point.

4. The intelligent path planning method for an unmanned loader according to claim 2, wherein Initial global path Energy consumption optimization function of driverless loaders including: ， Among them, is the speed of the driverless loader at the th path point, is the acceleration of the driverless loader at the th path point, is the weight of the traction force, is the traction force of the driverless loader at the th path point, is the weight of the resistance, is the resistance of the driverless loader at the th path point.

5. The intelligent path planning method for an unmanned loader according to claim 2, wherein Initial global path Security optimization function including: ， Among them, is the distance from the th waypoint to the obstacle, is the weight of the turning radius, is the th turning radius of the waypoint, is the weight of the slope angle, is the th slope angle of the waypoint, is the weight of the speed difference between the dynamic obstacle and the driverless loader, is the speed of the dynamic obstacle at the th waypoint, is the speed of the driverless loader at the th waypoint.

6. The intelligent path planning method for an unmanned loader according to claim 2, wherein, Initial global path Job efficiency optimization function including: ， Among them, is the travel time of the driverless loader from the th path point to the th path point . is the adjustment factor of the operation efficiency optimization function. is the travel load of the driverless loader from the th path point to the th path point . is the mass of the driverless loader at the th path point .

7. The intelligent path planning method for an unmanned loader according to claim 1, characterized in that Generating an occupancy grid map by rasterizing the fused environmental data includes: decomposing the fused environmental data into uniform or non-uniform grid cells, and updating the grid status in real time according to sensor data to adapt to environmental changes.

8. The intelligent path planning method for an unmanned loader according to claim 1, characterized in that, Perform trajectory verification on the new global path through static logic verification, dynamic simulation verification, and / or hardware-in-the-loop testing.

9. The intelligent path planning method for an unmanned loader according to claim 1, wherein It also includes optimizing the local path of the driverless loader through a multi-objective optimization algorithm and assisting in screening the local path through a reinforcement learning model.

10. An intelligent path planning system for an unmanned loader, characterized in that, Including: A map generation module for obtaining environmental data through sensors, performing fusion processing on the environmental data, and generating an occupancy grid map by rasterizing the fused environmental data; A global path generation module for performing an initialization operation on the driving path of the driverless loader, obtaining an initial global path, setting a global path planning objective function, and adjusting the initial global path by minimizing the global path planning objective function to generate a new global path; A trajectory verification module for performing trajectory verification on the new global path, thereby completing the path planning of the driverless loader.

Images