A method of loading, a system of loading, a loader, and an electronic device

Abstract

The application discloses a loading machine loading method, system, loading machine, electronic equipment and computer readable storage medium. The method comprises: in response to a loading instruction, starting a parent agent corresponding to a loading operation; in response to monitoring that the operation state of the loading machine meets the calling condition of a first sub-agent at a first operation time, the parent agent calling the first sub-agent and issuing a first operation task to the first sub-agent; the first sub-agent determining a first control amount of a first sub-action at the first operation time according to the first operation task and the operation state at the first operation time; and the first sub-agent controlling the loading machine to execute the first sub-action at the first operation time with the first control amount. The method solves the technical problem of low bucket filling rate and poor universality caused by making the loading machine track a fixed loading track to complete automatic loading in the prior art.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, and in particular, relates to a loading machine shoveling method and system, a loading machine, an electronic device, and a computer readable storage medium. BACKGROUND

[0002] As an indispensable equipment in construction and resource exploitation, the loading machine often faces extreme working conditions, and unmanned operation has attracted widespread attention because it can reduce the safety risk of operators.

[0003] Shoveling is a very key operation content in the operation task of the loading machine. At present, the unmanned shoveling operation of the loading machine usually relies on an automatic shoveling algorithm. The automatic shoveling algorithm completes automatic shoveling by making the loading machine track a fixed shoveling track. In the application process, the automatic shoveling algorithm has the following defects: first, the accuracy requirement for the shoveling point is high. If the shoveling point is not suitable, it is difficult to achieve a high bucket filling rate; second, it is not universal. For high-hardness materials, it is difficult to make the loading machine track the shoveling track.

[0004] Therefore, how to achieve a high bucket filling rate for different materials has become a technical problem to be solved in the process of unmanned operation of the loading machine. SUMMARY

[0005] The present application provides a loading machine shoveling method and system, a loading machine, an electronic device, and a computer readable storage medium to solve the technical problems of low bucket filling rate and poor universality caused by making the loading machine track a fixed shoveling track to complete automatic shoveling in the prior art.

[0006] In a first aspect, the embodiments of the present application provide a method for loading a shovel of a loader, the method comprising: in response to a loading instruction, starting a parent agent corresponding to a loading operation; wherein the loading operation comprises a plurality of sub-actions, the parent agent corresponds to a plurality of sub-agents that can be invoked, each of the sub-agents has a preset invocation condition, and the sub-agents are invoked by the parent agent to control execution of at least one of the sub-actions; in response to monitoring that a working state of the loader meets an invocation condition of a first sub-agent at a first working time, the parent agent invokes the first sub-agent and assigns a first working task to the first sub-agent, the first sub-agent is one of the plurality of sub-agents that meets the invocation condition, the first working task is determined by the parent agent according to the working state at the first working time and the first sub-agent, and the first working time is any working time in a process of executing the loading operation of the loader; the first sub-agent determines a first control amount of a first sub-action at the first working time according to the first working task and the working state at the first working time, the first sub-action is at least one of the sub-actions controlled and executed by the first sub-agent; and the first sub-agent controls the loader to execute the first sub-action at the first working time with the first control amount.

[0007] In a second aspect, the embodiments of the present application provide a loading machine shovel system, the system comprising: a positioning module, a sensor module, a parent agent module, a child agent module, and an actuator module; the positioning module comprising at least a first sensor and a positioning algorithm, the first sensor being configured on the loading machine to collect scene data of a working scene in which the loading machine is located at a first working time, the positioning algorithm being configured to calculate scene state information of the working scene according to the scene data, the first working time being any working time during which the loading machine performs the shovel operation; the sensor module comprising a plurality of second sensors, the second sensors being configured to collect self-state information of the loading machine at the first working time; the parent agent module comprising a parent agent, the parent agent being configured to determine a working state of the loading machine at the first working time according to the self-state information and the scene state information, and further configured to monitor whether the working state of the loading machine at the first working time meets a calling condition of a first child agent, and further configured to call the first child agent when it is determined that the working state of the loading machine at the first working time meets the calling condition of the first child agent, and further configured to determine a first working task for the first child agent according to the working state of the loading machine at the first working time, and further configured to assign the first working task to the first child agent, the first child agent being one of a plurality of child agents included in the child agent module and meeting the calling condition and the working state; the child agent module comprising a plurality of child agents, the first child agent being configured to determine a control amount of a first sub-action at the first working time according to the first working task and the working state of the loading machine at the first working time, the first sub-action being at least one of a plurality of sub-actions included in the shovel operation and controlled by the first child agent, and further configured to control the loading machine to perform the first sub-action at the first working time by a first actuator at the first control amount, the first actuator being at least one of a plurality of actuators included in the actuator module and corresponding to the first sub-action; and the actuator module comprising a plurality of actuators, the plurality of actuators being configured on the loading machine, the first actuator being configured to be controlled by the first child agent to enable the loading machine to perform the first sub-action at the first working time at the first control amount.

[0008] In a third aspect, the embodiments of the present application provide a loading machine, the loading machine being configured with the above shovel system.

[0009] In a fourth aspect, the embodiments of the present application provide an electronic device, comprising: a memory, a processor; the memory is configured to store one or more computer instructions; the processor is configured to execute the one or more computer instructions to implement the above method.

[0010] In a fifth aspect, a computer readable storage medium is provided, and one or more computer instructions are stored on the computer readable storage medium. The instructions are executed by a processor to perform the method described above.

[0011] Compared with the prior art, the loading machine loading method provided by the application divides the loading operation into multiple sub-actions, monitors the real-time operation state of the loading machine by a parent intelligent agent, calls a sub-intelligent agent according to a calling condition of the sub-intelligent agent, determines a control amount of a sub-action controlled by the called sub-intelligent agent by the called sub-intelligent agent, and controls the loading machine to perform the corresponding sub-action with the control amount. The method realizes hierarchical decision-making for the loading operation of the loading machine by calling different sub-intelligent agents by the parent intelligent agent and determining the control amount of the sub-action by the sub-intelligent agent. In the method, a fixed loading track is not tracked, the parent intelligent agent can adjust the called sub-intelligent agent according to the real-time operation state of the loading machine, and the sub-intelligent agent can adjust the control amount of the sub-action according to the real-time operation state of the loading machine. Therefore, it is an adaptive operation process and is not affected by the loading point and the hardness of the material, and can achieve a high full-bucket rate for various materials. The technical problem of low full-bucket rate and poor versatility caused by making the loading machine track a fixed loading track to complete automatic loading in the prior art is solved. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 FIG. 1 is a schematic diagram of a loading track when a loading operation is performed based on an automatic loading algorithm, provided by an embodiment of the application;

[0013] Figure 2 FIG. 2 is an application system diagram of the loading machine loading method, provided by an embodiment of the application;

[0014] Figure 3 FIG. 3 is a flowchart of the loading machine loading method, provided by a first embodiment of the application;

[0015] Figure 4 FIG. 4 is a schematic diagram of the loading machine loading system, provided by a second embodiment of the application;

[0016] Figure 5 FIG. 5 is a structural schematic diagram of the loading machine, provided by a third embodiment of the application;

[0017] Figure 6 FIG. 6 is a structural schematic diagram of the electronic device, provided by a fourth embodiment of the application. DETAILED DESCRIPTION

[0018] In the following description, numerous specific details are set forth to provide a thorough understanding of the present application. However, the present application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the spirit of the present application. Therefore, the present application is not limited by the specific embodiments disclosed below.

[0019] Loaders are essential construction equipment and play an irreplaceable role in the fields of construction and resource extraction. Currently, most loader operations are still carried out by experienced operators controlling the loader on-site. The operating environment of loaders is often extreme, with a lot of dust and extreme temperatures, and there are also risks of disasters such as collapses and mudslides. Long-term exposure to this environment by operators will not only seriously affect their physical health but also pose a great threat to their personal safety. Therefore, the unmanned operation of intelligent loaders has received extensive attention.

[0020] The operating tasks of loaders include excavation, loading, unloading, and hauling. Among them, shoveling is a very crucial link in the operating tasks of loaders. A higher bucket fill rate can improve operating efficiency and reduce operating costs. In the traditional operation process, the shoveling operation is completed by relying on the combined control of the boom, bucket, throttle, and brake by the operator. Among them, the interaction between the bucket and the material is involved, and achieving a high bucket fill rate requires the operator to have excellent skills and experience. Currently, in the unmanned operation process, the shoveling operation of loaders mainly relies on an automatic shoveling algorithm. The automatic shoveling algorithm enables the loader to follow a fixed shoveling trajectory and makes the boom and bucket perform fixed shoveling actions to complete automatic shoveling.

[0021] Figure 1 It is a schematic diagram of the shoveling trajectory during the shoveling operation based on the automatic shoveling algorithm provided by the embodiments of the present application.

[0022] As Figure 1 shown, the gray part represents the material, and the black line represents the shoveling trajectory (i.e., the movement trajectory of the bucket tip during the shoveling process). The shoveling process includes three stages: Stage 1, control the loader to move forward a distance L1, so that the bucket tip moves from point P0 to point P1 along the shoveling trajectory (black line); Stage 2, control the bucket to flip upward, and at the same time control the loader to continue to move forward a distance L2 in cooperation with the flipping action of the bucket, so that the bucket tip moves from point P1 to point P2 along the shoveling trajectory (black line). At this time, the bucket flips to the horizontal; Stage 3, control the boom to lift upward, so that the bucket tip moves from point P2 outside the material along the shoveling trajectory (black line), and the shoveling process ends.

[0023] The automatic shoveling algorithm has the following defects: First, it has a high requirement for the accuracy of the shoveling point. Once the shoveling point is inappropriate, it is very difficult to achieve a high bucket fill rate. As Figure 1As shown, when the perceived loading point is located to the left of P0 (such as P3), since the loader moves forward by a distance of L1 in stage one, the bucket tip cannot reach P1, and in subsequent stage two, the amount of material that can enter the bucket by turning the bucket over will decrease, and the full bucket rate will decrease. Second, it lacks universality. For materials with high hardness, it is difficult to make the loader track the loading trajectory, such as Figure 1 As shown, when the material has high hardness, it is impossible to control the loader to move forward by a distance of L1 in stage one, and it is impossible to track the loading trajectory (black line), so the loading operation for this material cannot be completed. Therefore, there is an urgent need for a loading method that has adaptability and robustness to achieve a high full bucket rate for different materials.

[0024] Therefore, the present application provides a loading method for a loader, which decomposes the loading operation into multiple sub-actions, monitors the real-time operation state of the loader by a parent agent, calls a sub-agent according to the calling condition of the sub-agent, and decides the control amount of the sub-action controlled by the called sub-agent to control the loader to perform the corresponding sub-action with the control amount. The method realizes hierarchical decision-making for the loading operation of the loader by calling different sub-agents by the parent agent and deciding the control amount of the sub-action by the sub-agent. In the method, a fixed loading trajectory is not tracked, the parent agent can adjust the called sub-agent according to the real-time operation state of the loader, and the sub-agent can adjust the control amount of the sub-action according to the real-time operation state of the loader. Therefore, it is an adaptive automatic operation process and is not affected by the loading point and the hardness of the material, and a high full bucket rate can be achieved for various materials.

[0025] The loading method for a loader, system, device, electronic equipment, and computer readable storage medium provided by the present application will be described in further detail below in conjunction with specific embodiments and the accompanying drawings.

[0026] Figure 2 is an application system diagram of the loading method for a loader provided by the present application. As shown in Figure 2 The system includes a loader end 201 and a server end 202. The loader end 201 at least includes a loader and a sensor deployed thereon, the sensor is used to capture scene state information of the operation scene and self-state information of the loader, and the server end 202 can be a processing device electrically connected with the loader end 201. The processing device can be deployed on the loader, or can be a server in communication connection with multiple loader ends 201, and the server is deployed with the loading method for a loader provided by the present application. When the server end 202 receives a loading instruction, the loader is controlled to perform a loading action based on the method.

[0027] The first embodiment of the present application provides a loading method of a loader, which realizes automatic adaptive loading of unmanned loader based on hierarchical decision principle, and ensures high bucket filling rate for various types of materials.

[0028] Figure 3 is a flowchart of the loading method of the loader provided by the present embodiment. The following detailed description of the loading method of the loader provided by the present embodiment is used to explain the technical solutions of the present application, and does not limit the actual use. Figure 3 The loading method of the loader provided by the present embodiment is described in detail. The embodiments described below are used to explain the technical solutions of the present application, and do not limit the actual use.

[0029] As shown in Figure 3 The loading method of the loader provided by the present embodiment includes the following steps S310 to S340:

[0030] Step S310, in response to a loading instruction, starting a parent agent corresponding to the loading operation; wherein the loading operation includes a plurality of sub-actions, the parent agent corresponds to a plurality of callable sub-agents, each sub-agent has a preset calling condition, and the sub-agent is used to control the execution of at least one sub-action after being called by the parent agent.

[0031] Step S320, in response to the monitoring that the working state of the loader meets the calling condition of the first sub-agent at the first working time, the parent agent calls the first sub-agent, and assigns the first working task to the first sub-agent, the first sub-agent is one of the plurality of callable sub-agents whose calling condition meets the working state, the first working task is determined by the parent agent according to the working state at the first working time and the first sub-agent, and the first working time is any working time in the process of the loader executing the loading operation.

[0032] Step S330, the first sub-agent determines the first control amount of the first sub-action at the first working time according to the first working task and the working state at the first working time, and the first sub-action is at least one sub-action controlled and executed by the first sub-agent.

[0033] Step S340, the first sub-agent controls the loader to execute the first sub-action with the first control amount at the first working time.

[0034] In a specific implementation, the above method is implemented by Figure 2The server 202 shown in the figure executes, and the server 202 is deployed with the loading machine's shoveling method provided by the embodiment, and is also deployed with a parent agent and multiple child agents that can be called by the parent agent. The parent agent completes the shoveling operation by calling the child agents. For example, after the loading machine's shovel contacts the ground, the parent agent will first call child agent A (responsible for aligning the loading machine) to ensure the safety of the shoveling action; to prevent the loading machine from being too close to the wall and causing safety problems, the parent agent will monitor the distance between the vehicle and the wall in real time according to the positioning information and map information obtained, and will call child agent B (responsible for suppressing the throttle output and adjusting the brake size according to the distance) when the distance is less than 0.5 meters.

[0035] In a specific implementation mode, the parent agent can be an application program edited by a developer, which at least includes a monitoring method of the loading machine's operation state, a calling method of the child agents, and a preset calling condition of each child agent. Specifically, during the execution of the loading machine's shoveling operation, the parent agent monitors the operation state of the loading machine at each operation time in real time, and determines whether the operation state of the loading machine meets the calling condition of a certain child agent according to the calling condition of each child agent. When it is determined that the operation state of the loading machine meets the calling condition of a certain child agent, the parent agent will call the child agent according to the calling method. The child agent has multiple implementation modes, such as a neural network model obtained based on machine learning, a traditional logic control method, etc. Different child agents can be set with one or more implementation modes to achieve their functions according to their operation tasks in the shoveling operation and the control-executed sub-actions. After being called by the parent agent, the child agent can calculate the sub-action control amount according to its implementation mode. For example, the moving child agent includes a path planning model and a path tracking model. When the parent agent calls the moving child agent, the moving child agent can calculate the moving path of the loading machine to the shoveling point based on the path planning model, and output the optimal control amount of the throttle sub-action, the control amount of the steering sub-action, the control amount of the brake sub-action, etc. based on the real-time operation state of the loading machine and the moving path as input data according to the path tracking model. The specific implementation mode of the parent agent and the child agent is not limited here.

[0036] The above steps are described in detail as follows:

[0037] For step S310:

[0038] In this step, the server starts the parent agent corresponding to the shoveling operation in response to the shoveling instruction.

[0039] The shovel instruction can be understood as a control information for conveying to the server that the loader needs to be controlled to perform the shovel operation. The shovel instruction can be sent by the operator to the server based on the remote control terminal, can be generated and sent by the client to the server in response to the triggering operation of the shovel operation control by the operator, or can be automatically generated and sent by the loader end to the server through the sensor sensing that the shovel operation is to be performed, and the specific implementation is not limited. The shovel instruction can include specific information related to the shovel operation, such as: position information of the shovel point, type information of the material, task quantity of the shovel operation, etc.

[0040] The shovel operation refers to the process of moving the loader to the shovel point and shoveling the material (such as soil, sand, coal, lime, ore, etc.) into the bucket, which includes a series of consecutive actions, such as approaching the material, inserting the bucket into the material, loading the bucket, lifting the arm to fill, etc., which are realized by cooperation of a plurality of basic sub-actions. In the embodiment, the shovel operation is divided into basic sub-actions, so that the sub-agent called by the parent agent can realize a higher fitness shovel operation by directly controlling the execution of these sub-actions. For example, the shovel operation includes throttle sub-action, brake sub-action, steering sub-action, boom lifting sub-action, bucket turning sub-action, etc.

[0041] In the embodiment, the sub-agent that can be called by the parent agent is determined by the specific operation content included in the shovel operation, and the increase or decrease of the operation content can be realized by adding or removing the sub-agent corresponding to the operation content. For example, under normal circumstances, the operation content of the shovel operation includes moving the vehicle body and shoveling, the moving of the vehicle body can be realized by the moving sub-agent and the safety sub-agent, and the shoveling can be realized by the power sub-agent, the boom sub-agent, the bucket sub-agent, and the safety sub-agent. When the operation content of shaking the material in the bucket is added in the shovel operation, only the bucket shaking sub-agent needs to be added, and the bucket shaking sub-agent controls the execution of the bucket turning sub-action. Therefore, the method provided in the embodiment can freely increase or decrease the sub-agent according to the actual operation content of the shovel operation, and the deployment is more flexible. In an optional implementation, the added sub-agent needs to be registered to the parent agent, and specifically, the calling condition of the added sub-agent is set in the parent agent.

[0042] The invocation conditions can be understood as the triggering conditions for the parent agent to invoke the child agent. These conditions may include the loader's own state conditions, the scene state conditions of the loader's operating environment, and the completion conditions of the preceding actions. The specific conditions need to be determined based on the child agent's operational content and the sub-actions controlled and executed by the child agent. For example, the invocation condition for the bucket child agent may be that the loader has moved to the vicinity of the shoveling point, such as the distance between the loader and the shoveling point being less than 6 meters; the invocation condition for the power child agent may be that the loader has reached the shoveling point and the bucket is in contact with the ground; the invocation condition for the bucket shaking child agent may be that the bucket has been tilted to a horizontal state and the throttle sub-action has ended.

[0043] In this embodiment, there is not a one-to-one correspondence between sub-agents and sub-actions. One sub-agent can control the execution of multiple sub-actions, and one sub-action can also be controlled by multiple sub-agents. For example, the sub-actions controlled by the motion sub-agent may include accelerator sub-actions, brake sub-actions, and steering sub-actions. The accelerator sub-action is also controlled by the power sub-agent, and the brake sub-action is also controlled by the safety sub-agent.

[0044] For step S320:

[0045] In this step, in response to the detection that the loader's operating status meets the calling conditions of the first sub-agent at the first operating moment, the parent agent calls the first sub-agent and sends the first operating task to the first sub-agent.

[0046] Once the parent agent is activated, it can monitor the loader's operating status in real time, determine whether the operating status meets the calling conditions of a certain child agent, and if so, call the child agent and issue the work task to it. In this embodiment, the first operating moment refers to any operating moment during the loader's material-shoveling operation, the first child agent refers to one of the multiple child agents that can be called by the parent agent whose calling conditions match the operating status, and the first work task refers to the work task issued to the first child agent at the first operating moment. Specifically, when the parent agent detects that the loader's operating status at the first operating moment meets the calling conditions of the first child agent, it calls the first child agent and issues the first work task to it.

[0047] The operation moment can be understood as a time node in the process of the loader performing the loading operation. The time node can be artificially set, for example, 1 second as a time node, and then 600 operation moments will be included in a 10-minute loading operation. In the embodiment, the time node is determined by the monitoring frequency of the parent agent, that is, each monitoring moment of the parent agent is regarded as a time node. The higher the monitoring frequency of the parent agent, the more time nodes, and the shorter the interval between the time nodes. For example, the monitoring frequency of the parent agent is 10 Hz, that is, the parent agent will monitor the operation state of the loader every 0.1 second in the process of the loading operation, and then 10 time nodes will be included in 1 second, and 6000 operation moments will be included in a 10-minute loading operation.

[0048] The operation task is determined by the parent agent according to the called sub-agent and the operation state of the loader when the sub-agent is called. After the operation task is issued to the sub-agent, the sub-agent will make a decision on the control amount of the sub-action with the goal of completing the operation task until the sub-agent is stopped.

[0049] For step S330:

[0050] In this step, the first sub-agent determines the first control amount of the first sub-action at the first operation moment according to the first operation task and the operation state at the first operation moment.

[0051] After the first sub-agent is called by the parent agent, the first sub-agent will make a decision on the control amount of the sub-action controlled and executed by the first sub-agent according to the operation task issued by the parent agent and the operation state. In the embodiment, the sub-action controlled and executed by the first sub-agent is defined as the first sub-action, and the control amount of the first sub-action determined at the first operation moment is defined as the first control amount.

[0052] The control quantity of the sub-action is autonomously decided by each sub-agent, and the implementation form of the sub-agent is different, and the decision method is also different. For example, the implementation form of the mobile sub-agent is a path planning model and a path tracking model, both of which are pre-trained neural network models. The path planning model can decide the moving path of the loader to the loading point according to the current position of the loader, the position of the loading point, and the map of the working scene. The path tracking model can take the working state of the loader and the moving path as input data, and output the control quantity of the throttle sub-action, the control quantity of the steering sub-action, and the control quantity of the brake sub-action. That is, the mobile sub-agent decides the control quantity of the sub-action based on the neural network model.

[0053] For step S340:

[0054] In this step, the first sub-agent controls the loader to perform the first sub-action with the first control quantity at the first working time.

[0055] After the first sub-agent determines the first control quantity of the first sub-action at the first working time, the first sub-agent controls the loader to perform the first sub-action with the first control quantity. Specifically, the first sub-agent controls the loader to perform the first sub-action through the actuators arranged on the loader. For example, the first sub-action is the throttle sub-action, the brake sub-action, or the steering sub-action. Then, the first sub-agent controls the loader to perform the first sub-action through the throttle pedal, the brake pedal, or the steering wheel of the loader.

[0056] In a specific implementation, the sub-agent outputs the control quantity of the decided sub-action in the form of a PWM (Pulse Width Modulation) signal quantity to the actuator corresponding to the sub-action, to realize automatic and accurate control of the sub-action.

[0057] Through the above steps, the loader completes the work content at a certain working time in the loading work process. Based on this method, the loader completes the work content at each working time, and the loading work is completed.

[0058] In the method provided in this embodiment, after the first sub-agent determines the control quantity of the first sub-action at the first working time (the working time when the first sub-agent is called), the decision process does not stop, and the control quantity of the first sub-action at each subsequent working time is determined according to the real-time working state of the loader, until the first work task is completed.

[0059] Based on this, in an optional implementation, the method provided in this embodiment can further include the following steps S351 to S352:

[0060] Step S351, the first sub-agent determines a second control quantity of the first sub-action at a second operation time according to the first operation task and the operation state at the second operation time, and the second operation time includes each operation time after the first operation time.

[0061] Step S352, the first sub-agent controls the loader to execute the first sub-action with the second control quantity at the second operation time.

[0062] In the embodiment, the second operation time represents each operation time after the first operation time, and the second control quantity represents the control quantity of the first sub-action decided by the first sub-agent at the second operation time. That is, after the first sub-agent determines the first control quantity of the first sub-action at the first operation time and controls the loader to execute the first sub-action with the first control quantity at the first operation time, the first sub-agent will continue to determine the second control quantity of the first sub-action at each subsequent operation time according to the real-time operation state of the loader at each subsequent operation time, and control the loader to execute the first sub-action with the corresponding second control quantity at each subsequent operation time.

[0063] For example, the moving sub-agent determines the first control quantity of the first sub-action as follows according to the first operation task and the operation state at T0 time (the first operation time): the throttle sub-action is 50% (in this example, in order to facilitate understanding, the throttle pedal / brake pedal down amplitude represents the control quantity of the throttle sub-action / brake sub-action), the brake sub-action is 0, and the steering sub-action is 0 (in this example, in order to facilitate understanding, the steering wheel rotation angle represents the control quantity of the steering sub-action). The moving sub-agent controls the loader to execute the first sub-action with the above first control quantity at T0 time through the throttle pedal, the brake pedal and the steering wheel, so as to realize straight acceleration. The moving sub-agent determines the second control quantity of the first sub-action as follows according to the first operation task and the operation state at T1 time (the second operation time): the throttle sub-action is 0, the brake sub-action is 0, and the steering sub-action is left turning 30 degrees. The moving sub-agent controls the loader to execute the first sub-action with the above second control quantity at T1 time through the throttle pedal, the brake pedal and the steering wheel, so as to realize deceleration turning. The moving sub-agent determines the second control quantity of the first sub-action as follows according to the first operation task and the operation state at T2 time (the second operation time): the throttle sub-action is 0, the brake sub-action is 50%, and the steering sub-action is 0 degrees. The moving sub-agent controls the loader to execute the first sub-action with the above second control quantity at T2 time through the throttle pedal, the brake pedal and the steering wheel, so as to realize deceleration straight. The moving sub-agent continues to determine the control quantity of the first sub-action and control the loader at each subsequent operation time until the loader reaches the loading point or the moving sub-agent is stopped by the parent agent.

[0064] In the method provided in the embodiment, the parent intelligent agent monitors the working state of the loader at the first working time, and after calling the first sub-intelligent agent, the monitoring and calling processes are not stopped, and the real-time working state of the loader is used to determine whether there is a sub-intelligent agent that can be called at each subsequent working time, and if there is, the sub-intelligent agent is called, so that the sub-intelligent agent performs the working task assigned by the parent intelligent agent while the first sub-intelligent agent performs the first working task.

[0065] Based on this, in an optional implementation, the method provided in the embodiment can further include the following steps S361 to S364:

[0066] Step S361, in response to the working state of the loader at the third working time meeting the calling condition of the second sub-intelligent agent, the parent intelligent agent calls the second sub-intelligent agent and assigns the second working task to the second sub-intelligent agent, the second sub-intelligent agent being any one of the multiple sub-intelligent agents that can be called and being in an uncalled state before the third working time, and the second working task being determined by the parent intelligent agent according to the working state at the third working time and the second sub-intelligent agent, the third working time being any working time after the first working time.

[0067] Step S362, the second sub-intelligent agent determines the third control amount of the second sub-action at the third working time according to the second working task and the working state at the third working time, the second sub-action being at least one sub-action controlled by the second sub-intelligent agent;

[0068] Step S363, the first sub-intelligent agent determines the fourth control amount of the first sub-action at the third working time according to the first working task and the working state at the third working time;

[0069] Step S364, the second sub-intelligent agent controls the loader to perform the second sub-action at the third working time with the third control amount; and / or, the first sub-intelligent agent controls the loader to perform the first sub-action at the third working time with the fourth control amount.

[0070] In the embodiment, the third working time represents any working time after the first working time when a sub-intelligent agent is called, and the second sub-intelligent agent represents a sub-intelligent agent called at the third working time, which can be any one of the multiple sub-intelligent agents that can be called by the parent intelligent agent and that is in an uncalled state before the third working time.

[0071] When the parent intelligent agent monitors the current working state of the loader to meet the calling condition of the second sub-intelligent agent at the third working time, the second sub-intelligent agent is called, and the working task to be completed by the second sub-intelligent agent is assigned to the second sub-intelligent agent. In this embodiment, the working task assigned to the second sub-intelligent agent is defined as a second working task. After the second sub-intelligent agent is called by the parent intelligent agent, the control amount of the sub-action controlled and executed by the second sub-intelligent agent is determined according to the second working task assigned by the parent intelligent agent and the working state. In this embodiment, the sub-action controlled and executed by the second sub-intelligent agent is defined as a second sub-action, and the control amount of the second sub-action determined at the third working time is defined as a third control amount. Since the first sub-intelligent agent is in the called state at the third working time, the first sub-intelligent agent also determines the control amount of the first sub-action at the third working time according to the first working task and the working state. In this embodiment, the control amount of the first sub-action at the third working time is defined as a fourth control amount. After the first sub-intelligent agent and the second sub-intelligent agent determine the control amounts of the corresponding sub-actions respectively, the loader is controlled to execute the corresponding sub-actions.

[0072] In an optional implementation, the first sub-intelligent agent and the second sub-intelligent agent can control the loader to execute the corresponding sub-actions synchronously, that is, the second sub-intelligent agent controls the loader to execute the second sub-action with the third control amount at the third working time, and the first sub-intelligent agent controls the loader to execute the first sub-action with the fourth control amount at the third working time. Specifically, when the first sub-action controlled and executed by the first sub-intelligent agent and the second sub-action controlled and executed by the second sub-intelligent agent do not repeat and conflict, the first sub-action and the second sub-action can be executed synchronously. For example, during the process in which the parent intelligent agent calls the moving sub-intelligent agent to control the loader to move to the loading point, when the distance between the loader and the loading point is less than 6 meters, the parent intelligent agent calls the boom sub-intelligent agent and the bucket sub-intelligent agent to control the bucket of the loader to be in contact with the ground. Since the sub-action controlled and executed by the moving sub-intelligent agent is the throttle sub-action, the brake sub-action and the steering sub-action, the sub-action controlled and executed by the boom sub-intelligent agent is the boom lifting sub-action, and the sub-action controlled and executed by the bucket sub-intelligent agent is the bucket turning sub-action, when the moving sub-intelligent agent determines the control amount A of the throttle sub-action, the control amount B of the brake sub-action, and the control amount C of the steering sub-action, the boom sub-intelligent agent determines the control amount D of the boom lifting sub-action, and the bucket sub-intelligent agent determines the control amount E of the bucket turning sub-action, since there is no repetition and conflict among these sub-actions, the moving sub-intelligent agent, the boom sub-intelligent agent, and the bucket sub-intelligent agent can control the loader to execute the throttle sub-action with the control amount A, the brake sub-action with the control amount B, the steering sub-action with the control amount C, the boom lifting sub-action with the control amount D, and the bucket turning sub-action with the control amount E at the same time.

[0073] In another optional implementation, the control of the first sub-agent and the second sub-agent on the loader can only be executed alternatively, that is, the second sub-agent controls the loader to execute the second sub-action with the third control amount at the third operation time, or the first sub-agent controls the loader to execute the first sub-action with the fourth control amount at the third operation time. How to determine whether the control of the loader at the third operation time is performed by the first sub-agent or the second sub-agent is planned in advance by the parent agent, that is, the parent agent also deploys the screening method of the sub-agent at the same time. For example, during the process in which the parent agent calls the moving sub-agent to control the loader to move to the loading point, the safety sub-agent is also called to ensure the safety of the moving process. The safety sub-agent senses whether there is an obstacle in the operation environment in real time through the sensor arranged on the loader, and judges whether the obstacle will be hit according to the moving direction, speed and steering angle of the loader. When the safety sub-agent judges that there is a risk of hitting the obstacle, the control amount of the throttle sub-action and the brake sub-action is output to reduce the moving speed of the loader or stop the moving of the loader. Assuming that the safety sub-agent decides the control amount F of the throttle sub-action, the control amount G of the brake sub-action, the control amount A of the throttle sub-action decided by the moving sub-agent, the control amount B of the brake sub-action and the control amount C of the steering sub-action at the third operation time, since the sub-actions controlled and executed by the safety sub-agent and the moving sub-agent are repeated, the safety sub-agent and the moving sub-agent cannot control the loader to execute the above sub-actions at the same time. Based on the screening method of the moving sub-agent and the safety sub-agent at the same time deployed in the parent agent, it is determined that the safety sub-agent takes over the throttle sub-action and the brake sub-action at the third operation time, that is, the safety sub-agent controls the loader to execute the throttle sub-action with the control amount F and the brake sub-action with the control amount G.

[0074] In an optional implementation provided in the embodiment, after the step of the first sub-agent controlling the loader to execute the first sub-action with the first control amount, the method provided in the embodiment can further include the following steps S371 to S372:

[0075] Step S371, the first sub-agent reports the execution result of the first sub-agent to the parent agent.

[0076] Step S372, the parent agent adjusts the calling state of the first sub-agent according to the execution result.

[0077] The first sub-agent not only makes a real-time decision on the control quantity of the first sub-action and controls the loader to execute the first sub-action with the control quantity in real time, but also reports the execution result of the loader after executing the first sub-action in real time. The parent agent can adjust the calling state of the first sub-agent according to the reported execution result. For example, the first sub-agent is a moving sub-agent, the moving sub-agent makes a decision on the control quantity A of the throttle sub-action, the control quantity B of the brake sub-action, and the control quantity C of the steering sub-action at the first work time, controls the loader to execute the throttle sub-action with the control quantity A, executes the brake sub-action with the control quantity B, and steers the sub-action with the control quantity C, and reports the distance between the loader and the material digging point to the parent agent after the loader executes the above sub-actions. The parent agent adjusts the calling state of the first sub-agent according to the distance between the loader and the material digging point.

[0078] In a specific implementation, the parent agent adjusts the calling state of the first sub-agent according to the execution result, including: in response to the parent agent judging that the execution result has met the first work task, stopping calling the first sub-agent.

[0079] The parent agent judges that the execution result has met the first work task according to the execution result reported by the first sub-agent, which means that the first sub-agent has completed the first work task issued by the parent agent. Therefore, the parent agent can stop calling the first sub-agent. For example, the first sub-agent is a moving sub-agent, and the moving sub-agent reports the distance between the loader and the material digging point to the parent agent. When the distance is 0, the first work task of the moving sub-agent is to move the loader to the material digging point. The distance between the loader and the material digging point is 0, which means that the execution result has met the first work task, and the moving sub-agent has completed the first work task. Therefore, the parent agent can stop calling the first sub-agent.

[0080] Of course, in another implementation, after the first sub-agent completes the first work task, the first sub-agent can perform a self-stopping operation and report the parent agent to inform the parent agent that the first sub-agent has completed the first work task and has automatically stopped the calling state.

[0081] In a specific implementation, the parent agent adjusts the calling state of the first sub-agent according to the execution result, including: in response to the parent agent judging that the execution result has not met the first work task, issuing an adjustment instruction for adjusting the first sub-action to the first sub-agent according to the work state at the fourth work time, and the fourth work time is the next work time of the first work time.

[0082] The parent agent determines that the execution result does not satisfy the first job task according to the execution result reported by the first sub-agent, and determines that the first sub-agent has not completed the job task issued by the parent agent, and the parent agent needs to maintain the calling state of the first sub-agent.

[0083] After the first sub-agent reports the execution result to the parent agent, the first job time has ended and enters the next job time of the first job time, which is defined as the fourth job time in this embodiment. The parent agent can issue an adjustment instruction for adjusting the first sub-action to the first sub-agent according to the job state of the loader at the fourth job time. For example, the first sub-agent is a moving sub-agent, and the moving sub-agent reports the distance between the loader and the material digging point to the parent agent. When the distance is 2 meters, the parent agent determines that the execution result does not satisfy the first job task (moving the loader to the material digging point), and the parent agent maintains the calling state of the moving sub-agent. The parent agent monitors that the bucket of the loader is not in contact with the ground. Therefore, the parent agent issues an adjustment instruction for reducing the speed to the moving sub-agent to ensure that the bucket is in contact with the ground before the loader reaches the material digging point. The moving sub-agent determines the control amount of the throttle sub-action, the control amount of the brake sub-action, and the control amount of the steering sub-action at the next job time based on the adjustment instruction.

[0084] The following provides an optional material digging job flow based on the material digging method described in this embodiment, which specifically includes the following steps S301 to S309:

[0085] Step S301, in response to a material digging instruction, starting a parent agent.

[0086] The material digging instruction contains position information of the material digging point, and the job target of the parent agent is to control the loader to go to the material digging point to dig material and achieve a high bucket filling rate.

[0087] Step S302, the parent agent calls a moving sub-agent and a safety sub-agent.

[0088] The moving sub-agent controls the loader to move to the material digging point by controlling the execution of the throttle sub-action, the brake sub-action, and the steering sub-action, and the safety sub-agent ensures that the loader does not collide with obstacles such as walls, other engineering vehicles, and people during movement by controlling the execution of the throttle sub-action and the brake sub-action.

[0089] The moving sub-agent comprises a path planning module and a path tracking module, and outputs control amounts of a throttle sub-action, a brake sub-action and a steering sub-action according to the real-time working state of the loader, controls the loader to perform the above-mentioned sub-actions with the output control amounts, and reports the execution result (the distance of the loader from the material loading point) of the above-mentioned sub-actions performed by the loader to the parent agent in real time.

[0090] The safety sub-agent is a neural network-based motion model, which judges whether the loader will collide with an obstacle in the subsequent working time according to the real-time working state of the loader, and takes over the throttle sub-action and the brake sub-action when the safety sub-agent judges that the loader is likely to collide with the obstacle, and outputs control amounts of the throttle sub-action and the brake sub-action in real time according to the distance between the loader and the obstacle, and controls the loader to perform the above-mentioned sub-actions with the above-mentioned control amounts.

[0091] In step S303, when the distance of the loader from the material loading point is less than 6 meters, the parent agent calls the boom sub-agent and the bucket sub-agent to place the bucket to be in contact with the ground.

[0092] The boom sub-agent is responsible for controlling the boom to lift to a preset angle, and two control modes are adopted, i.e. a position control based on PID and a speed control based on speed feedback, and the boom sub-agent switches to the appropriate control mode according to the hydraulic pressure of the boom (collected by a pressure sensor).

[0093] The bucket sub-agent is responsible for controlling the bucket to overturn to a preset angle, and a position control based on PID is adopted.

[0094] In step S304, when the distance of the loader from the material loading point is less than 2 meters, the parent agent monitors whether the bucket is in contact with the ground.

[0095] If not, an adjustment instruction of reducing the travel speed is issued to the moving sub-agent to ensure that the bucket can be placed to be in contact with the ground before the loader reaches the material loading point.

[0096] In step S305, when the bucket is placed to be in contact with the ground and the loader reaches the material loading point, the parent agent stops calling the moving sub-agent and calls the power sub-agent.

[0097] The power sub-agent controls the execution of the throttle sub-action to make the bucket of the loader continuously penetrate into the material, and the hydraulic pressure of the boom will continuously rise during this period.

[0098] When the safety sub-agent judges that the bucket is too close to the wall, the safety sub-agent takes over the throttle sub-action, outputs control amounts of the throttle sub-action and the brake sub-action in real time according to the distance between the loader and the wall, and controls the loader to adjust the throttle and the brake with the above-mentioned control amounts.

[0099] Step S306, when the hydraulic pressure of the boom reaches the preset value, the parent agent alternately calls the bucket sub-agent and the boom sub-agent to control the execution of the bucket turning sub-action and the boom lifting sub-action.

[0100] During this period, the power sub-agent controls the execution of the throttle sub-action to make as much material as possible enter the bucket.

[0101] Step S307, when the bucket turning reaches the horizontal state, the parent agent stops calling the power sub-agent and calls the material shaking sub-agent.

[0102] The material shaking sub-agent controls the execution of the bucket turning sub-action to achieve the effect of shaking the material in the bucket to make the material evenly distributed in the bucket.

[0103] Step S308, the parent agent calls the moving sub-agent, and the moving sub-agent controls the execution of the throttle sub-action, the brake sub-action, and the steering sub-action to make the loader drive away from the material loading point.

[0104] Step S309, the material loading operation is completed.

[0105] The first embodiment provides an optional material loading method of the loader, which has the following advantages: first, the method does not rely on accurate perception of the material loading point, and through the calling of different sub-agents by the parent agent and the decision of the control amount of the sub-action by the sub-agent, automatic adaptive material loading is realized, which can ensure that each material loading operation can achieve a high full bucket rate; second, the method does not rely on a fixed material loading track, and can be applied to various materials, and has universality; third, the method is convenient to deploy, and can adapt to different business demands by adding or reducing sub-agents.

[0106] It should be noted that the examples in the first embodiment are only for the explanation of the method described in the present application, and are not limited to the actual use. The material loading method of the loader provided in the present application includes but is not limited to the method described in the first embodiment.

[0107] The second embodiment of the present application provides a material loading system of a loader. Figure 4 is a schematic diagram of the material loading system of the loader provided in the present embodiment.

[0108] As Figure 4 shown, the material loading system of the loader provided in the present embodiment includes a positioning module 401, a sensor module 402, a parent agent module 403, a sub-agent module 404, and an actuator module 405.

[0109] The positioning module 401 includes at least a first sensor and a positioning algorithm. The first sensor is configured on the loader and is used to collect scene data of the working scene where the loader is located at the first working moment. The positioning algorithm is used to calculate the scene status information of the working scene based on the scene data. The first working moment is any working moment during the loader's material shoveling operation.

[0110] The sensor module 402 includes multiple second sensors, which are used to collect the loader's own status information at the first working moment.

[0111] The parent agent module 403 includes a parent agent, which is used to determine the working state of the loader at the first working moment based on its own state information and the scene state information; it is also used to monitor whether the working state of the loader at the first working moment meets the calling conditions of the first sub-agent; it is also used to call the first sub-agent when it is determined that the working state of the loader at the first working moment meets the calling conditions of the first sub-agent; it is also used to determine the first working task for the first sub-agent based on the working state of the loader at the first working moment; and it is also used to send the first working task to the first sub-agent, wherein the first sub-agent is any one of the multiple sub-agents included in the sub-agent module whose calling conditions meet the working state.

[0112] The sub-agent module 404 includes multiple sub-agents. The first sub-agent is used to determine the control quantity of the first sub-action at the first working time based on the first work task and the working state of the loader at the first working time. The first sub-action is at least one sub-action controlled by the first sub-agent among the multiple sub-actions included in the shoveling operation. It is also used to control the loader to execute the first sub-action with the first control quantity through the first actuator at the first working time. The first actuator is at least one actuator corresponding to the first sub-action among the multiple actuators included in the actuator module.

[0113] Optionally, the sub-agent module 404 includes at least the following sub-agents: mobile sub-agent 4041, power sub-agent 4042, boom sub-agent 4043, bucket sub-agent 4044, and safety sub-agent 4045; wherein,

[0114] The mobile sub-smart agent 4041 controls the execution of throttle, brake, and steering actions to enable the loader to move within the work environment.

[0115] The power sub-smart agent 4042 controls the execution of the throttle action to make the loader's bucket penetrate deeper into the material;

[0116] The boom sub-smart agent 4043 controls the execution of the boom lifting action so that the boom of the loader is raised or lowered.

[0117] The bucket agent 4044 controls the execution of the bucket tilting action to tilt the loader's bucket.

[0118] The safety sub-smart agent 4045 controls the execution of throttle and brake actions to maintain a safe distance between the loader and obstacles in the work environment.

[0119] Optionally, the sub-intelligent agent module 404 further includes: a data-shaking sub-intelligent agent 4046;

[0120] The material shaking agent 4046 controls the execution of the bucket rotation action to make the loader's bucket rotate.

[0121] The actuator module 405 includes multiple actuators configured on the loader. The first actuator is controlled by a first sub-agent to enable the loader to perform a first sub-action with a first control quantity at a first working moment.

[0122] The third embodiment of this application provides a loader. Figure 5 This is a structural schematic diagram of the loader provided in this embodiment.

[0123] like Figure 5 As shown, the loader provided in this embodiment includes: a loader body 501 and a shovel system 502 configured on the loader body 501.

[0124] The material shoveling system 502 includes: a positioning module 5021, a sensor module 5022, a parent intelligent agent module 5023, a child intelligent agent module 5024, and an actuator module 5025.

[0125] The positioning module 5021 includes at least a first sensor and a positioning algorithm. The first sensor is configured on the loader and is used to collect scene data of the working scene where the loader is located at the first working moment. The positioning algorithm is used to calculate the scene status information of the working scene based on the scene data. The first working moment is any working moment during the loader's material shoveling operation.

[0126] The sensor module 5022 includes multiple second sensors, which are used to collect the loader's own status information at the first working moment.

[0127] The parent intelligent agent module 5023 comprises a parent intelligent agent, which is configured to determine the working state of the loader at the first working time according to the state information of the parent intelligent agent and the scene state information; is further configured to monitor whether the working state of the loader at the first working time meets the calling condition of the first child intelligent agent; is further configured to call the first child intelligent agent when it is determined that the working state of the loader at the first working time meets the calling condition of the first child intelligent agent; is further configured to determine the first working task for the first child intelligent agent according to the working state of the loader at the first working time; and is further configured to issue the first working task to the first child intelligent agent, the first child intelligent agent being any one of the multiple child intelligent agents included in the child intelligent agent module and meeting the calling condition and the working state.

[0128] The child intelligent agent module 5024 comprises multiple child intelligent agents, and the first child intelligent agent is configured to determine the control amount of the first sub-action at the first working time according to the first working task and the working state of the loader at the first working time, the first sub-action being at least one sub-action of the multiple sub-actions included in the shovel operation and controlled by the first child intelligent agent; and is further configured to control the loader to execute the first sub-action at the first working time by the first child intelligent agent at the first control amount, the first child intelligent agent being at least one of the multiple child intelligent agents included in the child intelligent agent module and corresponding to the first sub-action.

[0129] The child intelligent agent module 5024 comprises multiple child intelligent agents, and the first child intelligent agent is configured to determine the control amount of the first sub-action at the first working time according to the first working task and the working state of the loader at the first working time, the first sub-action being at least one sub-action of the multiple sub-actions included in the shovel operation and controlled by the first child intelligent agent; and is further configured to control the loader to execute the first sub-action at the first working time by the first child intelligent agent at the first control amount, the first child intelligent agent being at least one of the multiple child intelligent agents included in the child intelligent agent module and corresponding to the first sub-action.

[0130] The fourth embodiment of the present application provides an electronic device, Figure 6 is a structural schematic diagram of the electronic device provided by the present embodiment.

[0131] As Figure 6 shown, the electronic device provided by the present embodiment comprises a memory 601 and a processor 602.

[0132] The memory 601 is configured to store computer instructions for executing the shovel method of the loader.

[0133] The processor 602 is configured to execute the computer instructions stored in the memory 601 to perform the following operations:

[0134] In response to the shovel instruction, a parent intelligent agent corresponding to the shovel operation is started; wherein the shovel operation comprises multiple sub-actions, the parent intelligent agent corresponds to multiple callable child intelligent agents, each of the child intelligent agents has a preset calling condition, and the child intelligent agent is configured to control the execution of at least one of the sub-actions after being called by the parent intelligent agent;

[0135] in response to monitoring that the working state of the loader meets the calling condition of the first sub-agent at the first working time, the parent agent calls the first sub-agent and assigns a first working task to the first sub-agent, the first sub-agent being one of the sub-agents that meets the calling condition and the working state, the first working task being determined by the parent agent according to the working state at the first working time and the first sub-agent, the first working time being any working time during the process of the loader performing the loading operation;

[0136] the first sub-agent determines a first control amount of the first sub-action at the first working time according to the first working task and the working state at the first working time, the first sub-action being at least one of the sub-actions controlled and executed by the first sub-agent;

[0137] the first sub-agent controls the loader to execute the first sub-action at the first working time with the first control amount.

[0138] Optionally, the following operations are further performed:

[0139] the first sub-agent determines a second control amount of the first sub-action at a second working time according to the first working task and the working state at the second working time, the second working time including each working time after the first working time;

[0140] the first sub-agent controls the loader to execute the first sub-action at the second working time with the second control amount.

[0141] Optionally, the following operations are further performed:

[0142] in response to monitoring that the working state of the loader meets the calling condition of the second sub-agent at the third working time, the parent agent calls the second sub-agent and assigns a second working task to the second sub-agent, the second sub-agent being any one of the sub-agents that is in an uncalled state before the third working time, the second working task being determined by the parent agent according to the working state at the third working time and the second sub-agent, the third working time being any working time after the first working time;

[0143] the second sub-agent determines a third control amount of the second sub-action at the third working time according to the second working task and the working state at the third working time, the second sub-action being at least one of the sub-actions controlled by the second sub-agent;

[0144] The first sub-agent determines a fourth control amount of the first sub-action at the third operation time according to the first operation task and the operation state at the third operation time.

[0145] The second sub-agent controls the loader to perform the second sub-action at the third operation time with the third control amount; and / or,

[0146] The first sub-agent controls the loader to perform the first sub-action at the third operation time with the fourth control amount.

[0147] Optionally, the following operations are further performed:

[0148] The first sub-agent reports an execution result of the first sub-agent performing the first sub-action to the parent agent;

[0149] The parent agent adjusts a calling state of the first sub-agent according to the execution result.

[0150] Optionally, the parent agent adjusts the calling state of the first sub-agent according to the execution result, including:

[0151] In response to the parent agent judging that the execution result has met the first operation task, the calling of the first sub-agent is stopped.

[0152] Optionally, the parent agent adjusts the calling state of the first sub-agent according to the execution result, including:

[0153] In response to the parent agent judging that the execution result has not met the first operation task, an adjustment instruction for adjusting the first sub-action is issued to the first sub-agent according to the operation state at a fourth operation time, the fourth operation time being a next operation time of the first operation time.

[0154] The fifth embodiment of the present application provides a computer readable storage medium, the computer readable storage medium includes computer instructions, the computer instructions are executed by the processor to implement the method described in the embodiments of the present application.

[0155] It should be noted that the "first", "second" and other relational terms in this paper are only used to distinguish one entity or operation from another entity or operation, and do not require or imply any actual relationship or order between the entities or operations. In addition, "including", "has", "contains" and "includes" and other similar forms of words are the same in meaning, and the end of any one or more items after any one of the above words is open, and any one of the above nouns does not mean that the one or more items have been listed exhaustively, or is limited to only these listed one or more items.

[0156] As used herein, the term "or" includes any and all possible combinations of the associated items, unless the possible combinations are mutually

[0157] It is noted that the above-described embodiments can be implemented in software or computer readable code, or a combination of hardware and software. If implemented in software, the software can be stored in a computer readable medium, as described above in store the programs of instructions. The software, when executed by the processor, can perform the disclosed method. The computing units and other functional units described in this disclosure can be implemented by hardware or software, or a combination of hardware and software. Those of ordinary skill in the art will also appreciate that the above-described modules / units can be combined into one module / unit, and each of the above-described modules / units can be further divided into a plurality of sub-modules / sub-units.

[0158] In the above detailed description, the embodiments have been described with reference to a number of specific details, which can vary from implementation to implementation. Certain adaptations and modifications of the described embodiments can occur to those skilled in the art. Other embodiments will be apparent to those of ordinary skill in the art from the disclosure, in light of the specific implementations disclosed, and are intended to be within the scope of the claims. The description and examples in this document are intended for purposes of illustration only and are not intended to limit the true scope and spirit of the application, which is defined by the claims. The sequence of steps shown in the drawings is also for illustrative purposes only and is not meant to imply any particular order of steps, sequence. Those of ordinary skill in the art will therefore appreciate that the steps can be performed in a different order.

[0159] In the drawings and detailed description of the application, exemplary embodiments are disclosed. However, many variations and modifications can be made to these embodiments. Accordingly, while specific terminology has been employed, such terminology is used in a generic and descriptive sense only, and not for purposes of limitation.

Claims

1. A method for shoveling material with a loader, characterized in that, The method includes: In response to a material shoveling command, the parent agent corresponding to the material shoveling operation is activated; wherein, the material shoveling operation includes multiple sub-actions, the parent agent corresponds to multiple callable sub-agents, each sub-agent has a preset calling condition, and the sub-agent is used to control the execution of at least one of the sub-actions after being called by the parent agent; In response to the detection that the working status of the loader meets the calling conditions of the first sub-intelligent agent at the first working moment, the parent intelligent agent calls the first sub-intelligent agent and sends the first working task to the first sub-intelligent agent. The first sub-intelligent agent is one of the plurality of callable sub-intelligent agents whose calling conditions meet the working status. The first working task is determined by the parent intelligent agent based on the working status at the first working moment and the first sub-intelligent agent. The first working moment is any working moment during the process of the loader performing the shoveling operation. The first sub-agent determines the first control quantity of the first sub-action at the first task and the task status at the first task time. The first sub-action is at least one of the sub-actions that are executed under the control of the first sub-agent. The first sub-agent controls the loader to execute the first sub-action with the first control quantity during the first working moment.

2. The method according to claim 1, characterized in that, The method further includes: The first sub-agent determines the second control quantity of the first sub-action at the second task time based on the task state at the first task time and the second task time. The second task time includes every task time after the first task time. The first sub-agent controls the loader to execute the first sub-action with the second control quantity during the second operation time.

3. The method according to claim 1, characterized in that, The method further includes: In response to the detection that the loader's operating status meets the calling conditions of the second sub-intelligent agent at the third operating time, the parent intelligent agent calls the second sub-intelligent agent and sends the second operating task to the second sub-intelligent agent. The second sub-intelligent agent is any one of the plurality of callable sub-intelligent agents that was in an uncalled state before the third operating time. The second operating task is determined by the parent intelligent agent based on the operating status at the third operating time and the second sub-intelligent agent. The third operating time is any operating time after the first operating time. The second sub-agent determines the third control quantity of the second sub-action at the third task time based on the second task and the task state at the third task time. The second sub-action is at least one of the sub-actions controlled by the second sub-agent. The first sub-agent determines the fourth control quantity of the first sub-action at the third task time based on the first task and the task state at the third task time. The second sub-agent controls the loader to execute the second sub-action with the third control quantity at the third operating moment; and / or, The first sub-agent controls the loader to execute the first sub-action with the fourth control quantity during the third operation time.

4. The method according to claim 1, characterized in that, The method further includes: The first sub-agent reports the execution result of the loader performing the first sub-action to the parent agent; The parent agent adjusts the calling state of the first child agent based on the execution result.

5. The method according to claim 4, characterized in that, The parent agent adjusts the calling state of the first child agent based on the execution result, including: In response to the parent agent determining that the execution result has satisfied the first job task, the invocation of the first child agent is stopped.

6. The method according to claim 4, characterized in that, The parent agent adjusts the calling state of the first child agent based on the execution result, including: In response to the parent agent determining that the execution result does not satisfy the first task, an adjustment instruction for adjusting the first sub-action is issued to the first child agent according to the task status at the fourth task time, wherein the fourth task time is the next task time after the first task time.

7. A loader's material-shoveling system, characterized in that, The system includes: a positioning module, a sensor module, a parent agent module, a child agent module, and an actuator module; The positioning module includes at least a first sensor and a positioning algorithm. The first sensor is configured on the loader and is used to collect scene data of the working scene where the loader is located at a first working moment. The positioning algorithm is used to calculate the scene state information of the working scene based on the scene data. The first working moment is any working moment during the loader's material shoveling operation. The sensor module includes multiple second sensors, which are used to collect the loader's own status information at the first working moment; The parent agent module includes a parent agent, which is used to determine the working state of the loader at the first working moment based on its own state information and the scene state information; it is also used to monitor whether the working state of the loader at the first working moment meets the calling conditions of the first sub-agent; it is also used to call the first sub-agent when it is determined that the working state of the loader at the first working moment meets the calling conditions of the first sub-agent; it is also used to determine a first working task for the first sub-agent based on the working state of the loader at the first working moment; and it is also used to send the first working task to the first sub-agent, wherein the first sub-agent is one of the multiple sub-agents included in the sub-agent module whose calling conditions match the working state; The sub-agent module includes multiple sub-agents. The first sub-agent is used to determine the control quantity of the first sub-action at the first working time based on the first work task and the working state of the loader at the first working time. The first sub-action is at least one of the multiple sub-actions included in the shoveling operation that is controlled by the first sub-agent. It is also used to control the loader to execute the first sub-action with the first control quantity through a first actuator at the first working time. The first actuator is at least one of the multiple actuators included in the actuator module that corresponds to the first sub-action. The actuator module includes multiple actuators configured on the loader. The first actuator is controlled by the first sub-agent to enable the loader to perform the first sub-action with the first control quantity at a first working moment.

8. The system according to claim 7, characterized in that, The sub-agent module includes at least the following sub-agents: a mobile sub-agent, a power sub-agent, a boom sub-agent, a bucket sub-agent, and a safety sub-agent; wherein... The mobile sub-agent controls the execution of throttle, brake, and steering actions to enable the loader to move within the work environment. The power sub-agent controls the execution of the throttle action to make the loader's bucket penetrate deeper into the material; The boom sub-smart agent controls the execution of the boom lifting action to raise or lower the boom of the loader. The bucket agent controls the execution of the bucket tilting action to tilt the bucket of the loader. The safety sub-agent controls the execution of throttle and brake actions to maintain a safe distance between the loader and obstacles in the work environment.

9. The system according to claim 8, characterized in that, The sub-intelligent agent module also includes: a material-shaking sub-intelligent agent; The material-shaking intelligent agent controls the execution of the bucket rotation action to make the loader's bucket rotate.

10. A loader, characterized in that, The loader is equipped with a material-shoveling system as described in any one of claims 7-9.

11. An electronic device, characterized in that, include: Memory, processor; The memory is used to store one or more computer instructions; The processor is configured to execute one or more computer instructions to implement the method as described in any one of claims 1-6.

12. A computer-readable storage medium storing one or more computer instructions thereon, characterized in that, When this instruction is executed by the processor, it performs the method as described in any one of claims 1-6.

Images