KR-102962050-B1 - INTELLIGENT REAL TIME EXCAVATION OPTIMIZATION APPARATUS AND METHOD FOR EARTHMOVING EQUIPMENT

KR102962050B1KR 102962050 B1KR102962050 B1KR 102962050B1KR-102962050-B1

Abstract

The present invention relates to a soil-adaptive autonomous excavation control device and method for earthmoving equipment based on reinforcement learning. An excavation control device according to one embodiment of the present invention may include a soil state modeling unit that models excavation resistance force according to soil parameters; a local trajectory generation unit that generates a local trajectory of an excavator by inputting soil parameters and excavation resistance force into a pre-prepared local trajectory control model based on reinforcement learning; a global path generation unit that generates a global path that flattens an excavation area to a target excavation height and minimizes the movement cost and direction change cost of the excavator using a pre-prepared global trajectory model; and an integrated control unit that controls the local trajectory to be executed on the global path.

Inventors

정소이
신민규
조준형

Assignees

아주대학교산학협력단

Dates

Publication Date: 20260507
Application Date: 20251001

Claims (15)

As an excavation control device for controlling an excavator, A soil state modeling unit that models excavation resistance according to soil parameters; A local trajectory generation unit that generates a local trajectory of the excavator by inputting the above soil parameters and the above excavation resistance force into a pre-prepared local trajectory control model based on reinforcement learning; A global path generation unit that uses a pre-prepared global trajectory model to level the excavation area to a target excavation height and generates a global path that minimizes the movement cost and direction change cost of the excavator; and It includes an integrated control unit that controls the local trajectory to be executed on the global path; and The above local trajectory control model is, A first learning stage that randomly changes soil parameters and target excavation height, and An excavation control device characterized by being learned through a curriculum learning process that includes a second learning stage that sequentially increases the difficulty level by reflecting the actual work environment.
In paragraph 1, The above integrated control unit transmits work information, including the excavator's working conditions and excavation results, to the soil condition modeling unit, and An excavation control device characterized in that the above-mentioned soil condition modeling unit models the excavation resistance force by reflecting the above-mentioned work information.
In paragraph 1, The above local trajectory control model is, The above soil parameters and current excavation height are defined as the state, and Defines the control angle that controls the cutting motion of the excavation bucket as a behavior, An excavation control device characterized by being learned through reinforcement learning that defines a value reflecting the excavation amount, excavation resistance, and the deviation between the target excavation height and the current excavation height as a reward.
In paragraph 3, The above compensation is, Excavation volume obtained in real-time at each stage of the excavation operation. Stage compensation calculated by weighting the excavation resistance and the deviation between the target excavation height and the current excavation height, and An excavation control device characterized by including a termination compensation calculated based on the deviation of the average excavation height of the excavation area and the target excavation height after the completion of the excavation work, and whether the bucket capacity is exceeded.
In paragraph 3, The above control angle is, Excavation control device characterized by including an angle between the ground failure surface and the terrain surface and an angle between the bucket cutting plate and the terrain surface.
delete
In paragraph 1, The above global path generation unit is, A cost function is defined based on Dynamic Programming (DP), the travel distance of the excavator, the number of direction changes, the excavation resistance, and the deviation between the target excavation height and the current excavation height. Excavation control device characterized by generating a global path that minimizes the above cost function.
In paragraph 1, The above global path generation unit is, Excavation control device characterized by generating a global path by applying a zigzag pattern coverage path.
A method for controlling an excavator, performed in an excavation control device, wherein Soil state modeling step for modeling excavation resistance according to soil parameters; A local trajectory generation step of generating a local trajectory of the excavator by inputting the soil parameters and the excavation resistance force into a pre-prepared local trajectory control model based on reinforcement learning; A global path generation step that uses a pre-prepared global trajectory model to level the excavation area to a target excavation height and generates a global path that minimizes the movement cost and direction change cost of the excavator; and It includes an integrated control step that controls the local trajectory to be executed on the global path; and The above local trajectory control model is, A first learning stage that randomly changes soil parameters and target excavation height, and Excavation control method characterized by being learned through a curriculum learning process that includes a second learning stage that sequentially increases the difficulty level by reflecting the actual work environment.
In Paragraph 9, The above integrated control step transmits work information, including the excavator's working conditions and excavation results, to the soil condition modeling unit, and Excavation control method characterized in that the above soil condition modeling step models the excavation resistance by reflecting the above work information.
In Paragraph 9, The above local trajectory control model is, The above soil parameters and current excavation height are defined as the state, and Defines the control angle that controls the cutting motion of the excavation bucket as a behavior, A method for controlling excavation, characterized by being learned through reinforcement learning that defines a value reflecting the excavation amount, excavation resistance, and the deviation between the target excavation height and the current excavation height as a reward.
delete
In Paragraph 9, The above global path generation step is, A cost function is defined based on Dynamic Programming (DP), the travel distance of the excavator, the number of direction changes, the excavation resistance, and the deviation between the target excavation height and the current excavation height. A drilling control method characterized by generating a global path that minimizes the above cost function.
In Paragraph 9, The above global path generation step is, Excavation control method characterized by generating a global path by applying a zigzag pattern coverage path.
A recording medium on which a computer program for implementing the excavation control method according to paragraph 9 is recorded.

Description

Intelligent Real-Time Excavation Optimization Apparatus and Method for Earthmoving Equipment The present invention relates to a device and method for real-time excavation optimization of intelligent earthmoving equipment, and more specifically, to a soil-adapting autonomous excavation control device and method. In the field of automated excavation technology, predefined models or simple rule-based control methods are primarily used to automate excavator operations. These technologies aim to achieve operational automation by controlling the excavator's trajectory according to a preset path. However, these methods fail to reflect the physical characteristics of the soil in real time and have limitations in applicability in actual working environments because they are designed based on simplified soil models. In particular, a lack of adaptability and precision under various ground conditions leads to reduced operational efficiency. Existing technology has structural limitations in that it cannot reflect the excavator's operating status and the physical characteristics of the soil in real time. This prevents the sufficient consideration of various variables and environmental factors occurring during excavation, consequently reducing operational precision and efficiency. In particular, under complex ground conditions, there is a higher likelihood of inaccurate excavator control, which can lead to a decline in work quality. Furthermore, these technical limitations can result in unnecessary energy consumption and wasted time during the work process. Additionally, existing technology suffers from a lack of connectivity between local and global control. Since the excavator's work path and bucket control are performed separately, operational consistency is compromised, negatively impacting work efficiency. This structural separation makes coordination and optimization during the work process difficult, resulting in increased working time and energy consumption. Consequently, the current level of technology still faces significant limitations in achieving efficient and precise excavation operations under various ground conditions. FIG. 1 is a diagram illustrating the schematic configuration of an excavation control device according to one embodiment of the present invention. FIG. 2 is an exemplary diagram illustrating the operation process of an excavation control device according to one embodiment of the present invention. Figure 3 is an example diagram showing the cutting resistance generated when a bucket cuts soil. FIG. 4 is an algorithm for the process of an excavation control device according to an embodiment of the present invention learning the local trajectory of an excavator using PPO. FIG. 5 is a flowchart of an excavation control method according to one embodiment of the present invention. The following detailed description of the invention refers to the accompanying drawings, which illustrate specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that various embodiments of the invention are different but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the invention in relation to one embodiment. It should also be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the invention. Accordingly, the following detailed description is not intended to be limiting, and the scope of the invention is limited only by the appended claims, including all equivalents to those claimed therein, provided appropriately described. Similar reference numerals in the drawings refer to the same or similar functions across various aspects. The components according to the present invention are defined by functional distinction rather than physical distinction, and can be defined by the functions each performs. Each component may be implemented as hardware or as program code and processing units that perform each function, and the functions of two or more components may be included and implemented in a single component. Therefore, it should be noted that the names assigned to the components in the following embodiments are not intended to physically distinguish each component but are assigned to imply the representative function performed by each component, and that the technical concept of the present invention is not limited by the names of the components. Preferred embodiments of the present invention will be described in more detail below with reference to the drawings. FIG. 1 is a diagram illustrating the schematic configuration of an excavation control device according to one embodiment of the present invention. The excavati