Excavator Operation Pattern Detection Using Sliding Training Windows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods face challenges in securing sufficient training information for excavator operations, particularly due to limited operation patterns and difficulty in capturing the unique control information, leading to inefficiencies in training and analysis.
Innovation Solution
A method and device that acquire excavator control signals, including joystick and pump pressure signals, over time, and utilize sliding time periods to determine operation patterns by combining first and second training information, with post-processing to identify main operation patterns based on frequency and angular changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If training information is acquired simply in proportion to analysis time, then the analysis process is straightforward, but it is difficult to secure a sufficient amount of training information
Solution Approach 1:
The system performs preliminary actions by acquiring and storing excavator control signals continuously over time before actual training is needed. This allows the accumulation of sufficient training data in advance, so that when training is required, there is already adequate information available without needing to spend additional time collecting data.
Solution Approach 2:
The system dynamically adjusts the time period used for training by introducing a sliding window mechanism. Instead of using a fixed time period, the system can adaptively select different time periods based on the amount of training information needed and the characteristics of the excavator operations, allowing flexible optimization between data quantity and acquisition time.
2Measurement precision
If a specific method of analyzing unique control information is required, then accurate analysis can be achieved, but the complexity of the analysis method increases
Solution Approach 1:
The system segments the excavator control signals into distinct time periods (first time period and second time period) and processes each segment separately. By dividing the continuous control signal data into manageable segments, the system can apply analysis methods to each segment independently, reducing the overall complexity while maintaining accuracy through systematic processing of divided data portions.
Solution Approach 2:
The system acquires more control signals than strictly necessary by using overlapping time periods. The first and second time periods partially overlap, meaning some data is processed multiple times. This excessive action ensures sufficient training information is available and improves accuracy by providing redundant data for analysis, while the modular segmented approach keeps the processing complexity manageable.
3Quantity of substance
If the time period for training is extended to cover complete operation cycles, then sufficient training information is secured, but the training time and computational load increase
Solution Approach 1:
The system dynamically adjusts the time period duration and sliding interval based on the excavator's operation cycle characteristics. By making the time period flexible rather than fixed, the system can optimize the balance between capturing complete operation cycles (for sufficient training information) and minimizing training time. The dynamic adjustment allows adapting to different operation speeds and cycle lengths.
Solution Approach 2:
The system performs preliminary acquisition of control signals covering multiple operation cycles before training begins. This preliminary action ensures that when training starts, there is already a buffer of complete operation cycles available, eliminating the need to wait for complete cycles during the actual training process, thus reducing training time while maintaining completeness.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
Disclosed are a method for determining the operation pattern of an excavator using training information, a device for performing the method, and a recording medium. According to an embodiment, the method for determining the operation pattern of an excavator using training information comprises the steps of: acquiring, over time, excavator control signals including joystick control signals for controlling an excavator; acquiring first training information for determining the operation pattern of the excavator on the basis of a first excavator control signal, corresponding to a first time period among the excavator control signals acquired over time; determining a second time period partially overlapping with the first time period and acquired by sliding the first time period by a unit time; acquiring second training information for determining the operation pattern of the excavator on the basis of a second excavator control signal corresponding to a second time period, among the excavator control signals; and determining the operation pattern on the basis of the first training information and the second training information.