AGV Traveling Parameter Tuning Using Bayesian Experiment Design
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automatic transport systems require significant time and effort for optimizing traveling parameters, as existing methods do not efficiently reduce the work and time needed for parameter optimization, despite using a large number of parameters for control.
Innovation Solution
A traveling parameter optimization system that includes an experimental traveling design creation device, a traveling instruction device, an acquisition device, an evaluation value calculation device, and a traveling parameter optimization device, utilizing Bayesian optimization and Gaussian process regression to efficiently optimize parameters based on experimental data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional traveling control methods are used with a large number of parameters, then control precision is improved, but optimization time and work effort increase significantly
Solution Approach 1:
The system performs preliminary actions by creating a comprehensive experiment design before actual parameter optimization begins. The experimental traveling design creation device pre-defines multiple travel experiments with different parameter combinations, allowing systematic data collection that reduces overall optimization time while maintaining control precision
Solution Approach 2:
The system implements feedback mechanisms where the evaluation value calculation device continuously assesses travel performance based on measurement values from sensors. This feedback is used by the traveling parameter optimization device to iteratively refine parameters, achieving high precision control through data-driven adjustments rather than extensive trial-and-error
2Reliability
If comprehensive experimental traveling is conducted to optimize parameters, then parameter appropriateness is improved, but work effort and time consumption increase
Solution Approach 1:
The optimization process is segmented into distinct functional modules: experiment design creation, traveling execution, measurement value acquisition, evaluation value calculation, and parameter optimization. This segmentation allows parallel processing and systematic management of the optimization workflow, improving productivity while ensuring comprehensive parameter optimization through structured experimentation
Solution Approach 2:
The system performs self-service optimization by automatically conducting experiments, collecting data, evaluating performance, and adjusting parameters without requiring manual intervention for each step. The traveling parameter optimization device autonomously processes measurement values and generates optimized parameters, significantly reducing work effort while maintaining high parameter appropriateness
Data Source
AI summary
An optimization system includes an optimization server and a management server, and the management server controls an AGV. The AGV transports a cargo and notifies the management server of a traveling state during traveling. The management server stores the traveling state of the AGV in a database. The optimization server estimates the traveling parameter to be used for the subsequent experimental traveling and repeats an experiment based on an experiment design for experimental traveling and the number of times of back-and-forth sway and the right-and-left sway width of the AGV when the AGV travels by using the traveling parameter estimated from the default value of the traveling parameter and an experiment result of the experimental traveling of the AGV. After the experiment ends, the optimization server optimizes the traveling parameter for each traveling state of the AGV based on the experiment result.