郑婷婷, 王东, 刘伟静,等.面向智能AGV调度系统的软件自适应机制研究[J].河南理工大学学报（自然科学版）,2025,44(4):74-82.

ZHENG T T, WANG D, LIU W J, et al. Study on software self-adaptive mechanism for intelligent AGV scheduling system [J]. Journal of Henan Polytechnic University (Natural Science) , 2025, 44(4): 74-82.

面向智能AGV调度系统的软件自适应机制研究

郑婷婷1, 王东1,2, 刘伟静1,2, 徐文丽1, 党兴1,2, 王俊德1

1.天津航天机电设备研究所，天津  300458；2.天津市宇航智能装备技术重点实验室，天津  300458

摘要: 目的 针对传统的AGV调度系统在实际应用中难以适应用户需求和环境的动态变化，存在需求响应慢、调度策略僵化等问题， 方法 提出一种集成“用户+软件+硬件”的多层级设计结构的智能AGV调度系统软件自适应机制，从“数据感知、设计决策、执行反馈”的自适应理论出发，将任务调度、路径规划、交通管制、电量管理以及通信管理等多个环节纳入软件自适应体系综合考虑，从全局角度提高系统的响应速度和调度效率，以实现对调度AGV任务需求和环境变化的实时感知和动态调整。首先，采用快速导入技术捕获用户需求，并基于规则映射转化为系统可识别的原始数据；其次，通过与预设调度策略池匹配，识别关键影响因子相关参数，从而选择当前场景下最合适的工作模式以执行AGV任务。此外，该自适应机制还可根据执行结果的实时反馈信息持续优化调度策略池，不断适应新的需求与环境的变化。 结果 实验结果表明，用户需求的收集和应用速度明显提高，响应时间不大于300 ms，且使用自适应机制设计框架后，AGV整体作业效率提升了36%。 结论 该软件自适应机制能显著提升AGV的响应速度、调度效率和数据准确性，有效适应特定工作场景下的复杂工作需求，提高系统灵活性和适应性。 

关键词:自适应机制;智能AGV;调度系统;环境感知;动态调整

doi: 10.16186/j.cnki.1673-9787.2024070019

基金项目:天津市科技计划项目（22JCYBJC01740）；天津市自然科学基金资助项目（S23DYYB9012）；中央军委装备发展部慧眼行动计划项目（5056D582）

收稿日期:2024/07/03

修回日期:2024/10/01

出版日期:2025/06/19

Study on software self-adaptive mechanism for intelligent AGV scheduling system

Zheng Tingting1, Wang Dong1,2, Liu Weijing1,2, Xu Wenli1, Dang Xing1,2, Wang Junde1

1. Tianjin Institute of Aerospace Mechanical and Electrical Equipment， Tianjin  300458， China； 2. Tianjin Key Laboratory of Aerospace Intelligent Equipment Technology， Tianjin  300458， China

Abstract: Objectives Traditional AGV scheduling systems were found to struggle with adapting to dynamic changes in user needs and environments in practical applications， exhibiting issues such as slow demand response and rigid scheduling strategies. Methods Integrating a multi-level design structure of “user + software + hardware”， a software adaptive mechanism for intelligent AGV scheduling systems was proposed，. Based on the adaptive theory of “data perception， design decision-making， and execution feedback”， multiple aspects， including task scheduling， path planning， traffic control， battery management， and communication management， were comprehensively considered within the software adaptive system. The system’s response speed and scheduling efficiency were improved from a global perspective to enable real-time perception and dynamic adjustment to changes in AGV task demands and environments. Firstly， rapid import technology was used to capture user needs， which were then converted into recognizable raw data for the system based on rule mapping. Secondly， by matching with a preset scheduling strategy pool， relevant parameters of key influencing factors were identified to select the most suitable working mode for executing AGV tasks in the current scenario. Additionally， the adaptive mechanism continuously optimized the scheduling strategy pool based on real-time feedback from execution results， adapting to new demands and environmental changes. Results The experimental results indicated that the speed of collecting and applying user needs was significantly improved. The response time was no longer than 300 ms. After using the adaptive mechanism design framework， the overall operational efficiency of AGVs was increased by 36%. Conclusions The software adaptive mechanism significantly enhanced the response speed， scheduling efficiency， and data accuracy of AGVs. It effectively adapted to complex work demands in specific working scenarios. The system flexibility and adaptability were improved. 

Key words: adaptive mechanism; intelligent AGV; dispatching system; environmental perception; dynamic tuning