SIA OpenIR  > 水下机器人研究室
Alternative TitleResearch on Shared Control Method of Autonomous and Remotely Operated Underwater Vehicle
Thesis Advisor田宇
Keyword自主遥控水下机器人 共享控制 基于行为控制 多目标优化 避碰
Degree Discipline控制工程
Degree Name硕士
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract本文的研究依托于国家重点研发计划项目课题“全海深ARV系统基础理论问题研究(2016YFC0300801)”,针对在环境探索和目标观察任务中如何融合ARV自主控制和操作人员遥控的问题,基于共享控制理论(Shared Control)开展研究,在广泛调研共享控制文献的基础上研究和设计了ARV的共享控制方法,并基于本文研发的仿真环境验证了所提出方法的有效性,为ARV共享控制的研究和应用提供了必要的理论基础和技术储备。 本文主要的研究内容和结果如下:(1)针对ARV如何在环境探索任务中和目标观察任务中实现共享控制的问题,采用基于行为控制的思想,提出一种基于行为的ARV共享控制方法。在环境探索任务中,设计了遥控行为和自主避障行为以使ARV在执行操作人员操作意图的同时保障安全;在目标观察任务中,基于提出的路径跟踪方法和围绕目标观察的超椭圆路径,设计了人机协同路径跟踪控制行为以使ARV围绕目标观察;最后通过本文设计的基于优先级的行为组织和融合方法协调各个行为的输出,实现了ARV基于设计的行为以“人主机辅”模式执行环境探索任务继而以“机主人辅”模式执行目标观察任务的有效共享控制。(2)为进一步提高ARV共享控制在环境探索任务中的效果,面向ARV的环境探索任务,提出了一种基于多目标优化的共享控制方法。根据服从操作人员控制意图、提升ARV安全性、降低操作人员操作复杂性并优化ARV运动路径的要求,将ARV艏向角的控制命令作为决策变量,分别设计了服从度、自主度和稳定度三个目标函数;根据栅格地图表示的局部环境中障碍物的分布信息设计了艏向角控制命令的安全性评估函数,并由安全性评估函数确定约束条件;根据目标函数和约束条件将ARV艏向角的共享控制转化为多目标优化问题,并使用最小最大法求解以得到最优的共享控制命令。(3)针对ARV共享控制方法的研究需求,采用模块化的设计思想和面向对象的开发方式,研发了ARV共享控制仿真研究环境。仿真环境由视景显示、控制算法、实物仿真三个模块组成,具有动画与数据演示、障碍物环境仿真、结果输出与分析等功能,提供了算法接入的通用接口,便于对本文提出的共享控制方法进行研究、验证与分析。基于ARV共享控制仿真研究环境,本文对两种共享控制方法进行了验证与分析,证明了所提出方法的有效性。仿真结果表明,基于行为的ARV共享控制有效的提高了ARV在环境探索任务和目标观察任务中的安全性和任务表现,减轻了操作人员的工作负担;基于多目标优化的共享控制方法在环境探索任务中进一步提高了ARV的任务表现并减轻了操作人员的工作负担。
Other AbstractSupported by the program of the National Key R&D Program of China “Research on the basic theory of the whole sea deep ARV system (Grant No. 2016YFC0300801)”, the topic of this thesis is how to integrate the control capabilities of the autonomous control system and the human operators suitably in the environmental exploration and object observation missions. Two shared control methods for ARV based on the shared control theory are proposed in this thesis. This methods are implemented and evaluated in a developed computer simulation environment, and the results demonstrate their effectiveness. The research of this thesis provides the necessary theoretical basis and technical reserve for the research and application of shared control in ARV. The main researches and results obtained in this thesis are as follows: (1) Based on the behavior-based approach, a shared control method for ARV is proposed in this thesis for human-vehicle collaborative environmental exploration and object observation missions. In the environmental exploration task, the remote control behavior and the autonomous object-avoidance behavior are designed to perform the operator's intention and improve the ARV’s safety. In the object observation mission, the human-vehicle collaborative path-following control behavior is designed based on the proposed superellipse path around the target and path tracking method to observe the object. And to coordinate the designed behaviors to implement the two missions, a control architecture with priority-based behavior fusion mechanism is developed. The proposed method effectively implements the human-centered and vehicle-centered shared control in simulated human-vehicle collaborative environmental exploration and object observation missions. (2) A multi-objective optimization based shared control method is proposed in this thesis to control an ARV’s heading angle to improve the performance of ARV shared control in the environment exploration task. In this method, three objective functions including Obedience, Autonomy and Stability are designed, to meet the presented sub-goals of optimization in the task, including improving the obedience of the ARV to human-control’s intention, the ARV’s safety in environments with obstacles, and the easiness of human control and smoothness of ARV’s motion path, respectively. These objective functions take the control command of ARV’s heading angle to be optimized as variable and evaluate the respective performance. This method also incorporates a designed Security function, which determines the feasible control commands of ARV’s heading angles along which the ARV moves will not collide with obstacles represented in an occupancy grid map to obtain the constraint condition. The objective functions and the constraint condition together defines the multi-objective optimization problem and the minimax method is employed to solve the problem to obtain the optimal control command of ARV’s heading angle. (3) Based on the modular design ideas and object-oriented development methods, the computer simulation environment for ARV shared control research is designed and developed. The simulation environment consists of three modules: visual display module, control algorithm module and practicality simulation module. It provides a general interface for algorithm access and has functions such as animation and data display, obstacle environment simulation, result output and analysis. These functions are quite convenient for this thesis to study, verify and analyze the ARV shared control method. The proposed shared control methods are implemented and evaluated in the developed computer simulation environment, and the results demonstrate their effectiveness. The simulation results show that the behavior-based ARV shared control method improves the security and task performance of ARV and reduces the workload of human operators effectively in environmental exploration and object observation missions. The multi-objective optimization based shared control method further improves the mission performance of ARV and reduces the workload of the human operators in environmental exploration tasks.
Contribution Rank1
Document Type学位论文
Recommended Citation
GB/T 7714
王兴华. 自主遥控水下机器人共享控制方法研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2019.
Files in This Item:
File Name/Size DocType Version Access License
自主遥控水下机器人共享控制方法研究.pd(3584KB)学位论文 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[王兴华]'s Articles
Baidu academic
Similar articles in Baidu academic
[王兴华]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[王兴华]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.