移动机器人节点的路径规划及轨迹优化算法

发布时间：2018-06-03 16:53

本文选题：移动传感网络 + 路径规划　；参考：《湖南科技大学》2017年硕士论文

【摘要】：煤矿监测监控系统是保证现代化煤矿安全生产的重要手段,其中无线传感网络技术为解决井下煤矿安全监控问题提供了有效技术手段和途径。相比传统无线传感器网络,基于移动机器人节点的移动传感器网络更能满足煤矿正常生产与救援条件下的信息获取需求。针对煤矿井下的复杂环境及其安全生产需求,本文将移动机器人作为传感器网络的节点,研究节点定向移动和节点部署的行为规划与控制方法。本文首先介绍了课题的背景及意义、国内外发展现状以及研究的主要内容,并简单说明了论文的章节安排。第二,本文研究了在局部动态环境下,单移动机器人节点的路径规划问题。针对传统势场法中,机器人对移动障碍物避障效率较低,路径中存在局部极小点等缺陷,本文构建了一种基于势流理论的势场模型,势流理论中的概念与路径规划中的概念一一对应。为保证对移动障碍物避障的可靠性,应用茹科夫斯基变换对势场分布及函数进行了改进,引入点涡和虚拟点源的概念以解决局部极小问题并优化轨迹,该部分最后讨论了若干势场加权叠加方法。在本部分的仿真中,在多种避障情景下对比了改进前后的势场法。结果表明,改进势场能够引导机器人对移动障碍物进行灵活避障,在避免局部极小点时较传统方法更为有效。第三,研究提出基于logistic曲线的移动机器人优化轨迹规划方法。在描述了基本轨迹规划问题,介绍并完善logistic曲线的基本性质之后,进一步提出半S型曲线。讨论曲率界值、概念化障碍物避障和不连续通道三种约束,以满足实际情况中机器人所需的轨迹约束条件。在理论和仿真实验两个方面与其他类型曲线进行了对比,并在综合情景下对logistic曲线规划的有效性进行了验证。仿真结果表明,提出方法能够通过灵活控制设计参数满足各种类型约束条件,相较传统类型的轨迹,其设计算法更为简单直观,计算量更小,且方法具有互换性。最后,研究多机器人的改进势场法路径规划问题,对实际作业中需要满足的三个重要条件进行了讨论。为保证机器人之间的基础避碰,使用社群势场定义机器人之间的相互作用力;当人类与机器人在环境中共存时,使用行为势场,兼顾机器人移动效率和对人类避障,以寻求安全与效率之间的平衡;提出防堵势场识别并避免全局拥堵。在仿真中,结合三种势场,对提出的规划方法的有效性进行了验证。在实验中,通过投影方法表示实验结果并进行了对比,进一步说明了提出方法较传统方法效率更高。
[Abstract]:Coal mine monitoring and monitoring system is an important means to ensure the safety of modern coal mine production. Wireless sensor network technology provides an effective technical means and approach to solve the problem of underground coal mine safety monitoring. Compared with the traditional wireless sensor networks, the mobile sensor networks based on mobile robot nodes can meet the information acquisition needs of coal mine under normal production and rescue conditions. In view of the complex environment and the requirement of safe production in underground coal mine, this paper takes the mobile robot as the node of the sensor network, and studies the behavior planning and control method of the node directional movement and node deployment. This paper first introduces the background and significance of the subject, the development situation at home and abroad, as well as the main contents of the study, and briefly describes the chapter arrangement of the paper. Secondly, the path planning problem of single mobile robot nodes in local dynamic environment is studied in this paper. In view of the disadvantages of the traditional potential field method, such as the low efficiency of obstacle avoidance and the existence of local minima in the path, a potential field model based on the potential flow theory is constructed in this paper. The concepts in potential flow theory correspond to those in path planning. In order to ensure the reliability of obstacle avoidance for moving obstacles, the distribution and function of potential field are improved by using Zhukovsky transform. The concepts of point vortex and virtual point source are introduced to solve the local minimum problem and optimize the trajectory. In the end, some potential field weighted superposition methods are discussed. In the simulation of this part, the potential field method before and after the improvement is compared in many kinds of obstacle avoidance scenarios. The results show that the improved potential field can guide the robot to avoid obstacles flexibly and is more effective than the traditional method in avoiding local minima. Thirdly, an optimal trajectory planning method for mobile robot based on logistic curve is proposed. After describing the basic trajectory planning problem and introducing and perfecting the basic properties of the logistic curve, the semi-S-shaped curve is further proposed. This paper discusses three kinds of constraints, such as curvature bound, conceptualized obstacle avoidance and discontinuous channel, in order to satisfy the trajectory constraint conditions required by the robot in practice. The theoretical and simulation results are compared with other types of curves, and the effectiveness of logistic curve planning is verified in a comprehensive scenario. The simulation results show that the proposed method can satisfy all kinds of constraint conditions by flexibly controlling the design parameters. Compared with the traditional type of trajectory, the proposed method is simpler and more intuitive, less computational and interchangeability. Finally, the path planning problem of the improved potential field method for multi-robots is studied, and three important conditions to be satisfied in practical operation are discussed. In order to guarantee the basic collision avoidance between robots, the community potential field is used to define the interaction force between robots, and when human and robot coexist in the environment, the behavior potential field is used to take into account the robot's mobility efficiency and obstacle avoidance. In order to find the balance between safety and efficiency, the anti-blocking potential field is identified and global congestion is avoided. In the simulation, the effectiveness of the proposed programming method is verified by combining the three potential fields. In the experiment, the projection method is used to express the experimental results and the comparison is made, which further shows that the proposed method is more efficient than the traditional method.
【学位授予单位】：湖南科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】