项目名称: 基于人类驾驶行为的移动机器人无标定优化视觉伺服方法研究
项目编号: No.61203333
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 自动化学科
项目作者: 张雪波
作者单位: 南开大学
项目金额: 25万元
中文摘要: 移动机器人视觉伺服是指利用实时视觉反馈,控制机器人跟踪指定轨迹或到达期望位姿,在自动泊车/充电、物品搬运等方面具有广泛的应用前景。由于非完整约束条件的限制,以及单目视觉信号的各种特性,设计高效、鲁棒的移动机器人视觉伺服策略,在理论方法上也是一个极具挑战性的问题。本项目旨在提出一种基于人类驾驶行为的移动机器人视觉伺服策略,从而提高其运行效率,增强其对摄像机参数未知等因素的鲁棒性,并打破位姿估计算法对应用场景的限制。具体而言,本项目将针对如下内容进行研究:具有人类驾驶特征的运动轨迹实时提取与分析;基于云台与机器人主动控制器的摄像机参数最优辨识;能够打破场景限制的位姿估计算法;摄像机参数存在辨识误差时的自适应/鲁棒视觉伺服控制器;基于自主交互机制的同时摄相机参数辨识与视觉伺服。力争通过本项目的研究,解决一类不确定非完整系统的视觉伺服问题,并促进其在家庭服务、工业生产等场合中的实际应用。
中文关键词: 移动机器人;视觉伺服;位姿估计;运动规划;
英文摘要: Visual servoing of mobile robots uses real-time feedback of visual information to achieve trajectory tracking or stabilization tasks, and thus it has promising applications in several areas such as automatic parking/charging, transportation of objects, and so on. Due to the well-known nonholonomic constraint and the uncertain characteristics of monocular visual signals, it is still a very challenging theoretical problem to design a highly efficient and robust visual servoing strategy. This project aims to propose an optimized uncalibrated visual servoing scheme with quasi-human driving characteristics, which can be applied in both planar and non-planar scenes with improved efficiency and strong robustness to unknown camera parameters. Specifically, this project will study the following respects: extraction and analysis of trajectories with quasi-human driving characteristics; optimal identification of camera parameters based on active controller design for the pan-tilt unit and the mobile robot; the pose estimation technique which is not restricted to a specific kind of scene; adaptive/robust visual servoing controller design with inaccurate camera parameters; simultaneous camera parameter identification and visual servoing based on an autonomous interaction mechanism. According to the research for this project,
英文关键词: Mobile Robots;Visual Servoing;Pose Estimation;Motion Planning;