【作者】 杜春红；

【导师】 岳宏；

【作者基本信息】 河北工业大学 ， 机械电子工程， 2004， 硕士

【摘要】 随着机器人研究朝智能化方向的发展，机器人控制方法的研究已经成了人们关注的热点。其中又以有别与基于模型的传统控制方法的研究为主，加之视觉、超声、红外等传感的信息，机器人已经越来越具有人类的能力。 本课题来源于中国科学院沈阳自动化研究所机器人学重点实验室资助下的专项研究——“基于视觉感知机器人行为控制技术研究”(代号：RL200202)。本文完成了其部分内容的研究，包括：机器人基于行为控制方法和策略的探讨，移动机器人视觉技术的分析研究，主动视觉在机器人行为控制技术中的应用等。本文的主要工作和创新点包括： 在机器人行为控制技术分析的基础上，设计了RIRA-ROBOT基于行为机器人的模块、分层式行为体系结构。通过基于模糊逻辑行为控制和基于优先级的多行为决策相结合的控制策略，设计了机器人行为协调器，增强了多行为竞争和并发行为无碰撞的能力。 进行了面向对象的视觉系统研究。基于主动视觉，设计了一个具有上下俯仰、水平转动的视觉装置，和能够实时调整光圈、焦距等成像参数的摄像头控制系统。建立了一个能够进行实时图像采集、图像处理，目标匹配的图像处理系统。 基于颜色信息，利用HSV颜色模型和数学形态学进行了图像分割和增强。该方法所利用的信息量大，不受外界光照等条件变化的影响，抗噪声能力强，简化了图像特征提取，保证了图像处理质量。利用不变矩算法的前四个不变矩作为物体特征进行了匹配，在不降低精确度的情况下，提高了匹配速度(每幅图像处理时间≤0.7秒)。 成功的将基于模糊逻辑行为控制和基于优先级的多行为决策相结合的控制策略，以及图像处理、目标识别等方法应用在RIRA-ROBOT基于行为机器人上。更多还原

【Abstract】 Following the developing of robot to intelligentized direction, the research on control method of robot has been become the focus to research. In these research, the primary method is the method that different to the traditional control method. Add the informatio n of visual, ultrasonic and infrared sensors to robot control, robots has had the ability si milar to people increasingly.This paper from the research that supported by the prefect of the Shenyang Institute ofAutomation (SIA) of the Chinese Academy of Science (CAS)--"Research on Robot behaviorcontrol based vision perception " (No. RL200202). This paper has accomplished part work of it, include: the research on the behavior based control method, the visual technique of robot and the application of active vision on behavior based robot .The main work and innovative ideas include:On the base of technique analyze of behavior based robot, a behavior system structure of RIRA-ROBOT behavior-based robot and a behavior coordinator by combine the fuzzy logic control with multiple objectives propriety- based decision have designed, the ability of competence in multiple behavior and no collision within the intercurrent behavior have enhanced.A visual device with the function of pitching and tilt and a control system to camera with the function to adjust the aperture and focus in time have designed as well as a processor of image to capture, process and matching have been also designed.Utilized the HSV color model and mathematical morphology segmented and enhanced the image behavior color infor. This method has the vast infor. and robustly confront with the disturbance of external environment ,eg. illumination etc. It has simplify the abstraction of image character and assured the quality of image processing. Used the head four moments of the moment invariant arithmetic to do image matching, didn’t debased the precision but enhanced the speed of matching, about 0.7s of one image.This method that combined the fuzzy logic control with multiple objectives propriety- based decision and the method of image processing, object identify to RIRA-ROBOT based-behavior robot have been successfully applied.更多还原