【作者】 尹继超；

【导师】 袁赣南；

【作者基本信息】 哈尔滨工程大学 ， 导航、制导与控制， 2010， 硕士

【摘要】 各式各样的机器人在现代工厂、采矿、国防、海洋开发及外太空探索中发挥着越来越重要的作用。人工智能发展到今天,在某些领域已有长足的进步。但离真正实现“人一样的智能”,还有很长的路要走。机器人遥操作技术是机器人研究领域一项非常重要的课题,在机器人智能化程度不高的现实情况下,机器人很难离开人的控制,因此遥操作技术几乎是现在的机器人必须具备的一项技术。机器人遥操作技术在现在和将来都有重要的研究价值和广泛的应用前景。本文通过深入的分析和详细的研究,对遥操作机器人技术进行了如下研究:首先,介绍了造成控制系统不稳定的因素——通信延迟。本文对时延系统进行了详细研究,并分析了网络时延问题的处理机制。最后,设计了在实验室环境下进行通信延迟模拟的具体方式,以便于在实验室环境下对机器人遥操作中的时延问题进行仿真研究提供依据。然后,介绍并总结了人工智能的发展情况。在机器人朝着自动化、智能化发展的大环境下,本文提出了智能机器人系统的的实现方法。最后,对预测控制中的关键技术“DR算法”进行了详细研究。并在实验室环境下进行了软件仿真试验。DR算法使用的目的在于:一方面,在保证一定的仿真精度的前提下,降低仿真实体发送状态更新信息的频率,降低网络中的通信量;另一方克服网络中固有的传输延迟,对网络延迟进行补偿。实验证明本文提出的基于控制对象行为分析的改进DR算法,在降低网络通信负担,提高控制系统控制精度方面效果明显。更多还原

【Abstract】 A wide range of robots is playing an increasingly important role in modern factories, mining, defense, marine development and exploration of outer space. Artificial Intelligence developed to today have made great progress in some areas. But away from the truly "human intelligence", there is still a long way to go. Teleoperation robot technology is a very important field of robot. In the reality of not highly intelligent robots, it is difficult to leave the robot’s control, so teleoperation technology is now almost a must be with technical for a robot.Teleoperation technology in the present and future has important research value and broad application prospects.In this paper, in depth analysising and detailed studying, we have the following studies on teleoperation robot technology:First, we introduce the resulting control system instability - communication delays. In this paper, we make a detailed study of time-delay system, and analyzes the network’s handling of time-delay mechanism. Finally, we design ways of simulated delay in a laboratory environment in order to facilitate for the robot teleoperation in time-delay problems providing a basis for simulation in a laboratory environment.Then,we introduce and summarize the development of artificial intelligence. In the environment of the robot developing toward automation and intelligent, this paper put forward the design of intelligent robot system.Finally, we made a detailed study of a key technology " Dead Reckoning " for the predictive control,and made a software simulation in a laboratory environment. The purpose of the use of Dead Reckoning is: on the one hand, under the premise of the simulation accuracy reducing the frequency of simulation entity sending status updating information and reducing network traffic; the other side to overcome the inherent network delay and to compensate network delay. In this paper, we made improvements on Dead Reckoning based on object behavior analysis. Experiment proved that the method has evident result in reducing the network traffic burden and improving control system’s precision.更多还原