Research On Improvements And Enhancements Of The Control System For Search And Rescue Robot

Search and rescue robot(SAR Robot for shot) has a great potential usage in disaster rescuing. It can be used as a substitute for the rescue members to detect the dangerous areas after the disasters, which can provide useful information of the environment, and strong support for the planning and scheduling for the rescue operation. SAR Robot should have the ability to work in very special environment. So it must own multiple functions, such as flexible obstacle climbimg ability, numerous environment perceptions, reliable data transmission, and so on. Much work had been done for the SAR robot and it has foudamental functions for rescue work now. It is very necessary to design its localization system and safeguard system for its practical application in the future.There are many types of localization methods. One of them is dead reckoning. To locate the robot's correctly, a localization symtem using optical encode is designed. Not only the circuit is designed, but also tested using the SAR robot. The result of the experiment is showed and analysied in Chapter 2. The result shows that localization error is still existing and can be very big after a long time running of the system. In order to solve this problem, two compensation methods are given. One is compensated using programing, the other is adding a heading sensor to the system.SAR robot should work underground. The environment there is complex and unpredictable. So the robot must have certain ability to protect itself. A safeguard system using tilt sensor(SCA-100T) is designed. As a result, the robot can adjust its movement when it faces the danger of rolling over. At the same time it can still receive the safe moving order from its operator. The reason for choosing tilt sensor other than inertial measurement instruments is also given out.Recently, a new localization method called visual odometry rises many research workers'attention. Visual odometry has a lot of advantages, such as large information, rich features, wide adaptive, and so on. It can not only provide the location massage of the robot, but also can give the pose information. For the purpose of applying visual odometry on SAR robot, Chapter 4 shows out the common frame of visual odometry. The frame includs four steps:feature detection, feature matching, camera calibration and motion estimation. Chapter 4 introduces related theorys of the frame, realizes some methods which are widely used in every step by analysing and programing. Some pictures are given out to show the testing results. Those research work on visual odometry is the basis of integrated use of three techniques discussed in this paper on SAR robot in the future.Finally the summarization of the thesis and the future work are presented.