Đang chuẩn bị liên kết để tải về tài liệu:
Mobile Robots Navigation 2008 Part 3

Tham khảo tài liệu 'mobile robots navigation 2008 part 3', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | 68 Mobile Robots Navigation 5.2 Typical Results A typical result of applying the three behaviors during navigation is shown in Figure 8. The robot was put into an example scenario bounded by movable walls with an exit in the opposite corner. The experiment was repeated two times. In the first run the scanner was held in a horizontal position 2D perception comparable to a standard 2D laser range finder. In the second run the scanner was continuously rotated over the aforementioned area of interest in a nodding-like fashion 3D perception . In both experiments a constant translational velocity of 0.3ms 1 was set with no rotational velocity. That is the robot was commanded to move forward. The application of the behaviors successfully moved the robot through the exit and outside the scenario. With 2D perception small test objects lying on the ground were not perceived. The robot crushed into these objects and pushed them through the exit. With 3D perception the objects have been perceived and the robot successfully avoided them. Although it is not explicitly covered in this example it is to note that the robot robustly perceives sudden dynamic changes in the environment due to the fast pitch rate of the scanner while driving. It therefore detects and avoids dynamic obstacles like for instance suddenly opened drawers or people passing by. What can also be seen in the figure is that simply commanding the robot to move forward while enabling the three collision avoidance behaviors leads to an emergent behavior of wandering around. This strategy can amongst other applications be used to perform a random exploration. Fig. 8. Behavior-based collision avoidance in an example scenario by means of 2D perception 2D laser range finder in a fixed horizontal position and continuous 3D perception together with the concept of obstacles maps. The small test objects are successfully avoided by the robot when using 3D perception. A video showing the robot perform in a similar experiment