Tham khảo tài liệu 'multi-robot systems from swarms to intelligent automata - parker et al (eds) part 9', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | Motion Planning with Safe Dynamics 163 minimum of the robot s radius and the distance to the randomly chosen target. An image of the planner running in simulation is shown in Figure 3 and a photograph of a real robot controlled by the planner is shown in Figure 4. To simplify input to the motion control the resulting plan is reduced to a single target point which is the furthest node along the path that can be reached with a straight line that does not hit obstacles. This simple postprocess smooths out local non-optimalities in the generated plan. Figure 3. A robot on the left finds a path to a goal on the right using ERRT. Figure 4. A robot lower left navigating at high speed through a field of static obstacles. 3. Motion Control Once the planner determines a waypoint for the robot to drive to in order to move toward the goal this target state is fed to the motion control layer. 164 Bruce and Veloso Figure 5. Our motion control approach uses trapezoidal velocity profiles. For the 2D case the problem can decomposed into two 1D problems one along the difference between the current state and the target state and the other along its perpendicular. The motion control system is responsible for commanding the robot to reach the target waypoint from its current state while subject to the physical constraints of the robot. The model we will take for our robot is a three or four wheeled omnidirectional robot with bounded acceleration and a maximum velocity. The acceleration is bounded by a constant on two independent axes which models a four-wheeled omnidirectional robot well. In addition deceleration is a separate constant from acceleration since braking can often be done more quickly than increasing speed. The approach taken for motion control is the well known trapezoidal velocity profile. In other words to move along a dimension the velocity is increased at maximum acceleration until the robot reaches its maximum speed and then it decelerates at the maximum allowed .
