This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: CS221 Lecture notes Robotics and motion planning In the previous lecture, we saw two examples of intelligent systems–Flakey the robot, that could move around an office environment and talk to people, and Stanley, Stanford’s autonomous car. An intelligent systems can be a robot and have a physical presence, or it can also entirely live within your desktop computer. Today, lets talk about what it takes to physically control a robot. What is a robot? Informally, it is a physical system that interacts envi- ronment using physical sensors and effectors. Some examples of robotic sensors include video cameras; sonar (which measures distances to obstacles by measuring how long it takes sound to bounce off the obstacle and return to the robot); laser range scanners (which works similarly to sonar, except it bounces light rather than sound off the obstacles); microphones; odometers (which measure distance traveled); GPS; accelerometers; and many more. These sensors give the robot information about the state of the environment around it, as well as information about the robot’s location and orientation within that environment. Another important type of sensor are “proprioceptive sensors,” which tells the robot about the position or changes in position of its own joints. For most robots, their effectors can be divided into two categories based on their function: • Locomotors , such as wheels or legs, to allow the robot to move itself around. Wheels are popular since they are easier to control, but there many other possibilities, such as legs. • Manipulators , such as a robot arm and hand, which allow the robot to interact with the world and affect the world around it. Given a robot, how can we get it to drive from one place to another, without hitting obstacles? Alternatively, given a robot arm, how can we 1 2 generate a smooth motion for it to reach out and, say, pick up an object? At its basic level, a robot consists of a set of motors, and we have to decide (say) what sequence of joint angles to command each motor to go to. We would like to find a sequence of joint angles that will cause the robot to follow some “path” from its initial position to some goal position. To develop algorithms to accomplish this, we will need to introduce the concept of a configuration space. 1 Configuration space First consider a robot in a 2D plane. How would we describe its position? That depends on whether it can rotate, or just translate without rotating in the plane. In the latter case, we can describe the robot’s position with a pair of real numbers, such as its Cartesian coordinates ( x, y ). In the former case, we would need three real-valued parameters ( x, y, θ ), with θ giving the robot’s orientation....
View Full Document
This note was uploaded on 11/30/2009 for the course CS 221 taught by Professor Koller,ng during the Winter '09 term at Stanford.
- Winter '09
- Artificial Intelligence