As noted previously, the second path of robot development was concerned with mobile robots, devices capable of moving in their environment by means of legs or wheels. (More recent mobile robots are also capable of mobility in the air and under water.) It shows a commonly used research robot, the Pioneer 3-AT, made by ActivMedia Robotics, in Amherst, New Hampshire. Note that this robot has four wheels, a number of sonar proximity sensors along the front surface, and a complex of communication equipment on its top surface. Some Pioneers are equipped with laser range sensors, global positioning satellite (GPS) receivers, and other sensing devices.
Control of the position, orientation and, velocity of the Pioneer (and other wheeled robots) is obtained through electric motors driving the wheels. Generally,the control systems for the driving motors make use of feedback; careful design is needed to avoid the possibility of instability. Di¤erential control of a pair of wheels may allow the robot to turn in place, or the robot may have a minimum turning radius (like an automobile). Further control problems will arise when the robot is attempting to navigate from a starting to a goal position. The simplest way to control a trajectory is by counting wheel revolutions. Unfortunately, this method, known as odometry, can lead to very large trajectory errors, since the robot wheels may slip. Hence, more complex methods of trajectory control are needed (possibly involving Kalman filters or other statistical methods), or additional sensors (such as vision) may be used to identify landmarks for navigation.
