Many systems exist today for robot global position estimation. Perhaps the most familiar example is the Global Positioning System (GPS), a network of 24 satellites that can give an absolute position estimate accurate to within several meters. For smaller scale position estimation, high-end solutions such as robotic vision and laser range-finding can provide millimeter accuracy distance measurements, which can then be matched with map data to convert local distance measurements to global position. As an alternative to these more expensive systems, ultrasonic and infrared distance sensors can offer similar performance with modest compromises in speed and accuracy. Of these two, infrared distance sensors often have slightly narrower beam width and faster response.
Figure 1.2(a) shows the Sharp GP2Y0A21YK0F, a popular medium range (10-80 cm), infrared (IR) distance sensor. The Sharp sensor uses triangulation to calculate distance by measuring the angle of incidence of a transmitted IR beam reflected from a distant surface onto a receiving position sensitive device (PSD). Because the angle is a nonlinear function of the distance to the surface, the Sharp sensor has the nonlinear calibration curve shown in Figure 1.2(b). Given discrete calibration data, we can linearly interpolate voltage readings taken from the sensor to derive distance measurements.