Mobile Robot Position Determination

摘要

One of the most important reasons for the popularity of mobile robots in industrial manufacturing is their capability to move and operate freely. In order for the robots to perform to the expectations in manufacturing, their position and orientation must be determined accurately. In addition, there is a strong tendency to grant more autonomy to robots when they operate in hazardous or unknown environments which also requires accurate position determination. Mobile robots are usually divided into two categories of legged and wheeled robots. In this chapter, we focus on wheeled mobile robots. Techniques used for position determination of wheeled mobile robots (or simply, mobile robots) are classified into two main groups: relative positioning and absolute positioning (Borenstein, 1996, 1997). In relative positioning, robot’s position and orientation will be determined using relative sensors such as encoders attached to the wheels or navigation systems integrated with the robots. Absolute positioning techniques are referred to the methods utilizing a reference for position determination. The Global Positioning Systems, magnetic compass, active beacons are examples of absolute positioning systems. Calculating position from wheel rotations using the encoders attached to the robot’s wheels is called Odometry. Although odometry is the first and most fundamental approach for position determination, due to inherent errors, it is not an accurate method. As a solution to this problem, usually odometry errors are modeled using two different methods of benchmarks and multiple sensors. In this chapter, we will discuss odometry and two different methods to model and estimate odometry errors. At the end, an example for mobile robot position determination using multiple sensors will be presented.