A Comparison of Different Approaches for Formation Control of Nonholonomic Mobile Robots regarding Object Transport

Abstract

Controlling the formation of several mobile robots allows for the connection of these robots to a large virtual unit. This enables a group of mobile robots to carry out tasks that a single robot could not perform. For this purpose, the use of nonholonomic mobile robots is especially useful, as they often have a higher payload and are suitable for a wider range of terrains. However, most research in the area of formation control is focused on holonomic robots, since their superior mobility allows for better control and allows for the research on more sophisticated control techniques. The remaining articles explicitly dealing with nonholonomic robots often do cover common controllers, but do not include realistic simulations or comparison of different controls on the same trajectory. Therefore, in this paper, we present a comparative analysis of two frequently used control approaches. We compare the behavior of a l-?-controller and a Cartesian reference-based controller with different types of reference value generation and pose determination. The evaluation of all resulting control schemes is based on the task of collaborative object transport. To do so, we selected performance criteria geared towards applicability in real processes. In addition, we used an error model, which takes into account the noise and accuracy of all sensors (IMU and encoder) as well as the drift in odometry caused by the slip of the robot's wheels. The comparison includes a series of simulations using two trajectories with a changing number of robots and different formation geometries. In the simulations we got slightly better results for the Cartesian control law.