Dynamic motion control for compliant framed wheeled modular mobile robots

In this dissertation, a distributed robust controller, a motion control and sensing architecture, and a terrain feature localization are studied to solve general navigation problems for Compliant Framed wheeled Modular Mobile Robots (CFMMR). This type of wheeled mobile robot uses rigid axles coupled by compliant frame modules to provide both full suspension and enhanced steering capability without additional hardware.; First, a distributed nonlinear damping controller is developed for single-axle unicycle type robots. The controller is then extended to multiple-axle CFMMR configurations and is robust to disturbances created by modeling errors, especially highly nonlinear frame forces caused by axle interaction. In particular, the controller considers time-varying reference velocities and allows the robot to perform posture regulation, path following, or general trajectory tracking.; In order to achieve improved motion control of CFMMR, a distributed cooperative motion control and sensing architecture is developed by combining a kinematic controller for motion coordination and providing reference commands, dynamic motion controllers to follow these commands and reject disturbances, and a sensor fusion system to provide accurate posture estimates. Requirements for each subsystem and their respective interconnections are defined to optimize system performance.; Simulations and experiments on a two-axle CFMMR verify robust dynamic motion control of path following and improved motion control of posture regulation under the developed control and sensing architecture.; The largest source of error in the above work is due to localization. Thus, a novel terrain feature localization technique is proposed to allow the robot to identify its location relative to measurable terrain characteristics. A terrain inclination map is extracted from a given topographical map along a specific path. An EKF is used to fuse the measurement data with the robot motion based on the extracted terrain map. Simulations validate the proposed terrain localization technique.; In this dissertation, three major aspects of robot motion control are studied. The control architecture provides a frame work for implementing motion control. The robust controller is a key part of this architecture and improves accuracy while the localization algorithm improves precision. While the target platform is CFMMR, these strategies represent fundamental advancements in robot motion control.