Control of Wheeled Robots Using GNSS and Inertial Navigation: Control Law Synthesis and Experimental Results

摘要

The paper addresses the problem of automatic steering of the wheeled robot like a car to drive it through desired path stored in the controller memory. More specifically, a target point selected on the body of the robot must follow the target path. The target point must be driven to the target path from the arbitrary position and arbitrary orientation of the robot platform and its motion along the target path must be stabilized. The GNSS receiver and inertial sensors are used to measure the position, velocity, and orientation. The orientation of the body of the robot is necessary to translate position of the receiver antenna to the position of the target point. Orientation angles are also used in the synthesis of the control law.The problem setup described above is not new. An extensive literature is devoted to it (see [l]-[4] and references cited there). The control law synthesis is described in [3,4], but two-sided constraints on the front wheels angle are not taken into account accurately. On the other side, these constraints affect badly the stability of the motion of the robot and degrade quality of the automatic control. The properties of the transient processes happen to be dependent on the initial position and initial orientation of the robot. For some initial data the lateral displacement of the target point from the target trajectory decreases in the exponential manner predicted