Longitudinal control strategies for unmanned ground behicles in uneven terrains

• Autor(a) principal: Victor Ricardo Fernandes Miranda
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Institucional da UFMG
• Texto Completo: http://hdl.handle.net/1843/31027
• topic: Autonomous ground vehicle; Trajectory tracking; Robust control; Nonlinear control; Mobile robots; Engenharia elétrica; Veículos autônomos; Controle robusto; Teoria do controle não-linear

This work presents control strategies to track longitudinal trajectories by an unmanned ground vehicle, considering changes in terrain type and conditions. It also features the design and construction of a small, off-road, four-wheeled, Ackermann platform, able to travel in uneven terrains. In several scenarios of mobile robot’s application, as in mining and agriculture tasks, it is necessary to drive on irregular and unstructured terrains. This may directly affect the vehicle performance while executing these tasks at higher speeds. These terrain irregularities, as well as changes in the soil type and changes in the load carried by the robot, are disturbances that affect the robot dynamics, causing difficulties in the point of view of navigation and speed control. In view of these problems, a simple platform was built, which can be used in irregular terrains, to perform semi-autonomous tasks. Thus, to achieve this autonomous tracking, this thesis proposes two control techniques to reduce the effect of ground disturbances and changes in the robot mass during the execution of a task: Robust PID controller, tuned by an LMI approach, and a nonlinear controller, tuned by a Backstepping technique, with integral action. This platform is equipped with low-cost sensors and embedded computers. Simulation and real-life experiments are presented to illustrate the performance of each designed controller.