Nombre: LUIZ MIGUEL MONTEIRO NASCIMENTO PESSOTTI TAVARES
Fecha de publicación: 29/10/2024
Supervisor:
Nombre |
Papel |
|---|---|
| DANIEL KHEDE DOURADO VILLA | Asesor |
Junta de examinadores:
| Nombre |
Papel |
|---|---|
| DANIEL KHEDE DOURADO VILLA | Presidente |
| DIEGO NUNES BERTOLANI | Examinador Externo |
| ELIETE MARIA DE OLIVEIRA CALDEIRA | Examinador Interno |
| FELIPE NASCIMENTO MARTINS | Examinador Externo |
| MARIO SARCINELLI FILHO | Coorientador |
Sumario: This dissertation presents the development of a motion planner for terrestrial mobile robots in dynamic environments, using a 360 LiDAR. The proposed system integrates the RRT (Rapidly-Exploring Random Trees) algorithm with modified artificial potential fields to avoid collisions with static and dynamic obstacles. To address the limitations of RRT in dynamic environments, a local RRT is introduced, which replans trajectories in real-time without interrupting the robot’s navigation. Additionally, a new obstacle avoidance algorithm based on artificial potential fields is implemented. The combination of these two techniques overcomes classic limitations such as local minima and difficulty in navigating through narrow passages. Experiments were conducted using the LIMO mobile robot in differential and omnidirectional configurations, in scenarios with both static and dynamic obstacles. The results demonstrate that the proposed system enables the robot to navigate efficiently in dynamic environments, quickly adapting to changes in the surroundings. However, some collisions were observed during testing, particularly in the differential configuration, due to limitations such as the LiDAR’s blind zone and distortions in the local map during abrupt rotations. These collisions highlight the need for improvements in the perception system to enhance navigation safety and robustness. The simulations further reinforce the effectiveness of the method compared to traditional algorithms, demon-strating the potential of this approach for mobile robotics applications. Despite the observed collisions, the proposed method proved to be promising, pointing to future directions focused on adding sensors and improving the mapping algorithm to reduce incidents and increase system reliability.
