Development of a semi-autonomous terrestrial robot for forest management

Nowadays, the frequency of fires is increasing due to climate change. Effective forest management is vital in mitigating the intensity and spread of this natural disaster. Currently, the task of forest cleaning is carried out by tractors equipped with forestry mulchers, which are operated remotely using joysticks. This thesis is part of a project to automate the cleaning process using a fully sensor-equipped electronic box. Specifically, it addresses the crucial aspect of precise line following, which is essential for efficient cleaning operations. The unmanned ground vehicle (UGV) under study consists of two main components. The vehicle part comprises the tractor LV 600 Pro, which is responsible for the vehicle's movement. The second component is the sophisticated electronic box called Sentry, developed by Bold Robotics Lda, which allows the tractor to move autonomously. This thesis aims to develop software for the component mentioned above. Various algorithms will be presented to handle and filter raw sensor data, also employing fusion methods to achieve higher measurement precision. Special attention will be given to navigation aspects, including the presentation of different motion planning strategies. Extensive testing will be conducted in simulated environments using Gazebo to validate the proposed approaches. Subsequently, field tests will be carried out to provide real-world comparisons and assess the effectiveness of the developed solutions. Keywords: ROS, UGV, line following, obstacle avoidance, sensor acquisition, control.