Mapping and navigation for service robots
Low-cost localization, mapping and navigation for ground robots, drones and manipulators working in industrial plants.
This line of projects develops a low-cost mapping and navigation system that can be mounted on different service robots, including ground vehicles, drones and robot arms, in production plants. It combines sensor fusion for localization, point-cloud mapping with LiDAR SLAM, and planning and control modules (graph and sampling-based planners, vector fields and predictive control), with implementations aimed at embedded hardware.
Recent results include an embeddable mapping and navigation unit, LiDAR–robot and LiDAR–camera calibration for multi-sensor fusion, instrumented wheels with nanostructured sensors for weight sensing and terrain classification, and EKF-LOAM, which fuses LiDAR SLAM with wheel odometry and inertial data for spaces with few geometric features.
Funding: CNPq (Technological Development 2020 and 2023, Universal 2021) and FAPEMIG (Universal 2022). Coordinated by Gustavo Freitas.
Publications
- JINTTerrain Classification Based on RNNs for Ground Robots Using a Custom Instrumented Wheel with Nano-Structured SensorsJournal of Intelligent & Robotic Systems 2026
- T-ASEEKF-LOAM: An Adaptive Fusion of LiDAR SLAM With Wheel Odometry and Inertial Data for Confined Spaces With Few Geometric FeaturesIEEE Transactions on Automation Science and Engineering 2022
- CROSEnhancing the Autonomous Mapping and Navigation Unit for Service RobotsIn Brazilian Conference on Robotics (CROS 2025)
- LARSExtrinsic LiDAR-Robot And LiDAR-Camera Calibration For Multi-sensor FusionIn Latin American Robotics Symposium (LARS) 2024
- LARSNanostructured weight sensor for a service robot wheelIn Latin American Robotics Symposium (LARS) 2024
- SBAIUnidade de Mapeamento e Navegação Autônoma Embarcável em Diferentes Dispositivos Robóticos.In XVI Simpósio Brasileiro de Automação Inteligente (SBAI 2023)