Mobile Robotics
Work description
Literature Review: • Review of previous work on localization using 3D sensors; • Analysis of current technologies, identifying their limitations and areas for improvement. Data Acquisition: • Selection of the most suitable 3D sensors (e.g. LiDAR, stereo cameras); • Planning data collection in different environments; • Identification of relevant public datasets. Resource Engineering: • Feature Extraction: Identification of the most relevant features, such as curvature, normals and distances; • Feature Selection: Choose the most important features using techniques such as: o Correlation analysis; o Importance of features (feature importance); o Dimensionality reduction. Algorithm Development and Training: • Defining the model architecture, choosing the algorithm and carrying out tests with different configurations and hyperparameters; • Training: Division of collected data into training, validation and test sets. Experimenting with various optimization techniques such as gradient descent and stochastic descent. Performance Analysis and Refinement: • Assessment of model performance on the validation set, using relevant metrics (e.g., precision, accuracy, recall, F1-score); • Model Refinement: Adjusting the model by modifying hyperparameters, adding or removing features, or exploring different architectures; • Consideration of techniques such as transfer learning or ensemble methods. Compilation of Results and Final Report: • Compilation of results obtained throughout the project; • Writing of the final report, detailing the work carried out and the conclusions drawn
Academic Qualifications
Bachelor's degree in engineering or related field;
Minimum profile required
Enrollment in a Master's degree in Electrical and Computer Engineering, Computer Engineering, or related areas;
Preference factors
Experience in Artificial Vision, Artificial Intelligence, ROS;
Application Period
Since 26 Sep 2024 to 09 Oct 2024
Centre
Robotics in Industry and Intelligent Systems