Autonomous mobility for urban environments
FIRST CALL
APPLICATION ID: ALL8
What we are looking for:
We are seeking innovative research proposals that explore learning-enabled multi-sensor localization as a viable and promising alternative to conventional guidance methods for autonomous vehicles. The selected Postdoc will join a research team between the Centro de Automática y Robótica (CAR) and Instituto de Robótica Industrial (IRI).
The context:
Autonomous driving systems, primarily in urban areas, necessitate GNSS-denied solutions for a successful deployment. The proposed research aims to extend Bayesian filtering techniques by leveraging deep learning methodologies to bolster their performance and robustness. To that end, information from cameras, LiDARs and IMUs will be seamlessly integrated.
The problem to address:
Jorge Villagra, CAR and Juan Andradre, IRI, are joining efforts to address the use of learning-enabled multisensor data-fusion as a viable alternative to conventional localization techniques. Existing methods lack adaptability and accuracy, especially in dynamic urban landscapes with diverse features. The research proposal should address these limitations by developing advanced localization techniques capable of functioning effectively in GNSS-denied urban areas. This problem is of interest to the various companies and organizations that support our joint activities, such as Renault Spain or the Karlsruhe Institute of Technology (Germany).
Objectives:
- To investigate the potential combination of Invariant EKF with deep learning mechanisms to enhance adaptability and accuracy in urban environments lacking GNSS signals.
- To integrate visual-LiDAR-IMU odometry alongside multi-LiDAR and computer vision for robust landmark-based perception in complex urban settings.
Expected Outcomes:
- Improved localization accuracy and reliability in GNSS-denied urban environments, enabling safer and more efficient autonomous driving.
- Enhanced adaptability and performance of autonomous driving systems in challenging urban landscapes, contributing to the advancement and widespread adoption of autonomous vehicle technology
