Descrição
Scania is now undergoing a transformation from being a supplier of trucks, buses and engines to a supplier of complete and sustainable transport solutions.
Do you want to shape the forefront of self-driving technologies? At Autonomous Transport Solutions (ATS) Research, Scania R&D, we pursue top-quality research and development of future cutting-edge ATS concepts. We operate in agile and self-steered teams that work in close cooperation with TRATON, leading technology suppliers, and academic institutions, with the ambition to detect and evaluate upcoming technologies. Our culture is built based upon delivering added customer value through research and practical experiments that iteratively lead to concepts for industrialization.
Your responsibilities
At Autonomous Transport Solutions (ATS) Research, Scania R&D, we pursue top-quality research and development of future cutting-edge ATS concepts. We operate in agile and self-steered teams that work in close cooperation with academic institutions, with the ambition to detect and evaluate upcoming technologies. Our culture is built based upon delivering added customer value through research and practical experiments that iteratively lead to concepts for industrialization.
Background
Autonomous driving is a game-changer in the transportation industry, with huge potential to improve road safety. Scania plans to focus on novel technical solutions that boost the safety, robustness, and scalability of autonomous vehicles in order to thrive in this dynamic environment. A critical requirement for safe autonomous driving is Situational Awareness. This requires access to complete and highly accurate information regarding the surrounding environment. HD maps provide large-scale context information without being degraded by the real-time perception constraints. However, this dependency necessitates continuous validation and update of HD maps, since they may otherwise represent wrong information due to road changes or inaccurate map creation.
At the current time, there is no established theoretical and technological framework that enables detecting and updating the changes of HD map features automatically.
Project Description
You will be developing novel theoretical concepts and evaluating these in practice on an experimental testbed. The goal of this project is the automatic and offline change detection and map update of road features in the HD map. The changes of the road features are identified using different sensor pipelines such as lidar and camera. The information received from various feature detectors and local change detection algorithms, and a-priori information about feature distribution are all fused together to take a decision on valid changes in the HD map. Special geometric structures, such as parallel lanes, that are commonly found on city streets and highways have innate implicit geometric properties. The suggested map changes should also provide such geometric consistency.
To achieve these objectives, the industrial Ph.D. student will pursue a multi-disciplinary research in the areas of state estimation, optimization, machine learning, artificial intelligence and probabilistic modeling.
This research project will be conducted in an agile multi-cultural research team within R&D. The team is responsible for concept evaluation and software development of maps and localization for autonomous driving.
The results of this research project will be implemented in Scania’s development environment for autonomous driving and they will be experimentally evaluated in operating prototype vehicles (trucks and buses) equipped with sensors and a platform set up for autonomous driving.
Your profile
- Knowledge of modern C++ and developing within a Linux-based development environment
- Fundamentals in non-linear optimization, state estimation, probabilistic modeling
Extra meritorious if you have:
- Knowledge and experience within software languages such as Python and/or MATLAB/Simulink along with a keen interest in programming and strong programming skills
- A proven interest in interdisciplinary research and record of initiative
- Experience working with lidar, camera and associated sensor data (Point cloud registration, Sensor fusion, SLAM, Visual odometry)
This research will be conducted in collaboration with the Robotics, Perception and Learning department at KTH - The Royal Institute of Technology, in Stockholm, under the supervision of Professor Patric Jensfelt.