Research Associate in "Scene understanding and decision making for human - machine assistance systems" (m/w/d)
About us
The chair of Prof. Bajcinca focuses on research of modern methods and advanced applications of control and system theory, involving three main pillars: cyber-physical systems, complex dynamical systems and machine learning. Through networking with a large number of national and international research, academic and industrial partners, funding projects with exotic and highly interesting tasks regarding model-based and data-driven control have been acquired on a regular basis. The research work is supported with an excellent laboratory equipment and high-performance computation in the areas of autonomous systems, robotics and energy systems, which is continuously being further developed.
https://www.mv.uni-kl.de/mec/home.
Research Scope
A heterogeneous and dynamic environment is often a great challenge for people with visual impairments. Although they use all their other senses for this and learn basic techniques for basic techniques for orientation and mobility, many scenarios arise in everyday life in which they are dependent on external support. The concrete goal of the project is the development of a wearable assistance system that is not only
intended specifically for a particular application or situation, but rather different scenarios in different areas of everyday life. The assistance system includes input devices (cameras and microphone), a computing unit and output devices (headphones and tactile devices) to relay information to the wearer. At this research project, the aim is to develop scene understanding and decision making algorithms for wearable devices that can assist visually impaired and blind people, based on artificial intelligence (AI) and machine learning (ML). These algorithms assist people with visual impairments in perceiving their surroundings and in performing targeted actions for orientation, movement and action. The AI algorithms shall not only capture the dynamic environment, but rather also observe the wearer's behavior, and analyze it in relation to the current situation. Compared to the state of the art, the algorithms should not only be fast, but also energy efficient, in order to keep the weight of the device down and thus not to impair the mobility of the wearer.
Research Task / Work Description
The key motivation of this research stems from the achievements in the domain of robotics and autonomous driving. Computer - driven assistance systems in these domains comprise algorithms layers for perception, planning and decision making. Such autonomous systems must perceive and understand the surrounding real world environment as accurately as possible. Additionally they must generalise across different environments and have the ability to deal with a large variety of objects which may not be known a priori. Although there has been significant advancement in scene understanding methods due to fast development of Deep Learning-based algorithms, their performance often does not meet real-world requirements. In the case of human assistance systems, where the environment is very dynamic and complex, state-of-the-art approaches may lack precision due to sensor uncertainties, decision-making systems difficulties with highly dynamic scenarios and resource-constrained hardware.
Your key responsibility includes:
- Review the literature in the field of scene understanding, decision making for autonomous systems
- Train and evaluate deep/machine learning models like semantic scene segmentation and object detection
- Model the detected static and dynamic objects in an interactive graph and build graph neural networks for scene understanding
- Representing the interactive graph as a spatial-temporal graph of the environment that can be used for a safe planning and decision making
- Evaluate the developed models and publish the results
- Test the developed algorithms in real and complex scenarios
Qualification Requirements
- Above average college degree in computer engineering, computer science, or mathematics
- Advanced computer vision skills beyond the content of basic lectures
- Practical experience with convolutional neural networks, graph neural networks and deep learning libraries e.g. Pytorch, Tensorflow
- Knowledge of at least one programming language: C/C++, Python is expected
- Ability to organise and cooperate with scientific as well as industrial partners of different disciplines
- Business fluent German and English language skills in written and spoken language are an advantage.
We offer
- Payment according to TV-L E13 with an initial one-year time limit
- The possibility to do a PhD and to teach is given in case of scientific aptitude
- TUK strongly encourages qualified female academics to apply
- Severely disabled persons will be given preference in the case of appropriate suitability (please enclose proof)
- Electronic application is preferred. Please attach only one coherent PDF.
You can expect an interesting, diversified and responsible task within a young, highly motivated and interdisciplinary team of a growing chair with great personal creativity freedom.
Contact
Prof. Dr.-Ing. Naim Bajcinca
Phone: +49 (0)631/205-3230
Mobile: +49 (0)172/614-8209
Fax: +49 (0)631/205-4201
Email: mec-apps(at)mv.uni-kl.de
Keywords
Scene Understanding
Decision Making
Computer Vision
Graph Neural Network
Autonomous Systems
Application Papers
Cover Letter
CV
University Certificates
References
List of Publications
Application Deadline
31. October 2023
Job Availability
Immediate