PhD candidate “IMAGE-BASED LOCALIZATION IN UNSTRUCTURED ENVIRONMENTS” (75%) (m/f/d)

At the Fulda University of Applied Informatics, the position is as follows:

(m/w/d)

to the extent of a 75% office limited for the duration of 3 years according to §2 para. 1 WissZeitVG. (code AI-2026-II)

Your main tasks:

› The aim is to answer research questions (I) for the reliability of modern image-based localization methods in real robotic navigation systems under dynamic environmental conditions, (II) for the influence of specific visual representations on robustness and real-time capability of localization pipelines on mobile robots, (III) for suitable system architectures for robust integration of visual localization methods into autonomous navigation systems under limited hardware resources

› methodically, an iterative development process is to be implemented that systematically collects AI training data, trains on selected AI models and evaluates their quality; the following tasks are edited

› Evaluation of the suitability and quality of relevant processes, especially CNN (Convolutional Neural Networks) versus ViT (Vision Transformers)

› Development of a hybrid backbone that combines CNNs and ViTs

› iterative data acquisition, training of AI models and validation of models

› Construction of a modular perception and localization pipeline, e.g. using stereo cameras, GNSS RTK, ViT localization

› experimental evaluation, implementation in the “Living Lab” Campus HS Fulda

› Perception of relevant teaching tasks in the scope of 3 SWS

Your profile:

› a completed university degree (master degree or equivalent university degree) of applied computer science with at least good results

› very good knowledge in the field of design and implementation of system architectures according to current state of the art, under limited hardware resources, including especially for the construction of a modular perception and localization pipeline

› very good knowledge in the field of training of AI models, from iterative data acquisition, to validation and evaluation of results, among others experimentally in application scenarios

› good to very good knowledge of programming with Python, ROS/ROS2 and C++; as well as in DevOps and infrastructure tools like Docker, Kubernetes and GitLab (CI/CD)

› good knowledge in the field of CNNs and ViTs, for use in the field of image-based self-localization methods for robotic, autonomous navigation systems

› Knowledge in the collection and processing of data with mono- and stereo cameras, lidars, IMU, GPS/ GNSS and RTK

› Experience with VR/AR simulations, also in the field of robotics, such as Gazebo

› Experience in scientific work, including publication and presentation of results

› very good English knowledge in word and writing

› Pre-experiences in Higher Education

We offer you:

› A colourful and lively campus life, characterized by personal and value-adding interaction and a family atmosphere

› Flexibility in working life through individual working time models, home office opportunities and measures to promote work-life balance as well as the reconciliation of work and family life

› Variety of continuing and continuing training with an outstanding in-house training program

› Extensive university sports and other health promotion measures

› Our Mensa incl. cafeteria and a food truck with regional and varied dishes

› Our modern university and state library

› Compensation by collective agreement (TV-H), 30 days holiday, child allowance and special payment

› An occupational pension scheme (VBL) with a strong performance package

› The LandesTicket Hessen for free use of the ÖPNV

The grouping takes place in the pay group 13 TV-H if appropriate conditions exist. The Fulda University of Applied Sciences is committed to equal opportunities and explicitly welcomes the application of people in all their diversity. People with disabilities are preferred with the same suitability.

Please send your meaningful application documents to the President of the Fulda University of Applied Sciences by 06.07.2026, by e-mail in a PDF document ([for e-mail see the original ad]). For further information, please contact: Prof. Dr. Jan-Torsten Milde ([for e-mail see the original ad]). Please submit your documents only in a copy (no folders) when you apply by mail, as they cannot be sent back after the completion of the selection procedure. They are destroyed taking into account the data protection regulations.

Applications via the online portal of the jobbörse are not taken into account.

Joblocations:

36037 Fulda, Leipziger Str. 123, HESSEN, GERMANY