At Bosch, we shape the future by inventing high-quality technologies and services that spark enthusiasm and enrich people’s lives. Our promise to our associates is rock-solid: we grow together, we enjoy our work, and we inspire each other. Join in and feel the difference.

The Robert Bosch GmbH is looking forward to your application!

Robots are increasingly part of everyday life, from delivery robots on pavements to service robots in restaurants. Ensuring smooth motion while safely avoiding humans remains a challenge. Properly utilizing uncertainty-aware human trajectory prediction is one of the key aspects in achieving smooth robot motion among humans. Recent work on human trajectory prediction uses non-parametric distributions for human motion uncertainties, allowing for more diverse and accurate predictions. While leveraging advanced trajectory predictions could enhance robot motion planning performance, the literature on utilizing these results in mobile robot motion planning is sparse. Uncertainties modelled by Gaussian distributions (or Gaussian mixture models) can be treated in stochastic model predictive control (MPC) by tightening the constraints based on the uncertainty level sets quite straightforwardly. However, for more general (including non-parametric) distributions, parameterizing such uncertainties for constraint robustness is much more challenging. Scenario-based stochastic MPC provides a way to handle such uncertainty distributions, but this results in a large number of constraints, making the optimal control problem (OCP) difficult to solve.

• In this thesis, you will investigate appropriate ways of sampling and selecting the human uncertain trajectories to be imposed as constraints in the OCP, and assess the resulting computational efficiency and robot motion performance. The developed controller will be implemented as a nav2 plugin and deployed on a physical mobile robot. The objective is to showcase a demonstration in which the robot drives among people in the canteen at lunchtime, navigating through narrow, twist-and-turning corridors and avoiding moving people based on their motion predictions.
• You will get the prediction of human trajectories and the uncertainty measurements using state-of-the-art neural networks.
• Furthermore, you will formulate an OCP such that optimal robot trajectories probabilistically satisfy the collision avoidance constraints given the human trajectory predictions.
• In addition, you will solve the resulting OCP fast enough for real-time applications.
• Last but not least, you will experimentally validate your method on physical mobile robots around humans.

• Education: Master studies in the field of Robotics, Systems and Control or comparable
• Experience and Knowledge: affinity with programming in Python, C++, CUDA and Rust; experience with ROS2 is an advantage
• Personality and Working Practice: you have strong communication skills, especially when it comes to explaining technical concepts; you are enthusiastic about acquiring new knowledge
• Enthusiasm: eager to gain hands-on experience working with real robots
• Languages: fluent in written and spoken English

Start: according to prior agreement
Duration: 6 months

Requirement for this thesis is the enrollment at university. Please attach your CV, transcript of records, examination regulations and if indicated a valid work and residence permit.

Diversity and inclusion are not just trends for us but are firmly anchored in our corporate culture. Therefore, we welcome all applications, regardless of gender, age, disability, religion, ethnic origin or sexual identity.

Need further information about the job?
Yunfan Gao (Functional Department)
+49 711 811 46473
Niels Van Duijkeren (Functional Department)
+49 711 811 59140

#LI-DNI