Postdoc Opening: Modeling bicycle behavior from visual stimulus to neuromuscular control ability


TU Delft is seeking a postdoctoral researcher for a 1 year equivalent contract (full or part time). The researcher will develop and execute an engineering research project with the title "Modeling bicycle behavior from visual stimulus to neuromuscular control ability". This project is funded by a TU Delft internal Cohesion grant and is a collaboration between the Intelligent Vehicles section and the Bicycle Laboratorium, both part of the Mechanical Engineering faculty. The purpose of the project is to use Computer Vision and bicycle dynamics models to accurately model cyclist behavior to improve bicycle and road designs, as well as safety in sustainable transportation.

Accurate motion data on cyclists is scarce. Existing data has been largely collected via counting at specific locations or GPS. However, the accuracy of GPS is limited and there is no data that can track bicyclists with centimeter accuracy in urban environments. We recently developed a bicycle equipped with the latest automotive-grade sensors. By using visual odometry from IMUs, lidar and cameras, we are able to drastically improve the localization performance over GPS. We are further developing a perception system to detect other traffic participants from the bicycle.

In the next stage you will equip cyclists with eye tracking devices to capture their gaze direction during rides in crowded urban environments (e.g. the Delft campus). The collected data will serve as input to a motion planner and mapped to neuromuscular balance ability, i.e. reaction time and sensing quality, needed to safely control the bicycle. By modeling how gaze influences decision making and how vision interacts with vestibular and proprioceptive sensing to maintain balance in safety critical maneuvers, you will gain crucial insights into how individuals perceive risks such as stop signs and crossing pedestrians, and the corresponding actions they undertake, such as braking and counter-steering.


You should have:

  • PhD in an engineering discipline by the start date of the position - Experience with machine learning for computer vision
  • Strong scientific programming skills
  • Strong written and oral communication skills in English
  • The ability to act independently as well as to collaborate effectively with members of a larger team

The following aspects will help you stand out:

  • A strong academic track record with publications in topics relevant to this position
  • Knowledge of automated driving from perception to planning
  • Specific knowledge of deep learning and data efficient methods
  • Knowledge of human-machine control and two-wheeler vehicle dynamics
  • Knowledge of Python and PyTorch

Keep in mind that this describes the background we imagine would best fit the role. Even if you do not meet all of the requirements and feel that you are up for the task, we absolutely want to see your application!

Additional Information

In this role you will be embedded with the Intelligent Vehicles section in the Cognitive Robotics Department and also a member of the Bicycle Lab in the Biomechanical Engineering Department. Assistant Professors Holger Caesar and Jason K. Moore will be your supervisors. For more information about this vacancy, please contact Holger Caesar (

For information about the selection procedure, please contact the HR Advisor at

How to Apply

Apply via the TU Delft website application portal linked below. We will not process applications sent by email and/or post.

Visit TU Delft's website to find more employment details and apply for the position (vacancy number: TUD04759):

Apply by January 31, 2024 and include:

  • a letter of motivation explaining why you are the right candidate for this job,
  • a detailed CV,
  • a copy of your diploma's,
  • the names and contact addresses of two or three references,
  • any other information that might be relevant to your application.

Note that the available position will be filled as soon as possible (i.e. once a suitable candidate is found).