Integration of hardware accelerators in humanoid robot platform

Integration of hardware accelerators in humanoid robot platform

roboter KIT

Context

A new generation of assistance robots is being researched and developed as part of the JuBot research project. These are to support seniors in their daily lives and be able to react in real time to changes in their environment. To ensure this, hardware accelerators for Deep Learning (DL) are used. However, these need to be integrated into the robot framework in order to be used successfully.

Tasks

In the context of this Hiwi activity, existing hardware accelerators for camera data-based DL functions, such as face/object recognition, are to be integrated into the robot framework. In particular, the data exchange between framework and accelerator shall be ensured and optimized.

Requirements

  • Programming experience in Python and C++
  • Knowledge in hardware design is advantageous
  • Motivation and interest in solving technical problems independently