[Robotics] Visuotactile Perception and Deep RL for Contact-Rich Robotic Manipulation - [Robotics] Visuotactile Perception and Deep RL for Contact-Rich Robotic Manipulation - HRI-US
[Robotics] Visuotactile Perception and Deep RL for Contact-Rich Robotic Manipulation
Job Number: P20INT-32
This title includes multiple positions. The focus of the research is to use vision and tactile sensor data, exploiting finger-object contact information, to enable robots manipulate objects in unstructured environments using machine-learning approaches.
San Jose, CA
You are expected to:
- Explore temporal approaches to track the state of an object over time.
- Explore deep learning approaches that take advantage of object geometry to improve the estimate of the object's state.
- Explore deep reinforcement learning and learning from demonstration approaches to robotic manipulation.
- Implement planning and control algorithms on hardware.
- Ph.D. or highly qualified M.S. candidate in computer science, electrical engineering, mechanical engineering, or related field.
- Experience in deep learning and other machine learning methods.
- Good programming skills in either C++ or Python.
- Experience in Robot Operating System (ROS).
- Experience working with sensors and actuators.
- Experience with deep reinforcement learning and sim-to-real approaches.
- Experience in manipulation, grasping and tactile sensing.
- Experience with PyTorch and TensorFlow.
- Experience with game engines such as Unity and Unreal Engine.
- Experience with implementation of real-time control algorithms on robotic systems.
Duration: 3 months