Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Music-to-Dance trajectory generation for a dynamically stable mobile manipulator
This master thesis is about implementing a music-to-dance trajectory generator for a ball-balancing manipulator. It involves selecting a set of motion libraries which are feasible for the robot, interpolate them, and implement a probabilistic model to link the motions to dance moves.
Keywords: Motion library, Optimal Control, State Estimation
In this project, we want to develop a model that associates music to feasible dance moves, which should be executed online by a dynamically stable mobile manipulator (the ball-balancing robot shown in the picture). The student will have to: 1) select a finite set of motion primitives, that the robot is already capable of tracking; 2) extract the beat signal from songs using existing tools; 3) implement a (probabilistic) model to associate the beat to the motion primitives.
The output of the thesis will consist in one or two demos that the robot will be able to execute, based on the sound of one or two songs that will be decided during the thesis' time.
References: -) Bi, Thomas, et al. "Real-time dance generation to music for a legged robot." 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2018. -) https://www.youtube.com/watch?v=A7_e-UWkfXo -) https://www.youtube.com/watch?v=39rRhTqcQc0
In this project, we want to develop a model that associates music to feasible dance moves, which should be executed online by a dynamically stable mobile manipulator (the ball-balancing robot shown in the picture). The student will have to: 1) select a finite set of motion primitives, that the robot is already capable of tracking; 2) extract the beat signal from songs using existing tools; 3) implement a (probabilistic) model to associate the beat to the motion primitives.
The output of the thesis will consist in one or two demos that the robot will be able to execute, based on the sound of one or two songs that will be decided during the thesis' time.
References: -) Bi, Thomas, et al. "Real-time dance generation to music for a legged robot." 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2018. -) https://www.youtube.com/watch?v=A7_e-UWkfXo -) https://www.youtube.com/watch?v=39rRhTqcQc0
-) C++ programming -) python -) basic hardware experience
-) C++ programming -) python -) basic hardware experience
We are looking for extremely motivated and excellent students, that already have experience in programming, a good theoretical background, and interest in doing hardware experiments. Experience with reinforcement learning algorithms is also appreciated.
We are looking for extremely motivated and excellent students, that already have experience in programming, a good theoretical background, and interest in doing hardware experiments. Experience with reinforcement learning algorithms is also appreciated.