4691196
Description
Mind Map by Jhon Lagos, updated more than 1 year ago
|
Created by Jhon Lagos
about 8 years ago
|
|
Cheek to Chip: Dancing Robots
and Al’s Future
- INTRODUCTION
- Music and dancing are activities best
known human entertainment, and are
linked to the ability to socialize. The
sound and movement patterns
characterize the dance.
- Robots That Learn to Dance from Observation
- This
- Recent generations of humanoid robots
increasingly resemble humans
- This progress
- Has led researchers to design robots that can mimic
dancers complexity and style of dance choreography
human.
- Has led researchers to design robots that can mimic
dancers complexity and style of dance choreography
human.
- This progress
- Recent generations of humanoid robots
increasingly resemble humans
- This
- It allows
- the robot what to do and how to do it from
observation. The model is divided into two legs.
- At the end they concatenate the two and dynamic
filters are applied to prevent the robot hit with it.
- the robot what to do and how to do it from
observation. The model is divided into two legs.
- Robots That Learn to Dance from Observation
- Music and dancing are activities best
known human entertainment, and are
linked to the ability to socialize. The
sound and movement patterns
characterize the dance.
- for this
- Researchers at the University of Tokyo have developed
the learning-from-observation (LFO) training method
- Generating upper-body motion
- use the same task-model strategy as for leg
motions. In this case, each model represents
a key pose of the imitated dance.
- use the same task-model strategy as for leg
motions. In this case, each model represents
a key pose of the imitated dance.
- Generating whole-body motion
- To generate executable motion, we use a dynamic filter and
conduct skill refinement. The dynamic filter compensates the
zero-moment point and the yaw-axis moment
- To generate executable motion, we use a dynamic filter and
conduct skill refinement. The dynamic filter compensates the
zero-moment point and the yaw-axis moment
- Motion from Sound: Intermodal Neural Network Mapping
- Robots that interact with humans must be
able to react to multimodal sensory input.
- Intermodality mapping
- consists of
- two phases. First, in the learning phase, the robot
observes some events together with associated sounds.
The robot memorizes these sounds along with the
motions of the sound source
- Keepon
- uses signal processing techniques
and accelerometers to detect
movements of the person
- uses signal processing techniques
and accelerometers to detect
movements of the person
- Dance Partner
- It uses a database of
knowledge and estimates
to predict the movement
- It uses a database of
knowledge and estimates
to predict the movement
- Keepon
- two phases. First, in the learning phase, the robot
observes some events together with associated sounds.
The robot memorizes these sounds along with the
motions of the sound source
- Control system architecture
- uses
- Step transitions that is, when and how the human partner
changes steps in a dance sequence are important for the
robot to generate its own dancing motion
- Step transitions that is, when and how the human partner
changes steps in a dance sequence are important for the
robot to generate its own dancing motion
- uses
- consists of
- Robots that interact with humans must be
able to react to multimodal sensory input.
- Generating upper-body motion
- Researchers at the University of Tokyo have developed
the learning-from-observation (LFO) training method
Want to create your own Mind Maps for free with GoConqr? Learn more.