Reinforcement Learning-based Switching Controller for a Milliscale Robot in a Constrained Environment.
Abbas TariverdiUlysse Côté AllardKim MathiassenOle Jakob ElleHåvard KalvøyØrjan Grøttem MartinsenJim TørresenPublished in: CoRR (2021)
Keyphrases
- mobile robot
- reinforcement learning
- real robot
- control architecture
- autonomous robots
- simulated robot
- real time
- mobile robotics
- robotic systems
- motion control
- navigation tasks
- robot control
- indoor environments
- control system
- function approximation
- optimal control
- robotic agents
- adaptive control
- robotic manipulator
- learning algorithm
- model free
- biped walking
- reward signal
- visual landmarks
- robot localization
- dynamic environments
- control loop
- actor critic
- home environment
- unknown environments
- path planning
- service robots
- real environment
- robot navigation
- human robot interaction
- control law
- multi agent
- state space
- rough terrain
- position and orientation
- robot behavior
- changing environment
- legged robots
- learning agent
- robot manipulators
- autonomous learning
- optimal policy