Dexterous robotic manipulation using deep reinforcement learning and knowledge transfer for complex sparse reward-based tasks.
Qiang WangFrancisco Roldan SanchezRobert McCarthyDavid Cordova BulensKevin McGuinnessNoel E. O'ConnorManuel WüthrichFelix WidmaierStefan BauerStephen J. RedmondPublished in: Expert Syst. J. Knowl. Eng. (2023)
Keyphrases
- knowledge transfer
- reinforcement learning
- manipulation tasks
- transfer learning
- transferring knowledge
- real robot
- object manipulation
- cross domain
- robot navigation
- robotic systems
- learning tasks
- motion planning
- transfer knowledge
- previously learned
- knowledge sharing
- multi task
- service robots
- learning algorithm
- real world
- reinforcement learning algorithms
- mutual adaptation
- human robot interaction
- state space
- humanoid robot
- mobile robot
- active learning
- information retrieval
- machine learning
- data sets
- model free
- markov decision processes
- multiple tasks
- optimal policy