A model for biped walking is proposed. The model has a compass-like mechanism with telescopic type legs, and uses a simple leg stretch-contraction motion pattern. The model with a point shaped foot is able to represen...
详细信息
A model for biped walking is proposed. The model has a compass-like mechanism with telescopic type legs, and uses a simple leg stretch-contraction motion pattern. The model with a point shaped foot is able to represent human walking characteristics. A systematic conversion from the point shaped foot to the equivalent of a normal surface foot - resembling the shape that of humans, this extension preserves the same human-like walking characteristics while gaining the benefits of added stability and the reduction of impact forces during foot contact.
Smooth control using an active vision head's verge-axis joint is performed through continuous state and action reinforcement learning. The system learns to perform visual servoing based on rewards given relative t...
详细信息
Smooth control using an active vision head's verge-axis joint is performed through continuous state and action reinforcement learning. The system learns to perform visual servoing based on rewards given relative to tracking performance. The learned controller compensates for the velocity of the target and performs lag-free pursuit of a swinging target. By comparing controllers exposed to different environments we show that the controller is predicting the motion of the target by forming an implicit model of the target's motion. Experimental results are presented that demonstrate the advantages and disadvantages of implicit modelling.
This paper focuses on learning to select behavioral primitives and generate sub-goals from practicing a task. We present a novel algorithm that combines Q-learning and a locally weighted learning method to improve pri...
详细信息
This paper focuses on learning to select behavioral primitives and generate sub-goals from practicing a task. We present a novel algorithm that combines Q-learning and a locally weighted learning method to improve primitive selection and sub-goal generation. We demonstrate this approach applied to the tilt maze task. Our robot initially learns to perform this task using learning from observation, and then learns from practice.
暂无评论