Convolutional fitted Q iteration for vision-based control problems.
Dongbin ZhaoYuanheng ZhuLe LvYaran ChenQichao ZhangPublished in: IJCNN (2016)
Keyphrases
- control problems
- fitted q iteration
- reinforcement learning
- augmented reality
- human computer interaction
- optimal control
- machine learning
- vision system
- state space
- queueing systems
- deep learning
- learning algorithm
- computer vision
- real time
- markov decision processes
- model free
- sparse coding
- optimal policy
- adaptive control
- learning problems
- restricted boltzmann machine
- approximate dynamic programming
- driver assistance systems
- stochastic control
- hamilton jacobi bellman