Matches in SemOpenAlex for { <https://semopenalex.org/work/W2913446087> ?p ?o ?g. }
- W2913446087 abstract "In this project we will use reinforcement learning, the CACLA algorithm, to let an agent learn to control a robotic arm. Inspired by domestic service robots that have to perform multiple complex tasks, manipulation is only a small part of it. Using neural networks an agent should be able to learn to complete a manipulation task without having to calculate paths and grasping points. We will be using a 6 degree of freedom robotica arm, Mico, and make use of a simulator called V-REP to perform the experiments. We compare the results to a traditional simple inverse kinematic solver to see if there is a gain in speed, accuracy or robustness. Whilst most agents use one neural network to perform their task, we will experiment with different architectures, namely the amount of neural networks that each control a sub-set of the joints, to see if this can improve results. Whilst for reinforcement learning exploration is very important we test two different exploration methods; Gaussian exploration and Ornstein-Uhlenbeck process exploration, to see if there is any influence in the training. We experimented first with letting the end effector of the arm move to a certain position without grasping an object. It was shown that when using only 1 joint learning is very easy, but when controlling more joints the problem of simply going to a single location becomes more difficult to solve. While adding more training iterations can improve results, it also takes a lot longer to train the neural networks. By showing a pre training stage consisting of calculating the forward kinematics without relying on any physics simulation to create the input state of the agent, we can create more examples to learn from and improve results and decrease the learning time. However when trying to grasp objects the extra pre training stage does not help at all. By increasing the training iterations we can achieve some good results and the agent is able to learn to grasp an object. However when using multiple networks to control a sub-set of joints we can improve on the results, even reaching a 100% success rate for both exploration methods, not only showing that multiple networks can outperform a single network, also that exploration does not influence training all that much. The downside is that training takes a very long time. Whilst it does not outperform the inverse kinematic solver we do have to take into account that the setup was relatively easy, therefore making it very easy for the inverse kinematic solver." @default.
- W2913446087 created "2019-02-21" @default.
- W2913446087 creator A5005013711 @default.
- W2913446087 date "2018-03-26" @default.
- W2913446087 modified "2023-09-23" @default.
- W2913446087 title "Learning to Grasp Objects with Reinforcement Learning" @default.
- W2913446087 cites W131069610 @default.
- W2913446087 cites W1487859867 @default.
- W2913446087 cites W1498436455 @default.
- W2913446087 cites W1510186039 @default.
- W2913446087 cites W1515881658 @default.
- W2913446087 cites W1518672027 @default.
- W2913446087 cites W1925816294 @default.
- W2913446087 cites W1996213539 @default.
- W2913446087 cites W2016876803 @default.
- W2913446087 cites W2041376653 @default.
- W2913446087 cites W2082511574 @default.
- W2913446087 cites W2102605133 @default.
- W2913446087 cites W2110304639 @default.
- W2913446087 cites W2110762409 @default.
- W2913446087 cites W2139137304 @default.
- W2913446087 cites W2154543439 @default.
- W2913446087 cites W2169136412 @default.
- W2913446087 cites W2201912979 @default.
- W2913446087 cites W2246023205 @default.
- W2913446087 cites W2341171179 @default.
- W2913446087 cites W2569836025 @default.
- W2913446087 cites W2790921766 @default.
- W2913446087 cites W2890250492 @default.
- W2913446087 cites W2964043796 @default.
- W2913446087 cites W3145506661 @default.
- W2913446087 hasPublicationYear "2018" @default.
- W2913446087 type Work @default.
- W2913446087 sameAs 2913446087 @default.
- W2913446087 citedByCount "1" @default.
- W2913446087 countsByYear W29134460872020 @default.
- W2913446087 crossrefType "dissertation" @default.
- W2913446087 hasAuthorship W2913446087A5005013711 @default.
- W2913446087 hasConcept C104317684 @default.
- W2913446087 hasConcept C119857082 @default.
- W2913446087 hasConcept C121332964 @default.
- W2913446087 hasConcept C127413603 @default.
- W2913446087 hasConcept C154945302 @default.
- W2913446087 hasConcept C171268870 @default.
- W2913446087 hasConcept C177264268 @default.
- W2913446087 hasConcept C17816587 @default.
- W2913446087 hasConcept C185592680 @default.
- W2913446087 hasConcept C187523126 @default.
- W2913446087 hasConcept C199360897 @default.
- W2913446087 hasConcept C201995342 @default.
- W2913446087 hasConcept C2780451532 @default.
- W2913446087 hasConcept C39920418 @default.
- W2913446087 hasConcept C41008148 @default.
- W2913446087 hasConcept C50644808 @default.
- W2913446087 hasConcept C55493867 @default.
- W2913446087 hasConcept C63479239 @default.
- W2913446087 hasConcept C74650414 @default.
- W2913446087 hasConcept C90509273 @default.
- W2913446087 hasConcept C97541855 @default.
- W2913446087 hasConceptScore W2913446087C104317684 @default.
- W2913446087 hasConceptScore W2913446087C119857082 @default.
- W2913446087 hasConceptScore W2913446087C121332964 @default.
- W2913446087 hasConceptScore W2913446087C127413603 @default.
- W2913446087 hasConceptScore W2913446087C154945302 @default.
- W2913446087 hasConceptScore W2913446087C171268870 @default.
- W2913446087 hasConceptScore W2913446087C177264268 @default.
- W2913446087 hasConceptScore W2913446087C17816587 @default.
- W2913446087 hasConceptScore W2913446087C185592680 @default.
- W2913446087 hasConceptScore W2913446087C187523126 @default.
- W2913446087 hasConceptScore W2913446087C199360897 @default.
- W2913446087 hasConceptScore W2913446087C201995342 @default.
- W2913446087 hasConceptScore W2913446087C2780451532 @default.
- W2913446087 hasConceptScore W2913446087C39920418 @default.
- W2913446087 hasConceptScore W2913446087C41008148 @default.
- W2913446087 hasConceptScore W2913446087C50644808 @default.
- W2913446087 hasConceptScore W2913446087C55493867 @default.
- W2913446087 hasConceptScore W2913446087C63479239 @default.
- W2913446087 hasConceptScore W2913446087C74650414 @default.
- W2913446087 hasConceptScore W2913446087C90509273 @default.
- W2913446087 hasConceptScore W2913446087C97541855 @default.
- W2913446087 hasLocation W29134460871 @default.
- W2913446087 hasOpenAccess W2913446087 @default.
- W2913446087 hasPrimaryLocation W29134460871 @default.
- W2913446087 hasRelatedWork W121431861 @default.
- W2913446087 hasRelatedWork W1838356191 @default.
- W2913446087 hasRelatedWork W2100370041 @default.
- W2913446087 hasRelatedWork W2257547405 @default.
- W2913446087 hasRelatedWork W2513373085 @default.
- W2913446087 hasRelatedWork W2565122071 @default.
- W2913446087 hasRelatedWork W2602553746 @default.
- W2913446087 hasRelatedWork W2735268712 @default.
- W2913446087 hasRelatedWork W2741926431 @default.
- W2913446087 hasRelatedWork W2774742309 @default.
- W2913446087 hasRelatedWork W2950622182 @default.
- W2913446087 hasRelatedWork W2963094133 @default.
- W2913446087 hasRelatedWork W2977416770 @default.
- W2913446087 hasRelatedWork W2979421998 @default.
- W2913446087 hasRelatedWork W2999490157 @default.
- W2913446087 hasRelatedWork W3095308078 @default.
- W2913446087 hasRelatedWork W3159526778 @default.