%0 Journal Article %1 vaxenburg2024wholebody %A Vaxenburg, Roman %A Siwanowicz, Igor %A Merel, Josh %A Robie, Alice A %A Morrow, Carmen %A Novati, Guido %A Stefanidi, Zinovia %A Card, Gwyneth M %A Reiser, Michael B %A Botvinick, Matthew M %A Branson, Kristin M %A Tassa, Yuval %A Turaga, Srinivas C %D 2024 %I Cold Spring Harbor Laboratory %K biology locomotion modeling reinforcement_learning sensorimotor %R 10.1101/2024.03.11.584515 %T Whole-body simulation of realistic fruit fly locomotion with deep reinforcement learning %U http://dx.doi.org/10.1101/2024.03.11.584515 %X The body of an animal determines how the nervous system produces behavior. Therefore, detailed modeling of the neural control of sensorimotor behavior requires a detailed model of the body. Here we contribute an anatomically-detailed biomechanical whole-body model of the fruit fly Drosophila melanogaster in the MuJoCo physics engine. Our model is general-purpose, enabling the simulation of diverse fly behaviors, both on land and in the air. We demonstrate the generality of our model by simulating realistic locomotion, both flight and walking. To support these behaviors, we have extended MuJoCo with phenomenological models of fluid forces and adhesion forces. Through data-driven end-to-end reinforcement learning, we demonstrate that these advances enable the training of neural network controllers capable of realistic locomotion along complex trajectories based on high-level steering control signals. With a visually guided flight task, we demonstrate a neural controller that can use the vision sensors of the body model to control and steer flight. Our project is an open-source platform for modeling neural control of sensorimotor behavior in an embodied context.

@article{vaxenburg2024wholebody, abstract = {The body of an animal determines how the nervous system produces behavior. Therefore, detailed modeling of the neural control of sensorimotor behavior requires a detailed model of the body. Here we contribute an anatomically-detailed biomechanical whole-body model of the fruit fly Drosophila melanogaster in the MuJoCo physics engine. Our model is general-purpose, enabling the simulation of diverse fly behaviors, both on land and in the air. We demonstrate the generality of our model by simulating realistic locomotion, both flight and walking. To support these behaviors, we have extended MuJoCo with phenomenological models of fluid forces and adhesion forces. Through data-driven end-to-end reinforcement learning, we demonstrate that these advances enable the training of neural network controllers capable of realistic locomotion along complex trajectories based on high-level steering control signals. With a visually guided flight task, we demonstrate a neural controller that can use the vision sensors of the body model to control and steer flight. Our project is an open-source platform for modeling neural control of sensorimotor behavior in an embodied context.}, added-at = {2024-03-18T11:50:35.000+0100}, author = {Vaxenburg, Roman and Siwanowicz, Igor and Merel, Josh and Robie, Alice A and Morrow, Carmen and Novati, Guido and Stefanidi, Zinovia and Card, Gwyneth M and Reiser, Michael B and Botvinick, Matthew M and Branson, Kristin M and Tassa, Yuval and Turaga, Srinivas C}, biburl = {https://www.bibsonomy.org/bibtex/2f518fa31b92aa57c95e7cac6f5b424f2/tabularii}, description = {Whole-body simulation of realistic fruit fly locomotion with deep reinforcement learning | bioRxiv}, doi = {10.1101/2024.03.11.584515}, interhash = {b596cfd4ee37aa2e1850d668c6a7a3e8}, intrahash = {f518fa31b92aa57c95e7cac6f5b424f2}, keywords = {biology locomotion modeling reinforcement_learning sensorimotor}, month = mar, publisher = {Cold Spring Harbor Laboratory}, timestamp = {2024-03-18T11:50:35.000+0100}, title = {Whole-body simulation of realistic fruit fly locomotion with deep reinforcement learning}, url = {http://dx.doi.org/10.1101/2024.03.11.584515}, year = 2024 }