Gazebo plug-in for simulating BioTac tactile sensors
- This repository contains a Gazebo plug-in for simulating BioTac tactile sensors.
- The plug-in publishes ROS messages containing simulated electrode and pressure sensor readings as if they were generated by real BioTac sensors.
- The sensor outputs are simulated using an artificial neural network.
- Example projects, training scripts, and a pretrained sensor model are included.
- Clone this reposity
- If you want to run the demos: get the Shadow hand packages and their dependencies
- If you want to train the plug-in yourself instead of using a pretrained model, get Keras, Tensorflow, Apriltags, and the Shadow hand packages
Running the demos
- Launch a simulation environment (eg.
roslaunch sim_biotac_motorhand sim_biotac_motorhand_1.launch, or
roslaunch sim_biotac_motorhand sim_biotac_motorhand_2.launch, or
roslaunch sim_biotac_motorhand ur5_motorhand.launch)
- Run one of the tests (eg.
rosrun sim_biotac_motorhand move_finger_1, or
rosrun sim_biotac_motorhand move_finger_2, or
rosrun sim_biotac_motorhand tripod_grasp)
- You can download the dataset associated with this network model at https://tams.informatik.uni-hamburg.de/research/datasets/index.php#biotac_single_contact_response
- If you use this work, please cite Philipp Ruppel, Yannick Jonetzko, Michael Görner, Norman Hendrich and Jianwei Zhang, Simulation of the SynTouch BioTac Sensor, The 15th International Conference on Intelligent Autonomous Systems, IAS-15 2018, Baden Baden, Germany.