Software developed by the team NAIST-Panasonic at the Amazon Robotics Challenge 2017 (version 2017.7.31).
The ROS packages are organized in online (those used live during the competition), offline (those used for data collection, calibration, etc.) and third-party.
-
Online:
- tnp_task_manager: Orchestrates all the nodes to complete the tasks.
- tnp_kuka_motion: Controls the robot arm motion.
- tnp_moveit_planner: Motion planner using MoveIt!
- tnp_tool: End effector tool for motion planning.
- tnp_deep_vision: Implements YOLOv2 for object recognition and grasping points estimation.
- tnp_feat_vision: Object recognition based on color histogram, HOG, weight and volume.
- tnp_recognition_space: Gets the RGB-D streams of four cameras.
- tnp_background_remover_bbox_extractor: Removes background and extracts the bounding box.
- tnp_bbox_and_background_substractor: Removes background and extracts the bounding box.
- tnp_end_effector: Control and sensing for the gripper and suction fuctions of the end effector.
- tnp_serial_com: Serial communication for the Arduinos.
- tnp_grasp_planner: Calculates suckable surfaces and saves the octomaps of the storage system.
- tnp_weight_events: Object recognition based on weight.
- tnp_optoforce: Manages the OptoForce sensors.
- tnp_monitor: GUI to visualize the whole system status, object recognition and grasping estimation results, voting results, task progress, etc.
- tnp_simple_state_machine: Conveys the nodes' status to the tnp_monitor.
- tnp_msgs: Messages to convey the status of some nodes for the tnp_monitor.
- tnp_led_control: Controls the illumination.
- tnp_database_loader: IO operations for json files.
-
Offline:
- tnp_image_taker: Collects image data for training purposes.
- tnp_hand_eye_calib: Hand-eye calibration for the end effector using multiple poses.
- tnp_ee_calibration: Hand-eye calibration for the end effector using a calibration cube.
- tnp_merge_capture_folder: Merges two folders.
- tnp_scanner: Implements a 3D scanner using the four cameras of the recognition space.
- tnp_recognition_space_calibration: Calibrates the four cameras using a calibration cube with markers.
- tnp_recognition_space_configuration: Configuration of the recognition space.
- tnp_simple_capture: Takes a single frame using the four cameras of the recognition space.
- tnp_svm: Implements a Support Vector Machine for the weight and volume.
- tnp_test_camera_orientation: Simple test to verify the orientation of the cameras.
-
Third-party:
- iiwa_stack: ROS package for the KUKA LBR iiwa R820 developed by Salvo Virga (TUM).
- yolo_light: ROS node implementing YOLOv2 for deep-learning-based object recognition.
- octomap_ros: ROS package to provide conversion functions between ROS and OctoMap's native types.
- octomap_rviz_plugins: RViz display plugins for visualizing octomap messages in ROS.
- phoxi_camera: Software for the time-of-flight camera developed by Photoneo.
The CAD models, desings and diagrams can be found in the tnp_cads_and_diagrams directory.
The docker folder provides a dockerfile to create an image including all the dependencies necessary to run our system.
The code in this repository is released under the Apache License version 2.0, with the exception of third-party code that keeps its original license.
- Gustavo A. Garcia R. (Captain)
- Lotfi El Hafi
- Felix von Drigalski
- Wataru Yamazaki
- Viktor Hoerig
- Arnaud Delmotte
- Akishige Yuguchi
- Marcus Gall
- Chika Shiogama
- Kenta Toyoshima
- Pedro Uriguen
- Rodrigo Elizalde
- Masaki Yamamoto
- Yasunao Okazaki
- Kazuo Inoue
- Katsuhiko Asai
- Ryutaro Futakuchi
- Seigo Okada
- Yusuke Kato
- Pin-Chu Yang