The DRC Plug Task involves unplugging a power cable from one socket and plugging it into another. The problem involves detecting a deformable cable in a noisy environment, grasping the cable with a feasible grasp pose, controlling the shape of the cable to match with the target pose, and inserting the cable-tip to the target socket.
[publication-1][publication-2][code-1][code-2]
Automatic manipulation of linear flexible objects is computationally expensive due to their infinite degrees of freedom in the free spaces. Manipulating such objects in a constrained space raises more challenges due to restricted robot motions.
[publication][code]
To bridge the gap between simulation and real-world, we came up with a new strategy called Simulation-to-Real-to-Simulation. The left figure shows the pipeline of the Sim2Real2Sim strategy we will use for bridging the gap. We start with a rough simulation environment with the estimated models. Then we test the system framework in the real world based on the methods developed in simulation and collect the data from the real world. Finally, we go back to the simulation and update the models and methods based on the data from the real world.
[publication][code]
The plug task, motivated by the 2015 DARPA Robotics Challenge Finals, involves inserting a plug connected with a cable into a paired socket. The human holds the socket while the robot is grasping the cable and manages to insert the plug. Experiments show that the plug task can be completed using our shared control method within 7 seconds.
[publication][code]
This challenge was aimed at looking for unique use cases for the Panasonic Prototype 3D LiDAR. We proposed to use the Panasonic prototype 3D LiDAR to help with developing intelligent wheel-chairs to provide disabled people with independent mobility from home to work and back to home. We mapped the environment by using SLAM with LiDAR and performed pedestrian detection and tracking using point cloud from the Panasonic LiDAR.
[news]
To complete three inspection tasks such as counting containers, recording their ID tags and performing gamma measurements in the nuclear power plant environment, we used a Jackal (UGV) robot with ROS, slam, mapping, real-time object detection, obstacle pose estimation and self-designed scissor lift in this competition. We were the only U.S. team selected to demonstrate the system on 20-24 November 2017 in Brisbane.
[publication][news-1][news-2]
To improve the safety, accuracy and reliability of a humanoid robot - NASA Valkyrie (R5) in completing manipulation tasks, the external contact force was estimated through the dynamic model of this robot. We demonstrated the capability of the force estimator when the task is to pick up a box by two arms under external force disturbances. Our approach results in estimated properties of the object being manipulated such as mass and stiffness, safety warning systems and robot actions being preseted, as well as stable motion control under external force disturbance.
[poster]
Our team explored the use of MATLAB to control a TurtleBot to perform path planning and obstacle avoidance autonomously in a given environment with a known map. To complete this goal, this project is divided in six parts: map acquiring, image processing, path planning, path following, map updating and obstacle avoidance.