Embedded Files
Trust-aware Interactive Reinforcement Learning for Human Robots Collaborative Bounding Overwatch
Trust-aware Interactive Reinforcement Learning for Human Robots Collaborative Bounding Overwatch
Multi-robot bounding overwatch in offroad faces no only traversability & visibility issues of a route but also the adversaries. It is crucial for a motion planning framework to consider all these factors.
A Time-series Trust Model for Human Multi-robot Collaborative Offroad Motion Task
A Time-series Trust Model for Human Multi-robot Collaborative Offroad Motion Task
In a human multi-robot collaborative task, an appropriate level of trust from the human operator in the multi-robot system (MRS) will increase the willingness of human to delegate tasks to the robots and thus reduce his/her cognitive workload and improves the collaborative task performance.
Bayesian optimization can concurrently infer the unknown individual human trust model in collaborating with the MRS and utilize the trust to plan a trajectory for the MRS. A human-MRS collaborative bounding overwatch task is deployed in the ROS Gazebo simulator to demonstrate that the Bayesian optimization is data-efficient and reduces human teleoperation workload.
Parallel Task and Motion Planning for Multi-Robot with Temporal Logic Constraints
Parallel Task and Motion Planning for Multi-Robot with Temporal Logic Constraints
Formal methods can deal with motion planning problems for a multi-robot with temporal logic constraints. But it has huge computation requirements.
A parallel task and motion planning framework and implemented a multi-robot experiment for manufacturing tasks in a lab setting. The framework can improve the scalability, computational complexity, and execution efficiency of multi-robot motion planning.
Trust-based Multi-Robot Motion Planning with Formal Methods and a Human-in-the-Loop
Trust-based Multi-Robot Motion Planning with Formal Methods and a Human-in-the-Loop
Formal methods are powerful to synthesize verifiably correct robot motion plans. High-level robot motion planning with formal methods can generate motion plans that are correct by construction. Embedding human factors analysis into robot motion planning under temporal logic constraints can improve the trustworthiness of robot motion planning.
Shared Control Based on Kinesthesis and Force Guiding Manipulation for Teleoperation
Shared Control Based on Kinesthesis and Force Guiding Manipulation for Teleoperation
A haptic device is designed as an interface to connect an operator's decision-making and an agent's autonomous navigation under image-based visual servo guidance. Operators could utilize the 6 DOF kinematics-driven haptic device to teleoperate a manipulator.
Page updated
Google Sites
Report abuse