Human-Robot Collaboration Through Personalized Task Learning in Multi-Robot Contexts
Presentation Type
Poster
Faculty Advisor
Weitian Wang
Access Type
Event
Start Date
26-4-2024 2:15 PM
End Date
26-4-2024 3:15 PM
Description
In everyday tasks there is a wide variety and range of ways to complete said task. This is an area of interest when it comes to human-robot collaboration (HRI) because of how robots can support and aid humans in completing various tasks. Robots are used to complete monotonous and repetitive tasks and can handle dangerous situations while humans are flexible and creative with their approach but require assured safety. Therefore, improving a robot’s ability to learn from human demonstrations using decision making models such as a hidden Markov model (HMM) foregoing the need for traditional reprogramming and other highly technical programming techniques can improve the efficiency of HRI and the tasks completed using it. This is especially true when there are multiple robots working together with a human compared to one. This study will create and detail a scalable approach based on the premise detailed above using HMM and a multi-robot setup. The approach will be centered around assembling a simple example object where a human and robot will take turns picking up pieces to assemble the object where it will learn a "policy" or the general way a human likes to assemble the object. We will also take into account safety features like detailing a way a robot should interact with the shared workspace. Ultimately this will result in a working approach which can be demoed and presented, which can be used as a model for various industries implementing HRI.
Human-Robot Collaboration Through Personalized Task Learning in Multi-Robot Contexts
In everyday tasks there is a wide variety and range of ways to complete said task. This is an area of interest when it comes to human-robot collaboration (HRI) because of how robots can support and aid humans in completing various tasks. Robots are used to complete monotonous and repetitive tasks and can handle dangerous situations while humans are flexible and creative with their approach but require assured safety. Therefore, improving a robot’s ability to learn from human demonstrations using decision making models such as a hidden Markov model (HMM) foregoing the need for traditional reprogramming and other highly technical programming techniques can improve the efficiency of HRI and the tasks completed using it. This is especially true when there are multiple robots working together with a human compared to one. This study will create and detail a scalable approach based on the premise detailed above using HMM and a multi-robot setup. The approach will be centered around assembling a simple example object where a human and robot will take turns picking up pieces to assemble the object where it will learn a "policy" or the general way a human likes to assemble the object. We will also take into account safety features like detailing a way a robot should interact with the shared workspace. Ultimately this will result in a working approach which can be demoed and presented, which can be used as a model for various industries implementing HRI.