This thesis explores the intersection of human-centric design and robotics, focusing on enhancing human-robot interaction through predictive, assistive, and planning methodologies. The research is structured around three key objectives: predicting human motion in shared spaces using semantic maps and advanced neural architectures; designing adaptive shared control frameworks for assistive robotic devices like the FriWalk; and developing human-aware motion planning for collaborative robotic manipulators such as the UR5e. Employing tools like Vision Transformers (ViTs) and Masked Autoencoders, the study achieves high-accuracy predictions of human trajectories and occupancy priors, which are essential for robots operating in dynamic environments. The shared control framework balances safety and user autonomy by dynamically adjusting robotic assistance based on behavioural analysis. For robotic manipulators, real-time human motion predictions integrate into trajectory planning algorithms, ensuring seamless and safe collaboration in mixed environments. The findings advance the field of human-aware robotics, contributing to safer, more intuitive interactions between humans and robots. This work lays the groundwork for future assistive technology and collaborative robotics developments, aiming to enhance safety, efficiency, and user autonomy in diverse applications.
Human-Aware Robotics: Predict, Assist, and Plan for Seamless Interaction
Falqueto, Placido
2025
Abstract
This thesis explores the intersection of human-centric design and robotics, focusing on enhancing human-robot interaction through predictive, assistive, and planning methodologies. The research is structured around three key objectives: predicting human motion in shared spaces using semantic maps and advanced neural architectures; designing adaptive shared control frameworks for assistive robotic devices like the FriWalk; and developing human-aware motion planning for collaborative robotic manipulators such as the UR5e. Employing tools like Vision Transformers (ViTs) and Masked Autoencoders, the study achieves high-accuracy predictions of human trajectories and occupancy priors, which are essential for robots operating in dynamic environments. The shared control framework balances safety and user autonomy by dynamically adjusting robotic assistance based on behavioural analysis. For robotic manipulators, real-time human motion predictions integrate into trajectory planning algorithms, ensuring seamless and safe collaboration in mixed environments. The findings advance the field of human-aware robotics, contributing to safer, more intuitive interactions between humans and robots. This work lays the groundwork for future assistive technology and collaborative robotics developments, aiming to enhance safety, efficiency, and user autonomy in diverse applications.File | Dimensione | Formato | |
---|---|---|---|
falqueto_phd_thesis.pdf
accesso aperto
Dimensione
14.34 MB
Formato
Adobe PDF
|
14.34 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/217523
URN:NBN:IT:UNITN-217523