This Thesis addresses the challenge of adaptive authority allocation in Bilateral Teleoperation by integrating Game Theory (GT) and Fuzzy Logic (FL) into a shared autonomy framework. The main scientific question explored is how to optimally balance Human Operator (HO) control and Autonomous Control System (ACS) intervention in complex manipulation tasks characterized by obstacles in the environment and varying operator skill levels. Motivated by the complementary strengths of humans and robots, the work develops a methodology where the interaction between HO and ACS is modeled as a differential game. Cooperative and non-cooperative game-theoretic formulations are employed to capture both collaborative and competitive dynamics, while role arbitration is implemented through a Fuzzy Inference System that dynamically adjusts the control authority based on task-relevant inputs such as end-effector distance to target, proximity to obstacles and workspace boundaries, and manipulability. The proposed framework enables smooth transitions between fully human, fully autonomous, and intermediate shared control regimes, thereby reducing cognitive load on the operator while maintaining task performance and safety. Trajectory generation and haptic feedback are incorporated to provide intuitive interaction. Simulation and real-world experiments demonstrate the effectiveness of the approach in minimizing operator effort, improving task completion time, and avoiding collisions. The results highlight the advantages of combining GT and FL: GT provides a strategy for dynamic negotiation of control, while FL allows the system to handle the uncertainty in human behavior related to the operator's perception of the remote environment. Overall, this work contributes in developing a shared autonomy architecture suitable for teleoperation scenarios where both safety and efficiency are critical. It establishes a foundation for future extensions to further enhance human-robot collaboration in the fields of remote surgery, hazardous environment manipulation, or cooperative industrial robotics.
Questa tesi affronta la sfida dell’allocazione adattativa dell’autorità nella teleoperazione bilaterale integrando la Teoria dei Giochi (TG) e la Logica Fuzzy (LF) in un framework di autonomia condivisa. La principale questione scientifica indagata riguarda come bilanciare in modo ottimale il controllo dell’operatore umano (HO) e l’intervento del sistema autonomo (ACS) in task di manipolazione complessi, caratterizzati da ostacoli nell'ambiente e differenti livelli di abilità dell’operatore. Motivata dai punti di forza complementari di operatore e robot, la metodologia sviluppata modella l’interazione tra HO e ACS come un gioco differenziale tramite la TG. Formulazioni cooperative e non cooperative sono impiegate per rappresentare sia dinamiche collaborative sia competitive, mentre l’arbitraggio dei ruoli è realizzato tramite un sistema di inferenza fuzzy che regola dinamicamente l’autorità di controllo in base a parametri rilevanti per il compito da svolgere, quali distanza dell’end-effector dall’obiettivo, prossimità a ostacoli e ai limiti del workspace, e grado di manipolabilità del robot. Il framework proposto permette transizioni fluide tra controllo completamente manuale, completamente autonomo e modalità intermedie di autonomia condivisa, riducendo il carico cognitivo sull’operatore e garantendo al contempo prestazioni e sicurezza. Gli esperimenti condotti mostrano che l’approccio proposto riduce lo sforzo dell’operatore, facilita il raggiungimento degli obiettivi e previene collisioni. I risultati evidenziano i vantaggi della combinazione della TG e della LF: la TG fornisce una strategia per la negoziazione dinamica del controllo, mentre la LF consente al sistema di gestire l'incertezza nel comportamento dell'operatore dovuta alla percezione dell'ambiente remoto. Complessivamente, questo lavoro propone un’architettura di autonomia condivisa adatta a scenari di teleoperazione in cui sicurezza ed efficienza sono cruciali, e apre la strada a future estensioni per migliorare ulteriormente la collaborazione uomo-robot in ambiti quali telechirurgia, manipolazione in scenari ostili e robotica industriale collaborativa.
A fuzzy game-theoretic assistance architecture for bilateral teleoperation of robots
DORIGO, MARCO
2024/2025
Abstract
This Thesis addresses the challenge of adaptive authority allocation in Bilateral Teleoperation by integrating Game Theory (GT) and Fuzzy Logic (FL) into a shared autonomy framework. The main scientific question explored is how to optimally balance Human Operator (HO) control and Autonomous Control System (ACS) intervention in complex manipulation tasks characterized by obstacles in the environment and varying operator skill levels. Motivated by the complementary strengths of humans and robots, the work develops a methodology where the interaction between HO and ACS is modeled as a differential game. Cooperative and non-cooperative game-theoretic formulations are employed to capture both collaborative and competitive dynamics, while role arbitration is implemented through a Fuzzy Inference System that dynamically adjusts the control authority based on task-relevant inputs such as end-effector distance to target, proximity to obstacles and workspace boundaries, and manipulability. The proposed framework enables smooth transitions between fully human, fully autonomous, and intermediate shared control regimes, thereby reducing cognitive load on the operator while maintaining task performance and safety. Trajectory generation and haptic feedback are incorporated to provide intuitive interaction. Simulation and real-world experiments demonstrate the effectiveness of the approach in minimizing operator effort, improving task completion time, and avoiding collisions. The results highlight the advantages of combining GT and FL: GT provides a strategy for dynamic negotiation of control, while FL allows the system to handle the uncertainty in human behavior related to the operator's perception of the remote environment. Overall, this work contributes in developing a shared autonomy architecture suitable for teleoperation scenarios where both safety and efficiency are critical. It establishes a foundation for future extensions to further enhance human-robot collaboration in the fields of remote surgery, hazardous environment manipulation, or cooperative industrial robotics.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_12_Dorigo_Thesis.pdf
solo utenti autorizzati a partire dal 19/11/2026
Descrizione: Full Thesis Dorigo Marco
Dimensione
7.31 MB
Formato
Adobe PDF
|
7.31 MB | Adobe PDF | Visualizza/Apri |
|
2025_12_Dorigo_Executive_Summary.pdf
solo utenti autorizzati a partire dal 19/11/2026
Descrizione: Executive Summary Dorigo Marco
Dimensione
1.19 MB
Formato
Adobe PDF
|
1.19 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/246707