In public perception, collaborative robotics often evokes visions of user-friendly robots working alongside operators in effective and efficient industrial processes. However, this does not reflect the current industrial landscape. In reality, most collaborative robots are used as automation devices with minimal human collaboration. Although the promise of cobots lies in their ability to work with workers, significant technical and operational limitations often prevent this ideal from being realised. For example, issues such as limited adaptability to changing work environments and the complexity of integrating cobots into human workflows often reduce their use to mere automation tools. From an industrial perspective, this shift is understandable: companies are investing in cobots with the expectation of flexibility and collaboration. However, when faced with these limitations, they often use cobots for tasks similar to traditional automation to justify the financial investment. This discrepancy between perception and reality highlights a critical gap in the field of robotics: a gap that this thesis attempts to fill. This thesis proposes a methodological framework, along with some methods and technological solutions, to overcome the gap of adopting cobots as mere automation tools. This gap has been investigated through a survey and a new industrial application, which have confirmed a few benefits along with the limitations of cobots in full automation. The results from the survey have highlighted the limitations from which some key challenges for future research have been identified that might help researchers and practitioners to navigate the complex landscape of collaborative robotics. The limitations, as well as the partial benefits, experienced by manufacturing companies when adopting cobots as conventional automation were investigated through the development of a new industrial work cell in a small and medium enterprise (SME). To increase the flexibility of the application, which the company already conceived to work as fully automatic, part program parameterisation and a user-friendly interface have been combined to perform welding, grinding, and polishing tasks on tweezers. The results show that such a method facilitates the introduction of cobots in SMEs. However, doubts arose about using the collaborative robot differently in more collaborative forms, which would further reduce the need for ad hoc developments and increase their flexibility. Based on the first two activities, a methodological framework has been proposed to help overcome the existing limitations in human-robot collaboration applications. The framework is based on the paradigm of a mutualistic and adaptive production system, which inspired the idea for this PhD thesis. Based on these principles, the framework helps to consider all critical factors that enable full collaboration between humans and cobots. In particular, the framework, alongside its guidelines, aims to support practitioners and researchers in developing human-robot collaboration applications that increase efficiency, but also the well-being of workers by ensuring they feel comfortable and safe working with cobots and enabling greater flexibility. The framework has been applied and validated in the development and operation of a screwdriving application in a SME. The results show an improvement in process performance and worker well-being. The digitalisation of the collaborative system and the workers interacting with it were examined using the concept of the Human Digital Twin. First, a meta-model was proposed to represent human workers in production systems. This meta-model facilitates the modelling of the digital representation of humans. Still, it does not support the creation of specific instances, as demonstrated by the validation scenario in an industrial collaborative parts handling scenario. To avoid the development of ad-hoc implementations with each new digital replica that lacks reusability, scalability and extensibility, the work proposes an extensible and flexible IIoT platform. One implementation of the platform has been tested at both industrial and laboratory scale. The introduction of true collaboration between humans and robots impacts human factors. For this reason, a series of experiments based on the Tower of Hanoi were conducted to complement the framework. These experiments, in which mental workload, trust, propensity to trust and anxiety were investigated, provide statistical evidence of the dynamics of the selected human factors in relation to the size, speed and collaborative modes. In addition, the same setup was used to investigate how cobot's anticipatory actions affect task performance and human perception. Finally, a new screwdriving application was developed starting from the industrial scenario to show that increasing the situational and operational awareness of collaborative robots and enabling simultaneous and adaptive actions and decisions can increase both the efficiency and effectiveness of collaboration. In a second scenario, the optimisation of a screwdriving task through simulation, Z-angle redundancy and the algorithm of the travel salesman problem was proposed to determine the optimal configuration of the robot and a sequence that minimises the cycle time. This method shows relevant results regarding cycle time reduction in both automatic and collaborative modes.
Nella percezione pubblica, i robot collaborativi evocano spesso l'immagine di robot che lavorano fianco a fianco degli operatori in modo efficace ed efficiente. Tuttavia, ciò non riflette la realtà del contesto industriale. La maggior parte dei robot collaborativi è utilizzata come dispositivi di automazione con un'interazione minima con gli esseri umani. Sebbene la promessa dei cobot risieda nella collaborazione con i lavoratori, numerose limitazioni tecniche e operative impediscono spesso il suo raggiungimento. Ad esempio, problemi come la scarsa adattabilità a contesti di lavoro variabili e la complessità dell'integrazione dei cobot nei flussi di lavoro umani ne riducono spesso l'uso a semplici strumenti di automazione. Dal punto di vista industriale, questo approccio è comprensibile: le aziende investono nei cobot aspettandosi flessibilità e collaborazione, ma, di fronte a queste difficoltà, li impiegano in compiti simili a quelli dell’automazione tradizionale per giustificarne l’investimento economico. Questa discrepanza tra percezione e realtà evidenzia una lacuna critica nel campo della robotica: una lacuna che questa tesi cerca di colmare. Questa tesi propone un framework, insieme a metodi e tecnologie, per superare l’adozione dei cobot come semplici strumenti di automazione. Questa lacuna è stata studiata attraverso un’indagine e una applicazione industriale, che hanno confermato alcuni benefici insieme ai limiti dei cobot come dispositivi di pura automazione. I risultati dell’indagine hanno messo in evidenza diverse limitazioni, dalle quali sono emerse alcune sfide chiave per il futuro, che potrebbero aiutare ricercatori e professionisti a orientarsi nel complesso panorama dei robot collaborativi. Le limitazioni, così come i benefici parziali riscontrati dalle aziende manifatturiere nell’adozione dei cobot come strumenti di automazione convenzionale, sono stati analizzati attraverso lo sviluppo di una nuova cella di lavoro in una piccola e media impresa (PMI). Per aumentare la flessibilità dell’applicazione, inizialmente concepita dall’azienda come completamente automatica, sono stati combinati la parametrizzazione e un’interfaccia utente intuitiva per eseguire operazioni di saldatura, smerigliatura e lucidatura di pinzette. I risultati mostrano che questo approccio facilita l’introduzione dei cobot nelle PMI. Tuttavia, sono sorti dubbi sull’utilizzo del robot collaborativo in forme più autenticamente collaborative, che potevano ridurre ulteriormente la necessità di sviluppi ad hoc e aumentare la flessibilità del sistema. Sulla base delle prime due attività, è stato proposto un framework per superare le limitazioni esistenti nelle applicazioni di collaborazione uomo-robot. Il framework si basa sul paradigma di un sistema produttivo mutualistico e adattivo. Paradigma che ha ispirato l’idea alla base di questa tesi di dottorato. Seguendo questi principi, il framework aiuta a considerare tutti i fattori critici che abilitano una piena collaborazione tra esseri umani e cobot. In particolare, il framework e le relative linee guida mirano a supportare professionisti e ricercatori nello sviluppo di applicazioni collaborative che migliorino l’efficienza, ma anche il benessere dei lavoratori. Il framework è stato applicato e validato nello sviluppo e nella gestione di un’applicazione di avvitatura in una PMI. I risultati mostrano un miglioramento delle prestazioni del processo e del benessere dei lavoratori. La digitalizzazione del sistema collaborativo e dei lavoratori che vi interagiscono è stata analizzata attraverso il concetto di Human Digital Twin. Inizialmente, è stato proposto un meta-modello per rappresentare i lavoratori umani nei sistemi produttivi. Questo meta-modello facilita la modellazione della rappresentazione digitale degli esseri umani, ma non supporta la creazione di istanze specifiche. Per evitare lo sviluppo di implementazioni ad hoc per ogni nuova replica digitale, che spesso risultano poco riutilizzabili ed estendibili, è stata proposta una piattaforma IIoT flessibile ed estensibile. Una delle implementazioni della piattaforma è stata testata sia in ambito industriale sia in laboratorio. L’introduzione di una collaborazione tra lavoratori umani e cobot ha un impatto sui fattori umani. Per questo motivo, è stata condotta una serie di esperimenti basati sulla Torre di Hanoi per integrare il framework. In questi esperimenti sono stati analizzati il carico mentale, la fiducia, la propensione alla fiducia e l’ansia, fornendo evidenze statistiche sulle dinamiche dei fattori umani selezionati in relazione a dimensioni, velocità e modalità collaborative del cobot. Inoltre, lo stesso setup è stato utilizzato per analizzare come l'anticipazione delle azioni del cobot influenzi le prestazioni del compito e la percezione umana. Infine, è stata sviluppata una nuova applicazione di avvitatura partendo dallo scenario industriale, per dimostrare che aumentare la consapevolezza situazionale e operativa dei robot collaborativi e abilitare azioni e decisioni simultanee e adattive può migliorare sia l’efficienza sia l’efficacia della collaborazione. In un secondo scenario, è stata proposta l’ottimizzazione del processo di avvitatura attraverso simulazione, ridondanza dell’angolo Z e l’algoritmo del problema del commesso viaggiatore, al fine di determinare la configurazione ottimale del robot e una sequenza che minimizzi il tempo ciclo. Questo metodo ha mostrato risultati rilevanti in termini di riduzione del tempo ciclo, sia in modalità automatica che collaborativa.
Achieving true human-robot collaboration: a methodological framework to solve challenges in industrial settings
Montini, Elias
2024/2025
Abstract
In public perception, collaborative robotics often evokes visions of user-friendly robots working alongside operators in effective and efficient industrial processes. However, this does not reflect the current industrial landscape. In reality, most collaborative robots are used as automation devices with minimal human collaboration. Although the promise of cobots lies in their ability to work with workers, significant technical and operational limitations often prevent this ideal from being realised. For example, issues such as limited adaptability to changing work environments and the complexity of integrating cobots into human workflows often reduce their use to mere automation tools. From an industrial perspective, this shift is understandable: companies are investing in cobots with the expectation of flexibility and collaboration. However, when faced with these limitations, they often use cobots for tasks similar to traditional automation to justify the financial investment. This discrepancy between perception and reality highlights a critical gap in the field of robotics: a gap that this thesis attempts to fill. This thesis proposes a methodological framework, along with some methods and technological solutions, to overcome the gap of adopting cobots as mere automation tools. This gap has been investigated through a survey and a new industrial application, which have confirmed a few benefits along with the limitations of cobots in full automation. The results from the survey have highlighted the limitations from which some key challenges for future research have been identified that might help researchers and practitioners to navigate the complex landscape of collaborative robotics. The limitations, as well as the partial benefits, experienced by manufacturing companies when adopting cobots as conventional automation were investigated through the development of a new industrial work cell in a small and medium enterprise (SME). To increase the flexibility of the application, which the company already conceived to work as fully automatic, part program parameterisation and a user-friendly interface have been combined to perform welding, grinding, and polishing tasks on tweezers. The results show that such a method facilitates the introduction of cobots in SMEs. However, doubts arose about using the collaborative robot differently in more collaborative forms, which would further reduce the need for ad hoc developments and increase their flexibility. Based on the first two activities, a methodological framework has been proposed to help overcome the existing limitations in human-robot collaboration applications. The framework is based on the paradigm of a mutualistic and adaptive production system, which inspired the idea for this PhD thesis. Based on these principles, the framework helps to consider all critical factors that enable full collaboration between humans and cobots. In particular, the framework, alongside its guidelines, aims to support practitioners and researchers in developing human-robot collaboration applications that increase efficiency, but also the well-being of workers by ensuring they feel comfortable and safe working with cobots and enabling greater flexibility. The framework has been applied and validated in the development and operation of a screwdriving application in a SME. The results show an improvement in process performance and worker well-being. The digitalisation of the collaborative system and the workers interacting with it were examined using the concept of the Human Digital Twin. First, a meta-model was proposed to represent human workers in production systems. This meta-model facilitates the modelling of the digital representation of humans. Still, it does not support the creation of specific instances, as demonstrated by the validation scenario in an industrial collaborative parts handling scenario. To avoid the development of ad-hoc implementations with each new digital replica that lacks reusability, scalability and extensibility, the work proposes an extensible and flexible IIoT platform. One implementation of the platform has been tested at both industrial and laboratory scale. The introduction of true collaboration between humans and robots impacts human factors. For this reason, a series of experiments based on the Tower of Hanoi were conducted to complement the framework. These experiments, in which mental workload, trust, propensity to trust and anxiety were investigated, provide statistical evidence of the dynamics of the selected human factors in relation to the size, speed and collaborative modes. In addition, the same setup was used to investigate how cobot's anticipatory actions affect task performance and human perception. Finally, a new screwdriving application was developed starting from the industrial scenario to show that increasing the situational and operational awareness of collaborative robots and enabling simultaneous and adaptive actions and decisions can increase both the efficiency and effectiveness of collaboration. In a second scenario, the optimisation of a screwdriving task through simulation, Z-angle redundancy and the algorithm of the travel salesman problem was proposed to determine the optimal configuration of the robot and a sequence that minimises the cycle time. This method shows relevant results regarding cycle time reduction in both automatic and collaborative modes.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_05_PhD_Montini.pdf
accessibile in internet per tutti a partire dal 11/05/2026
Descrizione: ACHIEVING TRUE HUMAN-ROBOT COLLABORATION: A METHODOLOGICAL FRAMEWORK TO SOLVE CHALLENGES IN INDUSTRIAL SETTINGS
Dimensione
112.32 MB
Formato
Adobe PDF
|
112.32 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/10589/238357