Towards Personalised Robotic Assessment and Response during Physical Human Robot Interactions
- Publication Type:
- Thesis
- Issue Date:
- 2022
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
The exponential growth of robotics in human environments have led to an explosion of human robot interactions. These interactions occur in proximity and have exposed the constraints and limitations of traditional models for robotic response which rely on task-centric measures. This has spurred on an area of research which focuses on understanding the capabilities and limitations of the human user during these interactions.
Humans are complex, autonomous agents that are difficult to model, and provide different categories of feedback that derive from biological systems. The current sensory paradigm requires an improved understanding of the limitations, the development of blended-measure models that employ human-centric measures, and a contextually connected biological human understanding into robotic frameworks.
This thesis presents a framework towards personalised robotic assessment and response with considerations on understanding the human user during physical human robot interactions. The framework approaches this by examining current limitations, enabling personalised models from human-centric measures, and enhancing the understanding of the human user through physiological and musculoskeletal models.
The implementation of a robotic system highlights the feasibility and limitations of using task-centric models during Physical Human Robot Interactions (pHRI). Further work investigates inertial effects of the user during interactions in the context of a prominent predictive model, Fitts’ Law.
Physical Human Robot Interaction Primitives extends upon Interaction Primitives by incorporating physical interaction forces between the human user and robot, enabling the inference of user intent when generating a personalised robotic response.
Finally, the enhancement of the link between biological human understanding and robotic frameworks is explored. A validation process for a popular musculoskeletal model is conducted, comparing computational results with experimental readings. The limitations for the complex model led to the generation of an empirical model correlating forearm muscle activity and grip strength. This physiological model captured co-contractions for antagonistic muscle pairs and supplemented motion analysis for the musculoskeletal model, enhancing the computational results.
The framework combines the topics which facilitate intuitive and adaptive human-robot interactions. The advancement of such collaborative intelligence enhances complementary strengths between human and robot, and work hand in end-effector towards a safer, more interactive future.
Please use this identifier to cite or link to this item: