Shaping Physical Interactions Between Humans and Robotic Manipulators
- Publication Type:
- Thesis
- Issue Date:
- 2023
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
Manipulators are becoming increasingly ubiquitous in human-centric environments, with some even operating in direct physical contact with humans. While consensus exists on the control strategies used for robots during physical Human–Robot Interaction (pHRI), the current landscape predominantly revolves around performance-based metrics, often neglecting the subtle nuances of physical interactions.
Recognising this gap, this thesis uses simulations and experiments to shape the development of control strategies. Since subtle variations in the force applied during physical contact can convey different meanings, the strategies developed enable a manipulator to modify its response to engage in purposeful high-level actions or reflex-like low-level responses.
A common approach to influencing high-level robot behaviour for pHRI is through the use of model-based metrics. Fitts law, an information-theoretic model, can predict human speed and accuracy based on target-directed reaching tasks. Using this model, the study found that humans demonstrate typical motions even when interacting with variable inertia interfaces.
It is recognised that interface inertia can impact the energy expended by humans during physical interaction. Given this understanding, redundancy resolution is leveraged to facilitate a Cartesian inertia-based optimisation strategy that minimises energy expenditure during interactions. Real-world experiments undertaken demonstrate a 25% reduction in expended kinetic energy.
Conventionally, redundancy resolution relies on kinematics and performance-based metrics. An alternative method applicable for pHRI utilises musculoskeletal models for evaluating human strength. However, it is not clear from the literature which of these methods is best suited for modelling and representing strength. Simulations detail the differences and advantages of the existing methods, discussing the appropriateness of each method for particular applications.
For addressing safe and efficient pHRI in unstructured environments, an oobleck-inspired control strategy that facilitates low-level reflex-like robot behaviour is proposed. An experimental study is conducted to elucidate the behaviours of oobleck during robot interaction, forming the basis of a parameter adaptation strategy. Observations demonstrate that the fluid permits compliant motion from external stimuli but can render the object of manipulation static when responding to undesired behaviour.
This thesis presents an examination of models to influence control strategies for pHRI. A control strategy exploiting redundancy is used to demonstrate a robot’s ability to exhibit purposeful responses, while the observations from the characteristics of a fluid are used to develop a reflex-like response. Insights from these techniques are leveraged to shape a manipulator’s physical feeling when interacting with humans. The synergy of these strategies is envisioned to enhance the interaction experience during pHRI.
Please use this identifier to cite or link to this item: