Lightweight, adjustable, and affordable devices are needed to enable the next generation of effective, wearable adjuncts for rehabilitation. Used at home or in a rehabilitation setting, these devices have the potential to reduce compound pressures on hospitals and social care systems. Despite recent developments in soft wearable robots, many of these devices restrict the range of motion and lack quantitative assessment of moment transfer to the wearer. The decoupled design of our wearable device for upper-limb rehabilitation successfully delivers almost the full range of motion to the user, with a mean maximum flexion angle of 149° (SD = 8.5). In this article, for the first time, we show that in tests involving a wide range of participants, 82% of the moment produced by the actuator is applied to the wearer. This testing of elbow flexion moment transfer supports the effectiveness of the device. This research is a step toward effective pneumatic soft robotic wearable devices that are adaptable to a wide range of users – a necessary prerequisite for their widespread adoption in health care.
28 February 2025 in Journal of Intelligent & Robotic Systems
Mobile robots are becoming increasingly popular in a wide array of applications: industrial, item delivery, search and rescue, space, social, and entertainment. A two-wheeled self-balancing mobile robot is a statically unstable non-linear system with strong coupling dynamics. Common practices in the development of control systems for such robots are either to linearise the region of application to be used with linear controllers or to use complex nonlinear controllers such as fuzzy logic, sliding mode, and neural networks. However, self-balancing robots are still restricted by the travelling distance needed to regain an upright stance, the length of settling time, high overshoot and lack of resilience to external disturbance. In this paper, we are proposing a novel genetic algorithm-based switching control to evolve more effective control parameters and increase autonomy. Differently from previous work a genetic algorithm has been used to select the parameters in a sliding mode control and a switching-algorithm-based controller of a two-wheeled self-balancing mobile robot. The performance of the proposed controllers is assessed in simulations using the CoppeliaSim environment. The tests used dynamic criteria (distance travelled, maximum angular deviation), control criteria (settling time, % overshoot). The results showed that the genetic algorithm-based control has better performance in the 55 degree recovery, impulse response and variable inclination tests and that switching algorithm-based control shows better performance in step response tests. The results produced by the evolutionary algorithm are often able to perform better than their analytic counterparts. This shows the potential of meta-heuristic algorithms to obtain solutions for optimization problems encountered by statically unstable non-linear systems in unstructured and fast-changing environments.
17 February 2024 in International Journal of Social Robotics
For efficient human–robot interaction, human operators need to be able to efficiently represent the robot’s movements in space and predict its next steps. However, according to frameworks of Bayesian multisensory integration, features outside the motion itself—like the sounds a robot makes while it moves—should affect how otherwise identical motions are perceived. Here, we translate an established psychophysical task from experimental psychology to a human–robot interaction context, which can measure these distortions to motion perception. In two series of preregistered studies, participants watched a humanoid robot make forward and backward reaching movements. When the robot hand suddenly disappeared, they reported its last seen location, either with the mouse cursor (Experiment 1a and 1b) or by matching it to probe stimuli in different locations (Experiment 2a and 2b). The results revealed that even small changes to the robot’s sound robustly affect participants’ visuospatial representation of its motions, so that the motion appeared to extend further in space when accompanied by slightly (100 ms) longer sounds compared to slightly shorter sounds (100 ms shorter). Moreover, these sound changes do not only affect where people currently locate the robot’s motion, but where they anticipate its future steps. These findings show that sound design is an effective medium for manipulating how people represent otherwise identical robot actions and coordinate its interactions with it. The study acts as proof of concept that psychophysical tasks provide a promising tool to measure how design parameters influence the perception and prediction of robot motion.
Robotics looks to nature for inspiration to perform effectively in unstructured environments and can be used as a platform to test biological hypotheses. Social animals often share information about food source locations: one example is tandem running in ants, where a leader guides a naive recruit to a known profitable food site. This is extremely advantageous as it allows the sharing of important information among colony members, but it also has costs, such as waiting time inside the nest when no leaders are around, and a reduced walking speed for the tandem couple compared to individual ants. Whether and when these costs outweigh the benefits is not well understood as it is challenging to observe complex social behaviours in nature. We developed two kilobot-based approaches to compare tandem running and lone scout foraging, where an ant searches for food without any previous knowledge of its location: one approach based on real-life experiments and one on computer simulations. We investigated the role that the size of the search arena played in the effectiveness of foraging. Tandem pairs were faster for all three arena sizes; however, this result was reversed in the simulations. These results highlight the inconsistencies between simulation and real-life kilobot experiments, previously reported for other systems and known as reality gap. Further testing is needed to inform on whether robotic applications should utilise agents with the same roles and capabilities for search, detection and repair-type tasks as simulations, or whether instead the two approaches should be treated separately.
10 April 2023 in IEEE Transactions on Medical Robotics and Bionics
Pathologies affecting the respiratory system can lead to a debilitating decrease in quality of life and can be fatal. To test medical devices and implants for the human respiratory system, a simulation system that can reproduce multiple respiratory features is necessary. Currently available respiratory simulators only focus on reproducing flow rate profiles of breathing while coughing simulators focus on aerosol analysis. In this paper we propose a novel, bioinspired robotic simulator that can physically replicate both breathing and coughing flow rate characteristics of healthy adults. We conducted a study on 31 healthy adult participants to gather the flow rate measurement of normal breathing, deep breathing, breathing while running and coughing. Coughing flow rate profiles vary considerably between participants, making an accurate simulation of coughs a challenge. To enable cough flow rate simulation, a new methodology based on the identification of four cough phases, Attack, Decay, Sustain and Release (ADSR) and their parametrization was devised. This methodology leads to the unprecedented ability to reproduce diverse and complex coughing flow rate profiles. Our simulator is able to reproduce respiratory flows with a root mean square error (RMSE) of 1.8 L/min between normal participant breathing and its simulation, 5% of the maximum flow rate simulated for that participant (pMFR), an RMSE of 10.08 L/min for deep breathing, 18% of the pMFR and an RMSE of 13.29 L/min for exertion breathing, 17% of pMFR. For the simulation of an average cough we recorded an RMSE of 51.43 L/min, 13% of the pMFR and for a low flow rate cough an RMSE of 12.38 L/min, 9.5% of the pMFR. The presented simulator matches the fundamentals of human breathing and coughing, advancing the current capability of respiratory system simulators.
Visual Perspective Taking (VPT) underpins human social interaction, from joint action to predicting others' future actions and mentalizing about their goals and affective/mental states. Substantial progress has been made in developing artificial VPT capabilities in robots. However, as conventional VPT tasks rely on the (non-situated, disembodied) presentation of robots on computer screens, it is unclear how a robot's socially reactive and goal-directed behaviours prompt people to take its perspective. We provide a novel experimental paradigm that robustly measures the extent to which human interaction partners take a robot's visual perspective during face-to-face human-robot-interactions, by measuring how much a robot's visual perspective is spontaneously integrated with one's own. The experimental task design of our upcoming user study allows us to investigate the role of robot features beyond its human-like appearance, which have driven research so far, targeting instead its socially reactive behaviour and task engagement with the human interaction partner.