A Human-like Upper-limb Motion Planner: Generating naturalistic movements for humanoid robots

Gulletta, Gianpaolo; Silva, Eliana Costa e; Erlhagen, Wolfram; Meulenbroek, Ruud G. J.; Costa, M. Fernanda P.; Bicho, Estela

doi:10.1177/1729881421998585

Cited by 20 publications

(17 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This approach fosters AI-generated movements that are not just mechanically efficient but also safe, adaptable, and truly human-like. It is especially crucial in contexts where AI systems interact with humans, such as in assistive robotics or virtual human simulations (55).…”

Section: The Reserve Of Joint Torque To Improve Effort Predictionmentioning

confidence: 99%

Force Reserve Predicts Compensation in Reaching Movement with Induced Shoulder Strength Deficit

Faity,

Barradas,

Schweighofer

et al. 2024

Preprint

View full text Add to dashboard Cite

Following events such as fatigue or stroke, individuals often move their trunks forward during reaching, leveraging a broader muscle group even when only arm movement would suffice. In previous work, we showed the existence of a 'force reserve' - a phenomenon where individuals, when challenged with a heavy weight, adjusted their motor coordination to preserve approximately 40% of their shoulder's force. Here, we investigated if such reserve can predict hip, shoulder, and elbow movements and torques resulting from an induced shoulder strength deficit. We engaged 20 healthy participants in a reaching task with incrementally heavier dumbbells, analyzing arm and trunk movements via motion capture and joint torques through inverse dynamics. We simulated these movements using an optimal control model of a 3-degree-of-freedom upper body, contrasting three cost functions: traditional sum of squared torques, a force reserve function incorporating a nonlinear penalty, and a normalized torque function. Our results demonstrate a clear increase in trunk movement correlated with heavier dumbbell weights, with participants employing compensatory movements to maintain a shoulder force reserve of approximately 40% of maximum torque. Simulations showed that while traditional and reserve functions accurately predicted trunk compensation, only the reserve function effectively predicted joint torques under heavier weights. These findings suggest that compensatory movements are strategically employed to minimize shoulder effort and distribute load across multiple joints in response to weakness. We discuss the implications of the force reserve cost function in the context of optimal control of human movements and its relevance for understanding of compensatory movements post-stroke.

show abstract

Section: The Reserve Of Joint Torque To Improve Effort Predictionmentioning

confidence: 99%

Force Reserve Predicts Compensation in Reaching Movement with Induced Shoulder Strength Deficit

Faity,

Barradas,

Schweighofer

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Looking at recent works, Guletta et al proposed HUMP [18], a novel motion planner that generates human-like upper-limb movement, He et al [19] solved an inverse kinematics problem to create legible motion, Gabert et al [20] used a sampling-based motion planner to create obstacle avoidant yet legible trajectories, and Miura et al extended the idea of legibility to stochastic environments [21]. Looking at legibility more generally, the idea has also been explored in high-level task planning [22] and other discrete domains [12], [23], [24], [25].…”

Section: A Legibilitymentioning

confidence: 99%

SLOT-V: Supervised Learning of Observer Models for Legible Robot Motion Planning in Manipulation

Wallkotter

Chétouani

Castellano

2022

2022 31st IEEE International Conference on Robot and Human Interactive Communication (RO-MAN)

View full text Add to dashboard Cite

We present SLOT-V, a novel supervised learning framework that learns observer models (human preferences) from robot motion trajectories in a legibility context. Legibility measures how easily a (human) observer can infer the robot's goal from a robot motion trajectory. When generating such trajectories, existing planners often rely on an observer model that estimates the quality of trajectory candidates. These observer models are frequently hand-crafted or, occasionally, learned from demonstrations. Here, we propose to learn them in a supervised manner using the same data format that is frequently used during the evaluation of aforementioned approaches. We then demonstrate the generality of SLOT-V using a Franka Emika in a simulated manipulation environment. For this, we show that it can learn to closely predict various hand-crafted observer models, i.e., that SLOT-V's hypothesis space encompasses existing handcrafted models. Next, we showcase SLOT-V's ability to generalize by showing that a trained model continues to perform well in environments with unseen goal configurations and/or goal counts. Finally, we benchmark SLOT-V's sample efficiency (and performance) against an existing IRL approach and show that SLOT-V learns better observer models with less data. Combined, these results suggest that SLOT-V can learn viable observer models. Better observer models imply more legible trajectories, which mayin turn -lead to better and more transparent human-robot interaction.

show abstract

“…The development of anthropomorphic robotic arms has received more and more attention in the robot industry in recent years [ 1 ]. By analyzing the human arm skeleton, joint movements and muscle forces in depth, the anthropomorphic robotic arm aims to develop a human-like mechanical structure, motion planning and control theory [ 2 , 3 ]. The practice has proved that anthropomorphic robotic arms can assist or even replace humans to complete auxiliary and heavy tasks in multiple fields, such as industry and biomedical and social services [ 4 , 5 ].…”

Section: Introductionmentioning

confidence: 99%

Sensor Fusion-Based Teleoperation Control of Anthropomorphic Robotic Arm

et al. 2023

View full text Add to dashboard Cite

Sensor fusion is a technique that combines information from multiple sensors in order to improve the accuracy and reliability of the data being collected. In the context of teleoperation control of an anthropomorphic robotic arm, sensor fusion technology can be used to enhance the precise control of anthropomorphic robotic arms by combining data from multiple sensors, such as cameras, data gloves, force sensors, etc. By fusing and processing this sensing information, it can enable real-time control of anthropomorphic robotic arms and dexterous hands, replicating the motion of human manipulators. In this paper, we present a sensor fusion-based teleoperation control system for the anthropomorphic robotic arm and dexterous hand, which utilizes a filter to fuse data from multiple sensors in real-time. As such, the real-time perceived human arms motion posture information is analyzed and processed, and wireless communication is used to intelligently and flexibly control the anthropomorphic robotic arm and dexterous hand. Finally, the user is able to manage the anthropomorphic operation function in a stable and reliable manner. We also discussed the implementation and experimental evaluation of the system, showing that it is able to achieve improved performance and stability compared to traditional teleoperation control methods.

show abstract

A Human-like Upper-limb Motion Planner: Generating naturalistic movements for humanoid robots

Cited by 20 publications

References 61 publications

Force Reserve Predicts Compensation in Reaching Movement with Induced Shoulder Strength Deficit

Force Reserve Predicts Compensation in Reaching Movement with Induced Shoulder Strength Deficit

SLOT-V: Supervised Learning of Observer Models for Legible Robot Motion Planning in Manipulation

Sensor Fusion-Based Teleoperation Control of Anthropomorphic Robotic Arm

Contact Info

Product

Resources

About