Trajectories and keyframes for kinesthetic teaching: A human-robot interaction perspective

By Barış Akgün, Maya Cakmak, J. Yoo, Andrea Lockerd Thomaz. 2012

294 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering, Human-Computer Interaction, Robotics & Control

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Control and Systems Engineering

In Human-Computer Interaction

In Robotics & Control

Citations in the corpus

One-Shot Imitation from Observing Humans via Domain-Adaptive Meta-Learning (2018, background): “Most imitation learning and learning from demonstration methods operate at the level of configuration-space trajectories [44, 2], which are typically collected using kinesthetic teaching [32, 1], teleoperation [8, 35, 62], or sensors on the”
Challenges in Shared-Environment Human-Robot Collaboration (2013, background): “Members of the HRI community have performed user studies on various LfD techniques, and continue to present results valuable to interaction designers [4]–[7].”
Autonomously constructing hierarchical task networks for planning and human-robot collaboration (2016, background): “LfD is particularly useful due to its vast applicability, as demonstrations can be effectively provided via teleoperation [8], observation [9], and kinesthetic teaching [10] to achieve task or skill proficiency.”
Incremental Semantically Grounded Learning from Demonstration (2013, methodology): “Current LfD methods take a variety of approaches to sequencing, ranging from associative skill memories [16], in which all primitives are made available for selection at each decision point, to other schemes in which primitives are executed in a fixed order [1, 13].”
Improved human–robot team performance through cross-training, an approach inspired by human team training practices (2015, methodology): “More recently, Gaussian mixture models (Akgun et al., 2012) have been used to teach a skill to a robot during experiments in which a human physically guides the robot through a trajectory; this approach is known as “kinesthetic teaching.””
Using dVRK teleoperation to facilitate deep learning of automation tasks for an industrial robot (2017, background): “Kinesthetic teaching has been used to give demonstrations for stacking towers [30], scooping melons [31], and writing on a whiteboard [32] among other tasks, and has recently become popular in LfD due to its intuitive interface for directly controlling joint angles and promising experimental…”
An Overview of Machine Teaching (2018, background): “This, in turn, can be useful in two ways: (i) insights gained here can be used to teach human users on how to optimally interact with and provide instructions to robots (e.g. for personalized robotic devices like Siri or Alexa) [1, 37,43], (ii) we could develop smarter learning algorithms for robots”
PEBBLE: Feedback-Efficient Interactive Reinforcement Learning via Relabeling Experience and Unsupervised Pre-training (2021, background): “While imitation learning is a powerful tool, suitable demonstrations are often prohibitively expensive to obtain in practice (Calinon et al., 2009; Pastor et al., 2011; Akgun et al., 2012; Zhang et al., 2018).”
Learning grounded finite-state representations from unstructured demonstrations (2015, background): “…take a variety of approaches to sequencing, ranging from associative skill memories [Pastor et al., 2012], in which all primitives are made available for selection at each decision point, to other schemes in which primitives are executed in a fixed order [Akgun et al., 2012; Niekum et al., 2012].”
Novel Interaction Strategies for Learning from Teleoperation (2012, methodology): “We presented a method that can learn from hybrid demonstrations in (Akgun et al. 2012a).”
AVID: Learning Multi-Stage Tasks via Pixel-Level Translation of Human Videos (2019, background): “The means by which these demonstrations can be obtained are often laborious and limited in applicability (Akgun et al., 2012; Calinon et al., 2009; Pastor et al., 2011; Zhang et al., 2018).”
Generative predecessor models for sample-efficient imitation learning (2019, background): “…partial demonstrations is a useful benchmark for the effectiveness of imitation learning methods but can also provide a convenient way of providing demonstrations and can free the agent to ﬁnd more optimal trajectories between states (see for example Akgun et al. (2012); Schroecker et al. (2016)).”