A Preliminary Study of Peer-to-Peer Human-Robot Interaction

Fong, Terrence; Scholtz, Jean; Shah, Julie A.; Flückiger, Lorenzo; Kunz, C.; Lees, David; Schreiner, John; Siegel, M.; Hiatt, Laura M.; Nourbakhsh, Illah; Simmons, Reid; Ambrose, Robert O.; Burridge, Robert R.; Antonishek, Brian; Bugajska, Magdalena D.; Schultz, Alan C.; Trafton, J. Gregory

doi:10.1109/icsmc.2006.384609

Cited by 26 publications

(17 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Accordingly, the study of human-robot interaction has seen rapid growth [29,34,39,47,64,85,92]. It is reasonable to assume that non-experts will increasingly interact with robotic systems and will have an idea of how the system should act.…”

Section: Motivation: Learning From Demonstration In Autonomymentioning

confidence: 99%

Bayesian Nonparametric Reward Learning From Demonstration

et al. 2015

View full text Add to dashboard Cite

Learning from demonstration provides an attractive solution to the problem of teaching autonomous systems how to perform complex tasks. Demonstration opens autonomy development to non-experts and is an intuitive means of communication for humans, who naturally use demonstration to teach others. This thesis focuses on a specific form of learning from demonstration, namely inverse reinforcement learning, whereby the reward of the demonstrator is inferred. Formally, inverse reinforcement learning (IRL) is the task of learning the reward function of a Markov Decision Process (MDP) given knowledge of the transition function and a set of observed demonstrations. While reward learning is a promising method of inferring a rich and transferable representation of the demonstrator's intents, current algorithms suffer from intractability and inefficiency in large, real-world domains. This thesis presents a reward learning framework that infers multiple reward functions from a single, unsegmented demonstration, provides several key approximations which enable scalability to large real-world domains, and generalizes to fully continuous demonstration domains without the need for discretization of the state space, all of which are not handled by previous methods.In the thesis, modifications are proposed to an existing Bayesian IRL algorithm to improve its efficiency and tractability in situations where the state space is large and the demonstrations span only a small portion of it. A modified algorithm is presented and simulation results show substantially faster convergence while maintaining the solution quality of the original method. Even with the proposed efficiency improvements, a key limitation of Bayesian IRL (and most current IRL methods) is the assumption that the demonstrator is maximizing a single reward function. This presents problems when dealing with unsegmented demonstrations containing multiple distinct tasks, common in robot learning from demonstration (e.g. in large tasks that may require multiple subtasks to complete). A key contribution of this thesis is the development of a method that learns multiple reward functions from a single demonstration. The proposed method, termed Bayesian nonparametric inverse reinforcement learning (BNIRL), uses a Bayesian nonparametric mixture model 3 to automatically partition the data and find a set of simple reward functions corresponding to each partition. The simple rewards are interpreted intuitively as subgoals, which can be used to predict actions or analyze which states are important to the demonstrator. Simulation results demonstrate the ability of BNIRL to handle cyclic tasks that break existing algorithms due to the existence of multiple subgoal rewards in the demonstration. The BNIRL algorithm is easily parallelized, and several approximations to the demonstrator likelihood function are offered to further improve computational tractability in large domains.Since BNIRL is only applicable to discrete domains, the Bayesian nonparametric reward learning framew...

show abstract

Section: Motivation: Learning From Demonstration In Autonomymentioning

confidence: 99%

Bayesian Nonparametric Reward Learning From Demonstration

et al. 2015

View full text Add to dashboard Cite

show abstract

“…In this setup, one robot has the role of welding, another robot has the role of quality inspection of the weld that is produced [33]. Two astronauts have various roles inside the workflow, ranging from relocating robots to checking the quality of results.…”

Section: J Fong Etalmentioning

confidence: 99%

Characterizing the State of the Art of Human-Robot Coproduction

2015

View full text Add to dashboard Cite

Abstract-It is expected that soon, systems consisting of a blend of humans and robots be devised in such a way that higher productivities will be achieved. The main enabler for this is expected to be the possibility of collaboration between workers and robots. HRI (Human Robot Interaction) is the field in which such phenomena are studied. A growing number of investigators treat the collaboration of robots and workers (humans) in many contexts, however attention towards the manufacturing industry is predominantly focused on full automation of human tasks. Industrial robots have long been unsafe to work in close vicinity to workers due to their duty to be fast and powerful. However, nowadays, with the drive from emerging technologies, this is changing. Safe worker-robot collaborations are beginning to take shape and the HRI community is beginning to study such scenarios. Despite being a very effective form of interaction, a key research question is whether collaboration is a suitable mode of interaction for manufacturing environments. To be able to address this question, we found a collection of ten workerrobot systems that constitute a first step in outlining coproduction characteristics. This collection allowed us to identify differences in task initiative and product handling and component handling, while we frame coproduction as an extension of man. Challenges that require additional attention are workflow planning and defining proper performance indicators. We conclude with the fact that, although the worker-robot collaboration systems are inspiring and redefine labor, no sufficient knowledge or tools exists to reproduce such qualities in different manufacturing settings. Further work will be focused on modeling and assessing the performance and bottlenecks of systems based on novel robotic systems.

show abstract

“…Previous studies in collaborative control have explored such cooperation between an operator and an automatic error detection system [4,5]. Kaupp et al have also recently done research in this direction [15], using robot-initiated notifications of error states to reduce teleoperation workload for navigational tasks.…”

Section: Error Detectionmentioning

confidence: 99%

“…Fong et al [4] proposed the term "Mean Time Between Interventions", describing the "mean time a human-robot system operates nominally (i.e. outside of intervention)".…”

Section: Mean Time Between Failures (Mtbf)mentioning

confidence: 99%

Field trial for simultaneous teleoperation of mobile social robots

Glas¹,

Kanda²,

Ishiguro

et al. 2009

Proceedings of the 4th ACM/IEEE International Conference on Human Robot Interaction

View full text Add to dashboard Cite

Simultaneous teleoperation of mobile, social robots presents unique challenges, combining the real-time demands of conversation with the prioritized scheduling of navigational tasks. We have developed a system in which a single operator can effectively control four mobile robots performing both conversation and navigation. We compare the teleoperation requirements for mobile, social robots with those of traditional robot systems, and we identify metrics for evaluating task difficulty and operator performance for teleoperation of mobile social robots. As a proof of concept, we present an integrated priority model combining real-time conversational demands and non-real-time navigational demands for operator attention, and in a pioneering study, we apply the model and metrics in a demonstration of our multi-robot system through real-world field trials in a shopping arcade.

show abstract

A Preliminary Study of Peer-to-Peer Human-Robot Interaction

Cited by 26 publications

References 14 publications

Bayesian Nonparametric Reward Learning From Demonstration

Bayesian Nonparametric Reward Learning From Demonstration

Characterizing the State of the Art of Human-Robot Coproduction

Field trial for simultaneous teleoperation of mobile social robots

Contact Info

Product

Resources

About