Interactive Tuning of Robot Program Parameters via Expected Divergence Maximization

Abstract

Enabling diverse users to program robots for different applications is critical for robots to be widely adopted. Most of the new collaborative robot manipulators come with intuitive programming interfaces that allow novice users to compose robot programs and tune their parameters. However, parameters like motion speeds or exerted forces cannot be easily demonstrated and often require manual tuning, resulting in a tedious trial-and-error process. To address this problem, we formulate tuning of one-dimensional parameters as an Active Learning problem where the learner iteratively refines its estimate of the feasible range of parameter values, by selecting informative queries. By executing the parametrized actions, the learner gathers the user's feedback, in the form of directional answers ("higher, " "lower, " or "fine"), and integrates it in the estimate. We propose an Active Learning approach based on Expected Divergence Maximization for this setting and compare it against two baselines with synthetic data. We further compare the approaches on a real-robot dataset obtained from programs written with a simple Domain-Specific Language for a robot arm and manually tuned by expert users (N=8) to perform four manipulation tasks. We evaluate the effectiveness and usability of our interactive tuning approach against manual tuning with a user study where novice users (N=8) tuned parameters of a human-robot hand-over program. CCS CONCEPTS• Computing methodologies → Active learning settings; • Human-centered computing → User centered design; • Computer systems organization → External interfaces for robotics.

Non-Dyadic Interaction: A Literature Review of 15 Years of Human-Robot Interaction Conference Publications

“…The typical context for migration is the programming of robots (e.g. [68,134,173]), see Figure 7(b). Stenmark et al [173], for example, developed an interface for robot programming for non-expert users.…”

“…Through RFID, the robot builds a history of interaction for each user, and adapts its behaviour through personalisation. (b) Illustration of the continuity theme from Racca et al[134]. The study uses the principle of migration to adjust the parameters of a robot on one artefact (the PC) and migrate the settings to the robot arm executing the migrated command.…”

Helping People Through Space and Time: Assistance as a Perspective on Human-Robot Interaction

“…A robot working collaboratively with a user can improve its efficiency by modeling the user's behavior, for example by determining specific poses to hold an object in to facilitate fluid collaboration during assembly (Akkaladevi et al, 2016) or by anticipating and delivering the next required item in assembly (Hawkins et al, 2013(Hawkins et al, , 2014Maeda et al, 2014) or cooking (Koppula et al, 2016;Milliez et al, 2016), or by providing help under different initiative paradigms during assembly (Baraglia et al, 2016). Collaborative environmental assistance can also be used to perform joint actions with a user, such as in handovers (Cakmak et al, 2011;Kwon and Suh, 2012;Grigore et al, 2013;Broehl et al, 2016;Canal et al, 2018;Cserteg et al, 2018;Goldau et al, 2019;Lambrecht and Nimpsch, 2019;Nemlekar et al, 2019;Newman et al, 2020;Racca et al, 2020), where the goal is to transfer an object from the robot's end effector to the user's hand; or comanipulation (Koustoumpardis et al, 2016;Nikolaidis et al, 2016;El Makrini et al, 2017;Goeruer et al, 2018;Rahman, 2019b;DelPreto and Rus, 2019;Rahman, 2020;Wang et al, 2020), where the aim is for the user and the robot to jointly move an object to a specified location or provide redundancy in holding an object in a joint assembly task (Parlitz et al, 2008) or safety critical situation such as surgery (Su et al, 2018).…”

A Survey on End-User Robot Programming

“…In addition to enabling specification of the overall logic and flow of a program, some systems may enable end-users to specify the parameters for commands or functions (e.g., [53,85,117]). In order to ease the process of specifying continuous parameters or parameters in 3-D space, systems may provide intuitive methods for parametrization, such as directional (e.g., [95]) or gesture-based specification (e.g., [86]).…”