TY - GEN
T1 - Levels of Explanation for Error Resolution in HRI
AU - Krakovski, Maya
AU - Kumar, Shikhar
AU - Edan, Yael
N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2026.
PY - 2026/1/1
Y1 - 2026/1/1
N2 - This research investigated the influence of different levels of explanation (LoE) for resolving errors in human-robot interaction. We compared different LoEs for different error types in two user studies involving different robot types (manipulator and humanoid) performing different tasks (sorting and physical training). The LoEs were opera-tionalized using two questions: “what” (verbosity) and “why” (justifica-tion) and combined to define four LoEs. Each robotic task was implemented with two LoEs, and different LoEs were compared for each task in a between-study design. The study, which included both younger and older adults, found that adding justification to verbosity did not significantly improve user perception or task performance in either robot/ task. Furthermore, the results showed that across both tasks, varying the LoEs did not have a significant effect on error resolution. Notably, most participants successfully resolved errors with the physical training robot, but error resolution was lower for the sorting task.
AB - This research investigated the influence of different levels of explanation (LoE) for resolving errors in human-robot interaction. We compared different LoEs for different error types in two user studies involving different robot types (manipulator and humanoid) performing different tasks (sorting and physical training). The LoEs were opera-tionalized using two questions: “what” (verbosity) and “why” (justifica-tion) and combined to define four LoEs. Each robotic task was implemented with two LoEs, and different LoEs were compared for each task in a between-study design. The study, which included both younger and older adults, found that adding justification to verbosity did not significantly improve user perception or task performance in either robot/ task. Furthermore, the results showed that across both tasks, varying the LoEs did not have a significant effect on error resolution. Notably, most participants successfully resolved errors with the physical training robot, but error resolution was lower for the sorting task.
KW - Levels of explanation
KW - justification
KW - understandability
KW - verbosity
UR - https://www.scopus.com/pages/publications/105046967160
U2 - 10.1007/978-981-95-2398-6_36
DO - 10.1007/978-981-95-2398-6_36
M3 - Conference contribution
SN - 9789819523979
T3 - Lecture Notes in Computer Science
SP - 544
EP - 557
BT - Social Robotics + AI - 17th International Conference, ICSR+AI 2025, Proceedings, Part 3
A2 - Staffa, Mariacarla
A2 - Cabibihan, John-John
A2 - Siciliano, Bruno
A2 - Rossi, Silvia
A2 - Sam Ge, Shuzhi
A2 - Bodenhagen, Leon
A2 - Tapus, Adriana
A2 - Cavallo, Filippo
A2 - Fiorini, Laura
A2 - Matarese, Marco
A2 - He, Hongsheng
PB - Springer Science and Business Media Deutschland GmbH
T2 - 17th International Conference on Social Robotics, ICSR+AI 2025
Y2 - 10 September 2025 through 12 September 2025
ER -