Building robot systems to interact with people in real environments

Abstract

Interacting with people poses a unique challenge for robots. While issues such as safety are well known, issues such as appearance and intuitive human-robot interfaces are also critical to success. The American Association for Artificial Intelligence Robot Competition has encouraged the development of robots that interact with people through competitions, challenges, and exhibitions. This paper describes the robot systems developed for eleven different events over eight years from 1998-2005 and examines the factors that led to both successes and failures. The primary lesson is that simplicity and robustness in sensing, behavior, and humanrobot interactions helps to counteract the uncertainty caused by people.

Social Momentum: Design and Evaluation of a Framework for Socially Competent Robot Navigation

“…These robots successfully served thousands of tour guide requests and navigated alongside museum visitors over a span of weeks. Their performance set the stage for real-world challenges, such as the AAAI Mobile Robot Challenge, organized by the American Association for Artificial Intelligence, which attracted wide interest and participation [54,56]. Although these deployments featured remarkable system development and integration efforts, their underlying navigation frameworks [24,57] treated humans as non-reactive obstacles without explicitly mod-eling interaction phenomena or humans' decision-making strategies.…”

Core Challenges of Social Robot Navigation: A Survey

“…The demonstrated capabilities of these systems motivated additional efforts in the tour-guide domain such as Robox [204], Mobot [166], Rackham [38], and CiceRobot [29]. Finally, progress in the field was incentivized through competitions such as the AAAI Mobile Robot Challenge, organized by the American Association for Artificial Intelligence (AAAI) [153,157].…”

Aprendizaje de y con robótica, algunas experiencias

“…En una propuesta para el aprendizaje de la robótica, el determinar cómo se aprende concierne a la metodología de trabajo en clase: planear el conjunto de actividades, medios y recursos a utilizar, determinar el tiempo necesario para el desarrollo de cada actividad, determinar los objetivos de cada una de las áreas involucradas y el tema a trabajar, las estrategias más utilizadas para definir la metodología, en este caso, son el aprendizaje colaborativo (Barkley, Cross y Howell, 2007;Plestina, Turic y Papic, 2007;Ahlgren, 2002), aprendizaje lúdico (Jiménez, 2005;Maxwel, 2007;Minguez, Javier, et al, 2005;Murphy, 2001) y el aprendizaje basado en problemas (Escribano y del Valle, 2008;Narváez S. Carlos H. Leo, et al, 2006). Estas tres estrategias pueden ser combinadas de acuerdo a las competencias que se quieran promover, los contenidos que se vayan a desarrollar y los recursos que disponga la institución educativa.…”

“…Maxwell (Maxwel, 2007) presenta su experiencia en la construcción de robots con la capacidad de interactuar con personas en entornos reales; este tipo de robots son desarrollados para participar en la exhibición y competencias robóticas de la Asociación Americana para la Inteligencia Artificial. La experiencia muestra la importancia de planificar el trabajo para alcanzar el objetivo final: la construcción de un robot robusto y el aprendizaje significativo que logran los estudiantes.…”