The Personal Rover Project The Comprehensive Design of a Domestic Personal Robot Emily Falcone, Rachel Gockley, Eric Porter, Illah Nourbakhsh The Robotics Institute, Carnegie Mellon University, Pittsburgh, PA 15213 Abstract In this paper, we summarize an approach for the dissemination of robotics technologies. In a manner analogous to the personal computer movement of the early 1980’s, we propose that a productive niche for robotic technologies is as a long-term creative outlet for human expression and discovery. To this end, this paper describes our ongoing efforts to design, prototype and test a low-cost, highly competent personal rover for the domestic environment. Introduction Robotics occupies a special place in the arena of interactive technologies because it combines sophisticated computation with rich sensory input in a physical embodiment that can exhibit tangible and expressive behavior in the physical world. In this regard, a central question that occupies our research group pertains to the social niche of robotic artifacts in the company of the robotically uninitiated public-at-large: What is an appropriate first role for intelligent human-robot interaction in the daily human environment? The time is ripe to address this question. Robotic technologies are now sufficiently mature to enable interactive, competent robot artifacts to be created [4,10,18,22]. The study of human-robot interaction, while fruitful in recent years, shows great variation both in the duration of interaction and the roles played by human and robot participants. In cases where the human caregiver provides short-term, nurturing interaction to a robot, research has demonstrated the development of effective social relationships [5,12]. Anthropomorphic robot design can help prime such interaction experiments by providing immediately comprehensible social cues for the human subjects. In contrast our interest lies in long-term human-robot relationships, where a transient suspension of disbelief will prove less relevant than long-term social engagement and growth. Existing research in this area is often functional, producing an interactive robot that serves as an aide or caregiver [13,19]. The CERO figure is of particular interest due to its evaluation as a robot interface representative in an office environment over a period of several months. Note that such long-term interaction experiments often revisit the robot morphology design question. Anthropomorphism can be detrimental, setting up long-term expectations of human-level intelligence or perception that cannot be met. Robots such as eMuu and Muu2 exemplify the same aesthetic principles of non-anthropomorphic expressiveness sought by our research group [3]. Most closely aligned to the present work are those projects in which the robot’s role is to be a vessel for exploration and creativity. Billard’s Robota series of educational robots provide rich learning experiences in robot programming [4]. Coppin’s Nomad rover serves as a telepresence vehicle for the public [8]. Although the human-robot relationship is secondary, the robot nonetheless provides displaced perception and exploration, inspiring users with regard to both robotics and NASA exploration programs. Educational robotics kits such as LEGO Mindstorms [14] also provide inspiration regarding science and technology. Such kits provide, in the best case, an iconic programming interface. Without depending upon previous programming experience, this enables a child to guide the behavior of their robotic creation over the short term. Teaching by example and durative scheduling are aspects of robot expression that are not addressed by these kits. Our aim is to develop a comprehensive example of long-term, social human-robot interaction. Our functional goal is to develop a robot that can enter a direct user relationship without the need for a facilitator (e.g. an educator) or a specially prepared environment (e.g. a classroom). We propose that an appropriate strategy is to develop a robot functioning within the human domestic environment that serves as a creative and expressive tool rather than a productive appliance. Thus the goal of the Personal Rover project is to design a capable robot suitable for children and adults who are not specialists in mechanical or electrical engineering. We hypothesize that the right robot will help forge a community of creative robot enthusiasts and will harness their inventive potential. Such a personal rover is highly configurable by the end user: a physical artifact with the same degree of programmability as the early personal computer combined with far richer and more palpable sensory and effectory capabilities. The challenge in the case of the personal rover is to ensure that there will exist viable user experience trajectories in which the robot becomes a member of the household rather than a forgotten toy relegated to the closet. A User Experience Design study conducted with Emergent Design, Inc., fed several key constraints into the rover design process: the robot must have visual perceptual competence both so that navigation is simple and so that it can act as a videographer in the home; the rover must have the locomotory means to travel not only throughout theinside of a home but also to traverse steps to go outside so that it may explore the back yard, for example; finally, the interaction software must enable the non-roboticist to shape and schedule the activities of the rover over minutes, hours, days and weeks. In the following sections, we present corresponding details of the comprehensive design of the robot mechanics, teaching interface and scheduling interface. Rover Mechanics and Control Rover Hardware The rover’s size and shape are born from practical constraints regarding the home environment together with the goal of emulating the aesthetics of the NASA exploratory rovers. Users should be able to easily manipulate the rover physically. And the rover must be small enough to navigate cramped spaces and large enough to traverse outdoor, grassy terrain and curbs. The fabricated rover’s physical dimensions are 18”x12”x24” (length, width, height). Four independently powered tires are joined laterally via a differential. Each front wheel is independently steered by a servomotor, enabling not only conventional Ackerman steering but also the selection of any center of rotation along the interior rear