_____________________________________________________ Proc. ‘ISAIRAS 2005 Conference’, Munich, Germany, 5-8 September 2005 (ESA SP-603, September 2005) MARS EXPLORATION ROVER PERFORMANCE AS A BASELINE FOR FLIGHT ROVER AUTONOMY TECHNOLOGY ASSESSMENT Edward Tunstel, Ayanna Howard, Mark Maimone, Ashitey Trebi-Ollennu NASA Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA 91109 USA, Email:tunstel@robotics.jpl.nasa.gov ABSTRACT Technology assessments rely on performance metrics to establish a basis for rating technologies. Metrics are also used to measure relative merit of similar technologies to state-of-the-art technology. Functional performance metrics are presented for mobility and robotic arm autonomy exercised on the Mars Exploration Rovers (MER) surface mission thus far. The metrics are used to apply an existing technology assessment method to establish a baseline for assessing future flight rover technologies. The methodology decomposes robotic activities into operational functions and addresses how technologies, based on performance metrics, can be systematically related to increases in science return. Considering the basic mission objective to maximize scientific yield, we can assess how relative performance of future technologies might impact science return. We provide a useful set of metrics and present an example application of the method to assess merit of hypothetical future Mars rover performance relative to the MER baseline. 1. INTRODUCTION The utility of autonomous rovers is a function of their ability to move about and explore without frequent contact with Earth-based mission operators. More increasingly, robotic vehicle autonomy is required to achieve aspects of overall success for planetary surface missions such as the 2003 Mars Exploration Rovers (MER) and the planned 2009 Mars Science Laboratory (MSL) and ExoMars missions. With MER, NASA landed two twin rovers, named Spirit and Opportunity, on Mars in January of 2004 (Fig. 1). These rovers were explicitly required to use robotic mobility and manipulator arm positioning functionality to achieve exploration mission objectives by serving as surrogate robotic field geologists for a science team on Earth. MSL and ExoMars may be required to do the same, and perhaps more, with greater demand on autonomy and lifetime. Investments in associated autonomy research or technology products are based in part on potential to maximize scientific yield of the missions. Mission planners or systems engineers tend to justify the inclusion of new technology by determining its effect on the utility of the mission, often computed by combining the utility of outcome with the probability of achieving the outcome [1]. Utility of autonomy technology should therefore be linked to achievement of mission goals by evaluating impact on science return. However, few systematic methodologies exist that quantify the concept of science return due to autonomy technology components. Fig. 1. Spirit and lander (computer models combined with 3-D surface data acquired by Spirit’s cameras). Recent work has proposed and developed a framework that can systematically relate technologies to science return in a structured fashion [2, 3]. Mission objectives are quantified in terms of science return and achievement of objectives is represented as a set of mission operational functions. Technologies are then linked to the mission through several levels of abstraction that associate performance metrics with science return. A fundamental requirement for applying this method is availability of a set of computed performance metrics that are associated with a selected baseline technology. For purposes of technology assessment, the baseline is viewed as the state-of-the-art (SOA) against which a new or alternative technology can be evaluated with respect to relative impact on science return. MER represents the longest deployment of planetary surface robots and a new benchmark in planetary robotic autonomy. As such, it is important to capture and document the rovers’ performance in ways that facilitate evaluation of similar technologies relative to