Operational performance metrics for mars exploration rovers

Abstract

The concept of operational performance metrics and associated measurement issues is explored for deployed robotic systems. The focus is on performance of mobility and robotic arm autonomy exercised on the NASA Mars Exploration Rovers surface mission. This planetary rover mission has been underway for nearly 3 years since January 2004 using two rovers performing separate missions. The autonomy functions of surface navigation, short-distance approach to surface science targets, and robotic placement of armmounted instruments on science targets are considered. Operational metrics that measure performance of these functions relative to system requirements are advocated. Specific metrics are defined and computed using telemetry from the rovers' multiyear operations on Mars and applied to rate performance during their respective missions. An existing methodology is applied to compute metrics with respect to required performance and to aggregate multiple metrics into a composite performance score for each rover. The formulation is augmented to accommodate importance weights that add flexibility in use of the metrics by different potential endusers, e.g., sponsors, program managers, systems engineers, and technologists. A brief example illustrates the application and effect of importance weights on overall rover performance scores.

Taxonomy, systems review and performance metrics of planetary exploration rovers

“…Locomotion mode metrics compare the performance between different locomotion types; [31] has a qualitative comparison between wheels, legs and tracks and [33] has a comparison between tracks and wheels. Operational metrics measure the required performance vs. actual in the field [40]. Metrics that evaluate the autonomy have also been proposed, such as [41].…”

Real-time assessment of robot performance during remote exploration operations

“…An important conclusion of our research is the identification and categorization of constraints affecting performance baselines (described in Section 6). We have seen examples of such constraints in other research, such as constraints on measuring traverse performance to a single regional terrain type (an example of an environmental constraint) proposed by Tunstel [14]. We know of no work, however, that characterizes these different types of constraints.…”

“…Much of the research on metrics for robotic performance has been for the purpose of assessing and comparing robot technologies [1,4,14]. Tunstel [14] defines technology metrics for the Mars Exploration Rovers.…”

“…Tunstel [14] defines technology metrics for the Mars Exploration Rovers. His objective is to define metrics characterizing the effects of technology on science return as a baseline for evaluating future rover technologies and establishing baselines for predicting technology impacts on science return.…”

Mobile robotic surveying performance for planetary surface site characterization

“…Other attributes can be adopted that are germane to surveying for planetary surface exploration but proposed in the context of mobility/navigation or task performance. These include terrain traversability measures such as obstacle abundance [17,18] and number or effort of rovers and human operators involved [11]. While identifying domain specific attributes that would enrich the relevance of existing geometric measures is helpful, the manner in which they would be incorporated into a metric formulation is also worth considering.…”