Figure 1. Zoë in the Atacama Desert.
or mapping strategy that adapts as data is collected.
Decisions about which samples to acquire and where
and when to travel next can be based upon metrics of
information gain. Similar metrics can also be used to
prioritize science data for download. Intelligent compression strategies use knowledge or models of content to interpret and summarize in a compact form.
The ultimate goal of science autonomy is to embody
sufficient understanding, quantified by models and
metrics, so that rovers can independently choose
actions that best support the scientific investigation
in which they are engaged. Rovers will take their
goals and guidance from scientists, but when isolated they should make scientifically rational decisions
and when in communication they should provide
the most relevant information possible.
Science autonomy is especially valuable for surface
rover operations because missions have finite lifetime
and rarely revisit sites after the first encounter — the
rover must make good decisions and get it right the
first time. Recent demonstrations on spacecraft show
increasingly sophisticated science autonomy capabil-
ities. Milestones include target tracking during the
Deep Impact comet flyby (Mastrodemos, Kubitschek,
and Synnott 2005); target detection and response by
the Mars Exploration Rovers (Castaño et al. 2008;
Estlin et al. 2012); and spectral detection, discovery,
and mapping by the EO- 1 spacecraft (Chien et al.
2005; Davies et al. 2006; Doggett et al. 2006; Ip et al.
2006; Thompson et al. 2013). At the same time, new
smart instruments are beginning to incorporate
autonomous science data analysis directly (Wagstaff
et al. 2013) and provide information that can be used
to guide the rovers’ targeting and operation.
These techniques and others will enable surface
rovers to achieve multiday autonomous operations.
Currently multiday rover plans do not travel over the
horizon of yesterday’s imagery, which limits the daily science yield. However, rover navigation already
permits safe over-the-horizon traverses, and in principle a rover could autonomously survey large areas
of terrain with its full suite of instruments. In one
natural arrangement, operators would direct the
rover using waypoints determined from satellite
images, relying on rover autonomy for low-level hazard avoidance and science target selection en route.