xGDS (Exploration Ground Data Systems) synthesizes real world data (from sensors, robots, ROVs, mobile devices, etc) and human observations into rich, digital maps and displays for analysis, decision making, and collaboration. xGDS evolved from our work developing tools to control, monitor and collect data from autonomous rovers supporting NASA's terrestrial field science research. xGDS is a highly collaborative, interactive suite of web software.
xGDS Project Lead Matt Deans gave this Google Tech Talk in March 2013.
xGDS processes and maps data (including video) in real-time during operations and uses it to support live role-based geolocated note taking. Notes can be used to search for and display important data. With xGDS, scientists can do real time analysis of data coming from their instruments, make inferences and plan new data collection operations while still in the field.
Our server infrastructure can run without access to the internet, which is crucial for remote deployments where communication is limited, unreliable, or unavailable. Our post-field deployment supports archiving and review of all data, the opportunity to augment and edit field notes, search and export in common formats (e.g. JPEG, KML, CSV) for specialized analysis.
Our web tools are written in Python on Django and run on standard Linux infrastructure (Ubuntu 14.04, Apache 2.4 and MySQL). We have NASA’s permission to open source the software under the Apache 2 license and plan to publish it on GitHib soon.
The xGDS team supports a science operation by deploying a custom web site that ties together a variety of apps that provide functions like planning, monitoring, and analysis. As much as possible, we reuse apps across a variety of operations, but generally some extensions are required; for example, to ingest new data formats.
Often, members of the xGDS team provide dedicated on-site support during field operations. These operations can become a great laboratory where we quickly iterate on new software features while co-located with the science team. We are dedicated to a user-centered design process where we work closely with scientists to understand how they think about their work before we begin to develop software to support them.
xGDS has developed a web-based software platform consisting of a standard stack of third-party software (Linux, Apache, MySQL, Python), a set of our own reusable apps, and a convention for how we organize new science sharing web sites. Here are a few of our reusable apps:
Planner. A simple map-based tool for laying out exploration plans that involve a sequence of "stations" or places where specific science tasks should be performed. The resulting plans may be passed to a rover for automated execution, or may be printed for use by an astronaut, ROV or submersible pilot.
Map server. A repository for organizing multiple layers of map data. Provides a global map feed that contains all the map data and updates automatically, so the entire operations team is always viewing the latest version.
Plot. A tool for plotting time series data. In live view, plots automatically scroll and update with new data. Scientists can choose which time series to display together, drag the plot window back in time to view historical data, and zoom in and out to view processes at time scales ranging from seconds to weeks. Time series associated with a moving platform like a rover can also be plotted as colored trails in a live updating map.
Notes. A place for console operators to enter notes while a test is in progress. Notes are automatically timestamped. Scientists can display a map of note markers, showing where the exploration platform was when the note was entered. They can also display notes in a list which includes thumbnails of camera frames captured around the time of each note.
NASA conducts a variety of planetary analog studies, where science teams practice operational techniques for future space missions while doing real science in extreme environments on Earth. It has been our privilege to support some wonderful collaborators on these projects:
NASA RESOLVE Rover
In Situ Resource Utilization (ISRU)
RESOLVE is a proposed mission to survey the lunar surface for water ice and other useful compounds with a rover. xGDS supported the RESOLVE / ISRU 2012 rover field test in Mauna Kea, Hawaii, and continues to support further testing.
Mojave Volatile Prospector (MVP)
MVP is a science-driven field program with the goal of producing critical knowledge for conducting robotic exploration of the Moon. MVP conducted a simulated lunar polar rover mission with the NASA KREX-2 rover operated in the Mojave Desert by a remote Science Operations Center located at NASA Ames Research Center and a Rover Operations Center located in the Mojave, with each location staffed with specific console positions.
Pavilion Lake Research Project (PLRP)
PLRP is an international, multi-disciplinary science and exploration effort to explain the origin of freshwater microbialites in Pavilion Lake, British Columbia, Canada, using divers, autonomous underwater vehicles, and crewed submersibles.
Planetary Lake Lander (PLL)
PLL is a multi-disciplinary study of the impact of rapid deglaciation at Laguna Negra, a high altitude lake in central Chile, using a probe buoy and a variety of auxiliary sensors.
NASA Extreme Environment Mission Operations (NEEMO)
NEEMO sends groups of astronauts, engineers and scientists to live in Aquarius, the world's only undersea research station, for up to three weeks at a time. NEEMO tests conduct seafloor surveys using operational techniques similar to future asteroid exploration missions, with divers, autonomous underwater vehicles, and crewed submersibles.
NASA Desert Research and Technology Studies (Desert RATS)
The Research and Technology Studies (RATS) team evaluates technology, human-robotic systems and extravehicular equipment for future human exploration missions in space. xGDS supported RATS tests at Black Point Lava Flow, Arizona, involving astronauts, autonomous rovers, and crewed surface vehicles.
K10 Robot Field Tests at Haughton Mars Project
K10 tested techniques for robotic follow-up of human exploration at Haughton Crater, an ancient impact crater in High Arctic Canada where geologists study questions relevant to future lunar exploration.