PLATO | CAMCEAD

The Cambridge Centre of Excellence for Astronomical Data (CamCEAD) at the Institute of Astronomy leads the development of the software framework and computing setup for the PLATO Exoplanet Analysis System (EAS), under Work Package 36 (WP36), which covers the discovery and characterisation of exoplanets. CamCEAD coordinates the infrastructure, backend services, pipelines, databases and distributed computing for this part of the mission.

WP36 and the EAS data flow

Within the PLATO data flow, CamCEAD is mainly responsible for the WP36 framework and for running the EAS. The data flow starts with photometry from the spacecraft and ends with catalogues of planetary systems for the science community.

PLATO SpacecraftHigh-precision stellar photometry

Ground SegmentL0/L1 baseline reduction

WP36 FrameworkEAS pipeline infrastructure and core software (CamCEAD)

EAS OperationsDistributed transit confirmation (CamCEAD)

PLATO DBCentral verification and parameters

Product DeliveryPlanet catalogues

Science CommunityAccess to confirmed planetary systems

WP36 framework

The framework carries the data through the steps that check exoplanet light curves and search for planets:

Operations: running the exoplanet pipeline code on Cambridge's computing systems.
Infrastructure: keeping software versions, packages and configurations consistent across the different analysis branches.
Modular pipelines: organising the code for automatic verification, detection of multiple transits and rejection of false positives.

EAS software

We write core modules of the Exoplanet Analysis System inside the WP36 framework:

Modelling: improving the transit modelling and the modules that spot light from a nearby star blending with the target.
Data handling: reading corrected light curves into the pipeline efficiently.
Testing: regression tests and checks run against each production version of the pipeline.

Operations

We make sure the processing stays stable under the load the mission will bring:

Readiness: large-scale tests on simulated transit data before real data arrives from the spacecraft.
Quality control: regular regression tests and performance benchmarks on the production pipelines.
Releases: planning software releases around ESA mission milestones.

CETRA: Cambridge Exoplanet Transit Recovery Algorithm

CETRA is a transit search algorithm developed for the EAS. It runs on NVIDIA GPUs using CUDA. It first searches the light curve for transits in time and then phase-folds the data using physically realistic transit shapes.

Sensitivity: in tests it found over 20% more low signal-to-noise and long-period transits than BLS and TLS.
Speed: running on GPUs makes the search much faster on large sets of high-cadence light curves.
Real data: on real datasets it gives higher detection significance, which lifts some faint candidates above the threshold for validation.
Open source: it is written in Python and available on PyPI and GitHub.

Computing on IRIS

CamCEAD uses the IRIS National Facility for the distributed cloud computing needed to process large volumes of PLATO WP36 data.

Kubernetes clusters

Kubernetes runs the exoplanet detection code and pipeline jobs in containers across many machines, so we can scale up as the mission's data grows.

Docker and Python

Each Python pipeline job runs in its own Docker container. This keeps the software environment the same every time, makes results reproducible and lets us add capacity quickly on IRIS.