Through ESA's Space Situational Awareness (SSA) Programme technologies are developed to maintain catalogued information of the population of man-made space objects. To achieve this task, up-to date data on these objects are crucial. Scheduling of new observations in short and long term and monitoring their execution is essential for most of the related services and functionalities. An efficient, easy-to-use and automated process of tasking various sensors allows one to tap into a potentially very large source of existing observatories, yet at the same time presents the challenge of integration of various sensor types. Such sensors will have different levels of automation and autonomy executing their existing programs, which will have higher or lower priorities than the priorities of currently needed observations.
In this contribution we will present WebPlan, a web-based sensor scheduling tool developed for ESA as a response to the above challenge. WebPlan is a marketplace and meeting platform for SST data consumers and data producers, with a possible capacity for 100-1000 sensors provisioning data every day and capable of <30 minutes request-to-delivery times (limited by the provisioning sensors capability for reaction and data processing). The system has been in a production environment at Sybilla since Dec 2019.
The system successfully addresses the following scheduling challenges:
• available computation time and resources (the algorithm has to perform on “a modern laptop”),
• multiple objectives (minimizing error envelope, cost),
• dynamic changes (i.e. weather, failures, other players),
• benchmark data to evaluate scheduling algorithms (set of rules prepared by experts),
Catalogue maintenance is performed with the use of the NSGA-II heuristic algorithm which produces schedules 20-30% better than prepared by experts with a set of the carefully selected rules. The algorithm is, e.g. tasked with minimization of idle time, time overheads and cost, maximizations of goals and observation quality, clustering effect and single or multiple observations impact on error envelope.
WebPlan web user interface integrates all of the system components into one web application. The following modules are present in the Web UI: Network (CentralHub), AstroDrive (storage), Scheduling (Timeslots, Programs, Scheduling Agents), Object catalogues (SpaceTrack/Celestrak, Simbad, SSO/NEO).
The system was initially tested on two use cases: catalogue maintenance and ad-hoc observations. Six sensors (located in Germany, Spain, Poland, and Australia) were used in two test campaigns and a total of 40 targets were observed during two test campaigns in Dec 2019 and March 2020. The system was rescheduled every 12/24 hours, and the envelope error was updated upon arrival of a new TDM file. A total of 25931 FITS were gathered and 242 TDMs generated.
WebPlan is a large (over 50 services on production environment) and complex (5 entities involved in the longest chain of tasking) distributed service based on Docker virtualization. It includes multiple custom scheduling agents executing specialized scheduling tasks for GNSS calibrations, GEO tracking, high interest events, re-entry or collision avoidance data provisioning. Support for multiple observing campaigns and programs is possible with easy book-keeping and results delivery.