Timely and efficient trending and analysis of satellite state-of-health telemetry is crucial to both pre-launch integration and testing, as well as on-orbit spacecraft and instrument operations. The Operations and Ground Systems group at UC Berkeley's Space Sciences Laboratory has been involved in the integration, test and operation of eight satellites. In response to the demand of monitoring a large fleet of spacecraft with a small staff, the operations group designed and implemented its own in-house trending software system - the Berkeley Trending Analysis and Plotting System (BTAPS). The original version of BTAPS decommutates CCSDS packet telemetry into mnemonics, converting raw values to engineering units, and stores each time-tagged value in its own column in a MySQL database for random-access lookup. This design has proved to be very reliable, however lessons learned eventually inspired a complete redesign of the database back-end. The BTAPS revision instead stores raw telemetry packets in MySQL and defers decommutation and conversion of individual data points until the time of data request. While this paradigm shifts some code and computational complexity onto client software, it makes the database schema vastly simpler, configuration more flexible, and improves performance. The revised BTAPS is agile enough to use during the earliest phases of mission development, when telemetry formats are still ill-defined and in flux. This paper explains the rationale for and benefits of redesigning the BTAPS back-end, quantifies the performance gains of the revised design, and presents real-world examples of BTAPS end-user application software.
展开▼