Accurate prediction of space station telemetry data can improve system monitoring. This paper presents a methodology for the acceleration of short-term forecasting of power data on an embedded platform designed for space. Initially, a long short-term memory (LSTM) network is trained to forecast voltage and current values from the ISS. This LSTM forecasts voltage and current minutes into the future while maintaining a low error rate. This LSTM network is then accelerated using the FPGA from a Zynq-7045 system on a chip (SoC). The Zynq-7045 was selected because it is the same SoC used on the SHREC Space Processor. Networks with differing parameters and dimensions were realized and evaluated withing the resource constraints of the Zynq-7045. The best performing LSTM networks were able to achieve over 3 × speedup against a software baseline with minimal increase in forecasting error.
Space Station Power Forecasting with LSTMs for an Embedded Platform
2021-08-16
903907 byte
Conference paper
Electronic Resource
English
Probabilistic traffic breakdown forecasting through Bayesian approximation using variational LSTMs
Taylor & Francis Verlag | 2023
|Partial Label Learning of RF Emitters with LSTMs
AIAA | 2020
|Monitoring spacecraft telemetry with a fleet of LSTMs
NTRS | 2017
|