ESA Open Invitation to Tender: 1-10685
Open Date: 16/06/2021 15:54 CEST
Closing Date: 29/09/2021 13:00 CEST

The objective of the activity is to improve the health monitoring of the satellite subsystems by applying artificial intelligence techniques to identify anomalies and unexpected behaviour of the different platform sub-systems and software elements of the spacecraft.Targeted Improvements: To enable predictive on-board health monitoring, increase of the dependability of the system and avoid operations to be performed at nights and weekends. Description: The current implementation of on-board fault detection, isolation and recovery is relying on monitoring of thresholds and on predefined models. Actions are only executed when these thresholds are reached. As these thresholds are defined before launch they are conservative and include margins. During the life of the communication satellite, a failure is only detected at the time a parameter reaches a threshold. Being able to detect a deviation early will make it possible to initiate warnings and to perform analyses before the threshold is reached. Abnormal behaviour within the defined thresholds is often a sign of anomalies and can be detected. As consequence, some mitigation actions could be taken earlier and avoid anyinterruption of service. Some artificial intelligence (AI) techniques targeting this problem have been successfully implemented on ground, even in the space domain for processing of housekeeping telemetry (novelty detection, in use in ESA operations since 2012). However, the data that is transferred to ground is a limited subset of the data available on-board and, in some cases, has been alreadyprocessed. This activity will improve the health monitoring of the satellite subsystems by applying artificial intelligence techniques to identify anomalies and unexpected behaviour of the different platform sub-systems and software elements of the spacecraft.Themission elements critical elements to be monitored will include mechanisms (e.g. reaction wheel, solar array drive antenna mechanism, etc.) and software (e.g. CPU load, health status, etc.). The on-board health monitoring system shall be able to provide processed information that will assist in the identification of the root cause of anomalies, considering knowledge and experience obtained from previous anomalies. The selected AI techniques shall be implemented in a representative on-board state-of-the-art computer using, if necessary, AI accelerators. Time and space partitioning is recommended to ensure that the execution of the on board healthmonitoring system will not affect the execution of critical tasks. The activity shall validate the approach by executing the AI techniques on data sets covering both nominal and degraded/failure cases and shall obtain an accuracy of 80% or more.

Directorate: Directorate Telecom Integrated Applica
Estabilishment: ESTEC
ECOS Required: No
Classified: No
Price Range: > 500 KEURO
Contracts Officer: Florence Odette Jeanne Glandieres
Initiating Service: TIA-TT
IP Measure: N/A
Prog. Reference: E/0534-01G – CC-AT 4.0.1
Tender Type: Open Competition
Open To Tenderers From: AT+BE+CA+CH+CZ+DE+DK+EL+ES+FI+FR+GB+HU+IE+IT+LU+NL+NO+PL+PT+RO+SE
Technology Keywords: 2-B-I-Methods and Tools for Onboard Software Engineering Processes
Products Keywords: 2-M-4-e-Other