Modern architectures are based on ‘three-layer’ architectures: one layer for slow deliberate activities (decision layer), one layer for fast reaction (executive and/or reactive layer) and one to achieve independency from the hardware layer (functional layer). SRC guidelines [RD.4] suggest the use of three-layer architecture.
In line with this approach and with the expressed requests of the SRC Guidelines for OG2, it is proposed for ERGO three-layer architecture (illustrated in Figure 1‑2). Its design benefits from the experience and lessons learnt gained within different projects by the team members and mostly from GOAC [RD. 12], the first Goal Oriented Autonomous Software Framework developed in Europa.
Figure 1‑2 shows the proposed ERGO architecture (OG2) and its relation with the other OGs (1, 3, 4 and 5).
As already mentioned, ERGO shall be able to handle all four level of autonomy (E1 to E4) as defined in section 1.3.1. Nevertheless one of the major objectives of ERGO is to design an autonomous framework, thus targeting the maximum level of autonomy, E4. For this purpose, a combination of knowledge and expertise in different areas of knowledge is required.
This consortium provides a strong background and heritage in the different areas of knowledge required, namely:
- Planning and Scheduling and Robotic controllers for autonomous systems.
- Critical SW development for space missions, both for the space and the ground segment, leveraging on flight experience such as the IXV and on state of the art OBSW developments of the on-board SW for ExoMars.
- Development of robotic applications (rovers, robotic arms).
- Multi-agent systems for space.
- Use of Formal Verification and validation techniques.
The proposal is coordinated by GMV; a company devoted to the aerospace sector for more than 20 years. All team members have also participated in the development of the autonomy software frameworks. The ERGO proposed architecture stems from the combined know-how of the consortium as a whole in all these areas.