Conceptual Framework for Adaptive Safety in Autonomous Ecosystems.
| Authors | |
|---|---|
| Year of publication | 2023 |
| Type | Article in Proceedings |
| Conference | Proceedings of the 18th International Conference on Software Technologies - ICSOFT |
| MU Faculty or unit | |
| Citation | |
| Web | https://doi.org/10.5220/0012086600003538 |
| Doi | http://dx.doi.org/10.5220/0012086600003538 |
| Keywords | Autonomous Collaborative Ecosystems; Adaptive Safety; Software Architecture; Trust; Security; Autonomous Vehicles; Smart Agents; Digital Twins |
| Description | The dynamic collaboration among hyper-connected Autonomous Systems promotes their evolution towards Autonomous Ecosystems. In order to maintain the safety of such structures, it is essential to ensure that there is a certain level of understanding of the present and future behavior of individual systems in these ecosystems. Adaptive Safety is a promising direction to control access to features between cooperating systems. However, it requires information about its collaborators within the environment. Digital Twins could be used to predict possible future behavior of a system. This paper introduces a conceptual framework for Adaptive Safety that is being triggered based on the trust score computed from the predictive simulation of Digital Twins, which we suggest to use in Autonomous Ecosystems to load and safely execute third-party Smart Agents. By quantifying trust towards the agent and combining it with a decision tree, we leverage this as a deciding factor to conceal or expose certain features among collaborating systems. |
| Related projects: |