Back to search

FRINATEK-Fri prosj.st. mat.,naturv.,tek

Co-tester: Collective-Adaptive Testing of Coevolving Autonomous Cyber-Physical Systems of Systems under Uncertainty

Alternative title: Co-tester: Kollektiv-Adaptiv tester av Coevolving Autonomous Cyber-Physical Systems av systemer under usikkerhet

Awarded: NOK 12.0 mill.

Nowadays, approaches for developing, operating, and evolving Cyber-Physical Systems (CPSs) to realize and optimize complex processes have started to change radically, mainly due to the use of advanced technologies such as Artificial Intelligence and Machine Learning. CPSs developed with such technologies have novel characteristics, including 1) enabling the integration of previously technologically isolated CPSs into a significantly larger, highly cooperative, and intelligent system, i.e., Cyber-Physical Systems of Systems (CPSoS); 2) constantly evolving individually and cooperatively to respond to ever-changing operating environment; 3) frequent interactions among CPSoS's CPSs. Such characteristics threaten the dependability of CPSoS, which consequently threatens their capability to provide trustworthy services. Though there exist testing methodologies for ensuring the dependability of individual CPSs, testing CPSoS as a whole with the consideration of the constantly evolving nature of individual CPSs and the CPSoS itself under ever-changing environment, and various levels of internal and external uncertainties inherent in the CPSoS and its operating environment, is a very new and challenging research area. The Co-tester project aims to addresses this challenge by developing novel testing strategies, empowered with advanced AI technologies, supported with novel strategies for discovering uncertain and unknown behaviors of CPSoS, its constituted CPSs, and operating environment, which will holistically and intelligently manipulate and manage evolutions of CPSoS, the CPSs, and their operating environment for extensive testing of CPSoS. Until now, we have started to investigate various advanced techniques for testing CPSoS including digital twins and Large-Language Models. In particular, we are working with software in the loop simulation platforms for self-driving cars and initial results are promising.

Nowadays, approaches for developing, operating, and evolving Cyber-Physical Systems (CPSs) to realize and optimize complex processes have started to change radically, mainly due to the use of advanced technologies, e.g., the Internet of Things and Artificial Intelligence (AI). CPSs developed with such technologies have novel characteristics, e.g., 1) enabling the integration of previously technologically isolated CPSs into a significantly larger, highly cooperative and intelligent system, i.e., Cyber-Physical Systems of Systems (CPSoS); 2) constantly evolving individually and cooperatively to respond to ever-changing physical and technological environment, under which the CPSoS and its constituted CPSs are operated; 3) frequent interactions within the CPSoS (among its CPSs) and with its environment foster new emergent behaviors and services that cannot be supported by individual CPS or a small number of CPSs. These characteristics threaten the dependability of CPSoS, which consequently threatens their capability to provide trustworthy services. Though there exist testing methodologies for ensuring the dependability of individual CPSs, testing CPSoS as a whole with the consideration of the constantly evolving nature of individual CPSs and the CPSoS itself under ever-changing environment, and various levels of internal and external uncertainties inherent in the CPSoS and its operating environment, is a very new and challenging research area. The Co-tester project aims to addresses this challenge by developing novel collective-adaptive testing strategies, which will be empowered with advanced AI technologies (e.g., machine learning and evolutionary computation techniques), supported with novel strategies for discovering uncertain and unknown behaviors of CPSoS, its constituted CPSs, and operating environment, which will holistically and intelligently manipulate and manage evolutions of CPSoS, the CPSs, and their operating environment for extensive testing of CPSoS.

Funding scheme:

FRINATEK-Fri prosj.st. mat.,naturv.,tek