A Computational Framework for Autonomous Self-repair Systems
This paper describes a novel computational framework for damage detection and regeneration in an artificial tissue of cells resembling living systems. We represent the tissue as an Auto-Associative Neural Network (AANN) consisting of a single layer of perceptron neurons (cells) with local feedback loops. This allows the system to recognise its state and geometry in a form of collective intelligence. Signalling entropy is used as a global (emergent) property characterising the state of the system. The repair system has two submodels - global sensing and local sensing. Global sensing is used to sense the change in whole system state and detect general damage region based on system entropy change. Then, local sensing is applied with AANN to find the exact damage locations and repair the damage. The results show that the method allows robust and efficient damage detection and accurate regeneration.
KeywordsSelf-repair Multi-cellular structures Regeneration Auto-Associative Neural Network Perceptron Signalling entropy Modeling
Authors gratefully acknowledge support of the the following: TNM - Doctoral Scholarship from VIED, Vietnam; J.A.- Sabbatical at Lincoln University, New Zealand; S.S.- Lincoln University Research Fund; M.L.- DARPA (#HR0011-18-2-0022), the Allen Discovery Center award from the Paul G Allen Frontiers Group, and the Templeton World Charity Foundation (TWCF0089/AB55 and TWCF0140).
- 4.Dinsmore, C.E.: A history of regeneration research: milestones in the evolution of a science. J. Hist. Biol. 26(1), 156–158 (1993)Google Scholar
- 6.Ferreira, G.B.S., Smiley, M., Scheutz, M., Levin, M.: Dynamic structure discovery and repair for 3D cell assemblages. In: Proceedings of the Fifteenth International Conference on the Synthesis and Simulation of Living Systems (ALIFEXV) (2016)Google Scholar