From Local Rules to Global Behavior: A Bottom-Up Framework for Directing Emergent Behavior in Autonomous Systems

Biological systems produce a remarkable variety of filamentous structures through processes such as secretion, aggregation, self-assembly, and polymer organization. Yet not all biological architectures arise purely through self-assembly. Some organisms actively extrude, deposit, and arrange fibers to construct functional structures, such as spider webs and silkworm cocoons. At a higher level of complexity, superorganisms collectively build structures that adapt to their environment and meet the functional needs of the colony, as seen in the tents of social caterpillars and the communal webs of social spiders. In these systems, each individual follows simple local rules, depositing material in response to immediate environmental cues. The resulting architectures exhibit complex, functional properties that emerge from purely local interactions. Inspired by these natural systems, we design systems of autonomous agents programmed to deposit fibrous materials according to local, rule-based interactions. A central challenge is that the mapping between local deposition rules and the global emergent properties of the resulting structures is both non-deterministic and non-unique. Nevertheless, by systematically designing and tuning the rule sets, we demonstrate that it is possible to direct the emergence of desired geometrical and mechanical properties in such structures. This approach offers a new paradigm for achieving global control in distributed, locally driven construction systems.