This study aims to explore how globally feasible foraging paths with collective consensus can emerge from individual rules and local perception. Inspired by natural ant colonies and based on existing biological research on their behavior, we develop a rule-based system that simulates the decision-making processes of individual ants. The rule system consists of three levels: instantaneous reactive rules based on antennal and limb sensing; short-term coordination rules informed by memory of task duration and collision frequency; and long-term optimization rules guided by pheromone dynamics. To theoretically model these processes, we employ a cellular automata framework characterized by spatiotemporal discreteness, local sensing, and parallel computation. This framework enables us to simulate how individual ants, leveraging local perception, interactions, and parallel task execution, collectively generate complex emergent collective behavior. To validate the effectiveness of the proposed model, we constructed a variety of terrains in NetLogo for simulation experiments. The results demonstrate that high-efficiency global foraging paths can emerge solely from individual-level rules and local interactions, without centralized control. Furthermore, these rules exhibit strong environmental adaptability, enabling the ant colony to dynamically adjust individual decisions in response to local environmental cues, thereby optimizing the overall foraging paths. This approach provides a novel and scalable framework for distributed multi-robot systems to address complex tasks, enhance coordination efficiency, and improve adaptability in dynamic and unstructured environments.