Comparative analysis of chaotic and deterministic methods to territory coverage by drone swarms

Abstract

This study proposes a hybrid chaotic-deterministic method for controlling a swarm of unmanned aerial vehicles, integrating social interactions such as separation, cohesion, alignment, and binding with an attraction-driven search strategy. The two-level anti-stagnation system operates at the level of individual agents (‘micro’) and the collective swarm (‘macro’) to prevent local minima and maintain controlled trajectory smoothing. The method uses double course conditioning to balance smooth trajectories with stochastic exploration. Meanwhile, an attractiveness function evaluates potential movement directions quantitatively based on territory novelty, distance factors and course stability. Social interaction forces – repulsion, cohesion, alignment and centroid binding – ensure swarm stability and collision avoidance throughout mission execution.

Comparative experimental validation was conducted through multiple simulation launches for each method in irregular polygonal territories. Both the chaotic and deterministic waypoint-based approaches demonstrated exceptional mission reliability, achieving a target coverage threshold in all trials, thereby confirming complete success rates. In terms of coverage efficiency, the chaotic method achieved superior average territory completeness compared to the deterministic approach, representing a measurable improvement. However, this enhanced coverage precision comes at a significant computational time cost: the chaotic method required substantially longer average mission duration compared to the optimized waypoint method. The chaotic approach also exhibited notably higher variability in results, reflecting the inherently stochastic nature of exploration-based methods. Thus, while the chaotic method demonstrates superior coverage efficiency, it exhibits inferior time efficiency compared to the deterministic baseline.

These findings quantify the fundamental trade-off between thorough exploration and time efficiency in unmanned aerial vehicle swarm operations, providing empirical evidence to inform mission-critical deployment decisions. The results suggest that chaotic methods are best suited to scenarios that prioritize comprehensive coverage and adaptability, such as search and rescue operations where undetected casualties would be a critical failure, while deterministic approaches are more effective in time-sensitive missions with predictable environments. The reliability of both methodologies, combined with the quantification of performance differences, enables the selection of methods based on evidence, aligned with specific operational requirements, mission constraints and acceptable risk-time-accuracy trade-off parameters.