Adaptive threshold-based task allocation for swarm search-and-rescue robotics in unknown environments with limited communication

Task allocation is a central challenge in swarm robotics, yet most existing approaches assume known target locations or global communication, limiting their applicability in real-world search-and-rescue (SAR). This paper proposes a distributed method, the adaptive dynamic response threshold model (A-DRTM), to address SAR tasks in unknown and noisy environments under communication constraints. A-DRTM defines a task selection probability function that dynamically adjusts thresholds according to task demand disparities. Stimulation and jitter mechanisms are further incorporated to enhance flexibility and avoid local optima, improving responsiveness to environmental changes and coordination efficiency. Comprehensive simulations across varying initial positions, map sizes, robot-to-task ratios, communication ranges, and moving targets show that A-DRTM consistently outperforms six representative threshold-based algorithms. It achieves higher task completion, better resource utilisation, and stronger scalability, particularly in large-scale and communication-limited conditions.