This article introduces a hybrid setup for cooperative multi-robot navigation that relies on concepts from Dynamic Learning Cuckoo Search (DLCS), Adaptive Neuro-Fuzzy Inference Systems (ANFIS) and Petri Net-based coordination to achieve effective navigation in a complex dynamic and constrained environment. The system intends to enhance global path planning, local trajectory following, collaborative coordination and obstacle avoidance techniques for multi-robot systems operating autonomously.
The system was validated through simulation and real-time experiments using 4 Khepera IV mobile robots in two different 2 × 2m environments, such as a structured obstacle environment with constrained corridors and safety margin inflation and an irregular obstacle environment to evaluate adaptability and local replanning capability. Performance was evaluated using path length, motion time, planning time, replanning frequency and trajectory tracking accuracy.
Results demonstrated that the proposed DLCS-based framework generated smoother and shorter paths than the conventional Cuckoo Search (CS). In the structured environment, the approach reduced path length by approximately 6–8%, motion time by 5–7%, planning time by 8–10% and replanning frequency by 10–14%. In irregular obstacle environments, the farmwork showed improved adaptability, smoother trajectory generation and reduced directional oscillations during multi-robot interactions. The ANFIS controller achieved stable trajectory tracking with rapid heading-error convergence and lower steady-state error. Simulations and experimental results show close agreement, validating the real-time feasibility of the proposed framework.
The proposed work integrates DLCS-based global optimization, ANFIS-based adaptive tracking and Petri Net-based cooperative coordination within a unified navigation framework.
