Dynamic Consensus‐Based Formation Control With Output Mask Privacy for Multi‐Robot Systems

ABSTRACT This article proposes a formation control scheme with privacy preservation for a multi‐nonholonomic mobile robot system based on the dynamic average consensus strategy. In the scheme, each robot can only access the data of its neighboring robots. Information sharing between adjacent agents inevitably poses a risk of private information leakage. A time‐varying output mask function is employed within the dynamic consensus algorithm to address the issue of preserving privacy information. The masking function performs a time‐varying transformation before a node transmits its data to neighboring nodes, concealing its actual values from other nodes while keeping the original convergence behavior of the dynamic system. Compared to traditional privacy‐preserving techniques, such as those relying on noise addition or cryptographic methods that affect the final convergence value, the output mapping masking algorithm significantly simplifies the algorithm's structure and reduces computational overhead. Theory proof and simulation results confirm that the proposed scheme can steer the multiple robot system converging into the predefined formation pattern and maintain the formation motion along the target trajectory.