Disturbance observer-based predictive formation control of mobile robots under switching topology via neurodynamic...

This paper addresses the formation control of mobile robots subject to switching topologies and persistent unknown disturbances. A novel disturbance observer-based robust distributed model predictive control framework is proposed to achieve formation coordination. Specifically, the formation task is decoupled into path-following and coordination sub-problems using an improved virtual structure. To ensure computational efficiency, the convex quadratically constrained quadratic programming problem is decomposed into quadratic programming subproblems via sequential quadratic programming and solved by an efficient neurodynamic optimizer. Theoretical analysis provides rigorous proofs of closed-loop stability and recursive feasibility, explicitly addressing the simultaneous effects of switching topologies and disturbance observer estimation residues. Comparative simulations and outdoor experiments on uneven terrain demonstrate that the proposed strategy ensures robustness in practical scenarios and real-time performance.