Robust Cooperative Formation Maneuver Control of Multi-Omnidirectional Robot Systems Considering Roller Effects and Ground Friction

Abstract

To address the robust formation maneuvering control problem of multi-omnidirectional robot systems under the combined effects of the roller effect and ground friction, this paper proposes a composite disturbance-rejection robust control architecture integrating a hierarchical observer and a sliding mode controller. First, a hierarchical observer composed of a finite-time disturbance observer and a variable-bandwidth extended state observer is designed to achieve modular estimation and differentiated compensation of disturbances acting through different channels. Second, by incorporating sliding mode control, a distributed formation-tracking protocol based on neighborhood interactions is developed to guarantee uniformly ultimately bounded stability of the formation states. Meanwhile, a Kalman filter-based high-frequency smoothing mechanism is introduced to attenuate chattering induced by sliding mode switching and high-frequency components arising from disturbance estimation errors, thereby enhancing the smoothness of the control inputs. Finally, numerical simulations and prototype experiments demonstrate that the proposed control strategy achieves better control accuracy and robustness under multiple disturbance conditions. Note to Practitioners—In practical engineering applications, multi-omnidirectional robot systems are simultaneously disturbed by roller effects, ground friction, and measurement noise, which seriously degrade the accuracy of formation control. In this paper, a robust formation control architecture for engineering applications is proposed, wherein a hierarchical observer is employed to achieve differentiated compensation for roller effects and ground friction. Within this architecture, a finite-time disturbance observer and a variable-bandwidth extended state observer are designed to operate in concert without reliance on an accurate robot model. On this basis, a sliding mode formation controller and a Kalman filter smoothing mechanism are integrated to realize distributed coordination via the inter-robot communication topology, thereby preventing individual disturbances from inducing formation collapse. Concurrently, high-frequency interference arising from sliding mode chattering and measurement noise is effectively suppressed, which enhances the smoothness of the control input. In practical deployment, only the observer and filter parameters are required to be adjusted by users based on the platform’s motion characteristics and sensor noise levels. Numerical simulations and experimental results of the proposed scheme are all presented.