Formation Control of Non-Holonomic Mobile Robots: Predictive Data-Driven Fuzzy Compensator

Jinfeng Wang,Hui Dong,Fenghua Chen,Mai The Vu,Ali Dokht Shakibjoo,Ardashir Mohammadzadeh

DOI: https://doi.org/10.3390/math11081804

IF: 2.4

2023-04-11

Mathematics

Abstract:A key research topic in the field of robotics is the formation control of a group of robots in trajectory tracking problems. Using organized robots has many advantages over using them individually, such as efficient use of resources, increased reliability due to cooperation, and better resistance against defects. To achieve this, a controller is proposed that steers the leader robot and subsequent follower robots asymptotically to a reference trajectory. The basic controller is feedback linearization. To ensure stability against perturbations, a compensator based on type-3 fuzzy logic systems (T3-FLSs) and a data-driven control strategy is designed. The approach involves employing a finite number of open-loop data and using the model-based predictive controller (MPC) approach to acquire sufficient criteria for stability. An infinite-horizon function is minimized online, which allows the data-based control policy to be considered the optimal control method. The gains of the constrained data-based control signal are computed at each time step to enhance accuracy. Applying the data-based state feedback controller to the system yields positive and stable state trajectories with appropriate transient responses. The suggested data-driven compensator is guaranteed to handle constraints. A practical example is simulated to evaluate the proposed strategy.

mathematics