S Sisdarmanto Adinandra

Saturated control of time-varying formations and trajectory tracking for unicycle multi-agent systems

We propose a synchronization approach to solve a problem where multiple unicycle agents are required to follow individual reference trajectories while maintaining a time-varying formation. Motions of the agents are synchronized thanks to coupling terms in their feedback control laws. Under saturation constraints on the control signals, our control laws guarantee global asymptotic zeroing of the synchronization errors and global asymptotic stability of the tracking error dynamics. For stronger controller couplings, the robustness of formation keeping to perturbations is increased. The proposed approach is successfully validated in experiments.

Collision-free tracking control of unicycle mobile robots

We propose a tracking control with collision avoidance for a group of unicycle mobile robots. A supervisory system assigns to each robot its reference path, together with the desired velocity profile as a function of the position along the path. The robot paths and velocity profiles do not necessarily prevent collisions of the robots. Based on the actual robot coordinates, reference velocities of individual robots are coordinated in real-time such as to ensure collision-free robot movements along the assigned paths. The robot motion controllers realize globally asymptotic stable tracking of the reference trajectories under constraints on the actuator inputs. The controllers proposed here offer several advantages with respect to those in the literature, such as better suppression of tracking errors due to more freedom in controller tuning and relaxed constraints on the reference velocities. The suggested tracking control with collision avoidance has successfully been validated in experiments. It can be used for the realization of autonomous transportation tasks in warehouses, harbors, factories, etc.

A Practical Model Predictive Control for A Group of Unicycle Mobile Robots

Abstract This paper presents the validation of a control algorithm for a group of unicycle mobile robots based on nonlinear model predictive control. The paper focuses on the practical application of the control algorithm. The architecture for the multiagent system employed in this paper is considered as a sequentially decentralized method. A combination of steepest descent and line search techniques is used to solve the optimization problem. Using transport in a warehouse as a case study, the strength of the proposed control algorithm is shown in experiments of up to 12 robots. A performance analysis of the influence of different control parameters is given.

Flexible Transportation in Warehouses

In recent years, autonomous mobile robots (AMR) have emerged as a means of transportation system in warehouses. The complexity of the transport tasks requires efficient high-level control, i.e. planning and scheduling of the tasks as well as low-level motion control of the robots. Hence, an efficient coordination between robots is needed to achieve flexibility, robustness and scalability of the transportation system. In this chapter, we present a methodology to achieve coordination in different control layers, namely high-level and low-level coordination. We investigate how the coordination strategies perform in an automated warehouse. We use simulation results to analyse the system performance. We take into account typical performance indicators for a warehouse, such as time to accomplish the transportation tasks and total cost of the system. In addition to the simulation results, we conduct experiments in a small-scale representation of the warehouse.