421 research outputs found
An experimental comparison of adaptive and robust control methods for precise positioning with tubular linear motors
No full textDirect drives with linear motors are attracting the attention of both industry and academia thanks to their advantages in terms of higher precision, higher acceleration/deceleration, and reduced dimensions. This paper presents a comparison between industrial PID controllers and state-of-art non-linear control strategies for accurate position tracking on a linear permanent magnet synchronous motor. Namely, the comparison considers an advanced sliding mode approach with scheduling laws related to the state trajectory in the phase plane, and an approximation-based adaptive scheme that relies on a neural network to cancel the non-linearites of the system so as to have almost linear residual dynamics. The feasibility of the control strategies is validated by an extensive experimental analysis. The schemes are both theoretically stable and guarantee accurate positioning, which are, in terms of average absolute position error, two times better than standard PI
A soft computing approach for task contracting in multi-agent manufacturing control
No full textThis paper describes a new task-contracting schema for multi-agent manufacturing control based on soft computing. It aims to apply fuzzy techniques to implement a real-time multi-criteria task-contracting mechanism for part flow control in manufacturing floor. For comparison purposes, the paper also considers other recently proposed evolutionary strategies to adapt and optimize agents' decision parameters to the changing conditions of the manufacturing floor. All the considered approaches are compared on a detailed simulation model of a hypothetical manufacturing system that was recently proposed in literature as benchmark for multi-agent control systems
A genetic approach for adaptive multiagent control in heterarchical manufacturing systems
No full textIn this paper, we apply genetic algorithms to adapt the decision strategies of autonomous controllers in a part-driven heterarchical manufacturing system. The control agents use pre-assigned decision rules only for a limited amount of time, and obey a rule replacement policy propagating the most successful rules to the subsequent populations of concurrently operating agents. The twofold objective of this approach is to automatically optimize the performance of the control system during the steady-state unperturbed conditions of the manufacturing floor, and to improve the reactions of the agents to unforeseen disturbances (e.g., failures, shortages of materials) by adapting their decision strategies. Results on a detailed discrete event model of a multiagent heterarchical manufacturing system confirm the effectiveness of the approach
New control policies preventing deadlock in automated manufacturing systems
No full textIn this paper we propose a hybrid approach to prevent deadlock between parts flowing in manufacturing systems. The approach uses both Petri nets and digraphs to exploit the higher simplicity of digraphs to detect deadlock combined with theoretic results known for Petri nets. The new PN-based prevention policies translate the information obtained from properly defined digraphs into empty siphons of a corresponding PN modeling the same system
An improved projection algorithm for direct adaptive fuzzy control
No full textAn effective variant of the projection algorithm frequently used in adaptive fuzzy control is presented. The proposed algorithm preserves the formal guarantees that the adapted parameters and some related system variables remain within the prescribed bounds, and improves the speed of convergence of the adaptation. The effectiveness of our approach is shown both in simulation and experimental case studies
A discrete-event system model for multi-agent control of automated manufacturing systems
No full textIn the area of multi-agent systems, many efforts have been recently devoted to find appropriate tools to model, and specify in a formal way the dynamics and the mechanisms of interaction of the various autonomous agents. In this paper, an approach based on the discrete event system specification technique is used. Typical agents used in manufacturing control systems can be viewed as discrete event systems, and analyzed with the proposed modeling tool. The models lend themselves both for developing a simulation platform, and for realizing the control software of the actual plant
Multi-agent fuzzy control of operation dispatching in flexible manufacturing environments
No full textThis paper describes a negotiation algorithm for task contracting in multi-agent manufacturing systems based on fuzzy techniques. In the proposed approach, operations on raw parts are executed on the workstation with a processing priority resulting from a negotiation between the agents controlling the machine and the parts flowing in the systems. Pricing and decision criteria encompass a set of variables taking into account different aspects of the available alternatives, as processing and setup times, workload in queues, buffer saturation or starvation. The proposed approach is compared with analogous strategies derived from literature on a simulation model of a hypothetical manufacturing system that was recently proposed as benchmark for multi-agent control systems
Recent developments in the application of computational intelligence to multi-agent manufacturing control
No full textThis paper surveys some recent approaches using Computational Intelligence (namely Fuzzy Logic and Evolutionary Algorithms) for manufacturing system control. In particular, the paper focuses on approaches oriented toward the Multi-Agent System design paradigms. Furthermore, this paper summarizes the results of a simulation comparison between-two fuzzy multi-agent architectures for job flow control, differing for the degree of interactions between agents, on a detailed case study that was recently proposed in literature to benchmark distributed multi-agent control approaches
- …
