Octavian STEFAN, Toma-Leonida DRAGOMIR
Politehnica University Timisoara,
2 Piata Victoriei, Timisoara, 300006, Romania
octavian.stefan@upt.ro, toma.dragomir@upt.ro

Abstract: The current study proposes a control-aware Quality of Service adaptation co-design method for networked control systems. The novel networked control structure is based on a remotely placed Quality of Service adapter that continuously changes the network parameters using the Next Steps in Signaling protocol suite for end-to-end Quality of Service. The control system’s design is developed gradually and the system’s stability is assessed by considering the overall system as a switched linear one. Finally, the results are validated on a numerical example.

Keywords: Networked control systems, co-design control methods, switched linear systems.

>Full text
CITE THIS PAPER AS:
Octavian STEFAN, Toma-Leonida DRAGOMIR, A Control-aware QoS Adaptation Co-design Method for Networked Control Systems, Studies in Informatics and Control, ISSN 1220-1766, vol. 24 (1), pp. 33-42, 2015. https://doi.org/10.24846/v24i1y201504

1. Introduction

Advances in digital communication technologies offered new possibilities for the development of telecontrol applications. Consequently, networked control systems (NCS) have gained an increasing amount of attention in the scientific community. Although NCS have several advantages ([11]), there are some network induced issues like time-varying delays, data loss, and limited data transfer capacity which can affect the system’s performance and stability ([24]). In order to overcome these issues, different control strategies are proposed in the specialized literature that can be divided in two categories: control over network solutions (e.g. gain scheduling ([15]), adaptive Smith predictor ([14]), optimal stochastic control ([18]), event based control ([12]), predictive control ([5]), communication disturbance observer ([17]) or robust control ([24])) and control of network solutions (e.g. end-to-end Quality of Service (QoS) ([8])). Although each category is proven to work in practice, best possible results are achieved when using co-design methods obtained by combining control solutions from both categories (best resource utilization) ([21]). Furthermore, when using a shared medium network to transport data for multiple NCS and other possible applications with an unknown traffic pattern, co-design methods are mandatory to assure the control objectives. According to [1], two types of co-design methods exist. The first one, referred to as “implementation-aware control law design”, presumes real-time continuous adaptation of the control parameters according to the ones of the network (e.g. continuous adaptation of the sample rate ([2]) or continuous adaptation of the controller gains ([22]) based on the values of the network time delays and packet loss number). The second one, named “control-aware QoS adaptation” presumes a real-time reallocation of network resources, by modifying the QoS parameters, in order to maintain the quality of control (e.g. dynamic bandwidth allocation in a Switched Ethernet Network ([9])).

Current paper proposes a new control-aware QoS adaptation co-design method for NCS using the Next Steps in Signaling (NSIS) protocol suite for end-to-end QoS. In addressing the control objective (tracking and stabilization) a networked control structure is considered, composed out of a local plant, a remotely placed controller and a network adaptation block. Based on the network parameters’ values, the network adaptation block performs a real-time continuous adaptation of the QoS parameters in order to assure the control objectives and to minimize the network resource utilization.

For the analysis stage, the study uses a switched linear system ([20]) as a network transmission model (NTM) – presented by the authors in [22] – that completely characterize network transmissions from an input-output perspective, taking into account time-varying delays, packet losses and irregular situations together with their handling strategy ([23]) that can occur when using unreliable networks and connectionless protocols.

The remainder of this paper is organized as follows. Section 2 describes the problem formulation. Section 3 presents the NTM. Section 4 presents the design methodology for the control structure. Section 5 analyses the system’s stability. Section 6 presents an illustrative example and Section 7 states some final conclusions.

REFERENCES

1. AUBRUN, C., D. SIMON, Y.-Q. SONG, Ed., Co-design Approaches for Dependable Networked Control Systems, Wiley, 2010.