Rainer Blind

Analysis of Networked Event-Based Control with a Shared Communication Medium: Part I – Pure ALOHA

Abstract In recent control literature, it was shown that event-based control requires less samples than time-triggered control. Consequently, it is argued that event-based control is superior to time-triggered control for Networked Control Systems (NCS) because it reduces the network utilization. The aim of this work is to show that it is not that simple. Therefore, we compare event-based and time-triggered control for a NCS consisting of subsystems with integrator dynamics and a shared communication medium, where packets are only lost due to collisions. We show that packet losses due to collisions drastically reduce the performance of event-based control if packets are transmitted whenever the event-based control generates an event. In contrast to previous findings, it turns out that in this setup time-triggered control is superior to event-based control.

Optimized Resource Dimensioning in an embedded CAN-CAN Gateway

The Controller Area Network (CAN) is a robust, low-cost, and simple event-triggered technology for connecting electronic control units in the manufacturing industry and vehicles. Todays real-time control systems are distributed over a multitude of CAN systems (domains) which are connected via embedded gateways. A failure or overload situation in the gateway can affect several domains. Furthermore, gateways often become bottlenecks between the domains and in the case of CAN buses they can pose further problems with respect to the priority-based network access method. Due to this access method and the limited resources (e.g., buffer capacity) embedded CANCAN gateways have to be dimensioned accurately. Otherwise, unacceptable processing delay and message loss within the gateway can occur. The main contribution of this paper is to investigate the optimized dimensioning of an embedded CAN-CAN gateway with regard to minimizing gateway resources in terms of processing and buffer capacity and decreasing message loss at the same time. For that purpose a CAN bus and a gateway model are described and used to investigate scenarios with two domains connected via a gateway.

Optimal and optimal-linear control over lossy, distributed networks

Abstract We consider an optimal control problem for networked control systems, where the loop is closed via a lossy, distributed network with an acknowledgment mechanism. The network is distributed in the sense that there are different sets of sensors and actuators that each communicate individually with the controller. We assume that all packets, i.e., the measurement packets, the control packets and the acknowledgment packets are sent over the lossy network and thus are subject to loss. We derive suboptimal controllers with respect to a quadratic cost criterion for the general case and optimal controllers for the case that all states are perfectly measured over a single link. Additionally, we present stability criteria for both cases.

On the optimal sending rate for Networked Control Systems with a shared communication medium

In this work, we consider the optimal sending rate for Networked Control Systems (NCS) with different communication schemes. Therefore, we compare time-triggered control with a Time Division Multiple Access (TDMA) or a Frequency Division Multiple Access (FDMA) communication protocol and event-based control with an ALOHA or Carrier Sense Multiple Access (CSMA) communication protocol. Our main interest is the optimal load and optimal performance of such systems. Nevertheless, we are also interested how the optimal performance scales with the number of agents, whether or not it is possible to add a new agent to an already running system, and the effects of an overloaded communication system.

Is it worth to retransmit lost packets in Networked Control Systems

Recent publications showed that packet loss significantly reduces the performance of Networked Control Systems (NCS). One approach to increase the reliability of the communication system is to retransmit lost packets. Obviously, such retransmissions take time and thus require to increase the sampling time. Increasing the sampling time generally degrades the performance of a control system. Thus, we raise the question whether it is worth to increase the sampling time to allow more packet transmissions per sampling interval.

Exact convex formulations of network-oriented optimal operator placement

Data processing tasks are increasingly spread across the internet to account for the spatially distributed nature of many data sources. In order to use network resources efficiently, subtasks need to be distributed in the network so data can be filtered close to the data sources. Previous approaches to this operator placement problem relied on various heuristics to constrain the complexity of the problem. In this paper, we propose two generic integer constrained problem formulations: a topology aware version which provides a placement including the specific network links as well as an end-to-end delay aware version which relies on the routing capabilities of the network. A linear programming relaxation for both versions is provided which allows exact and efficient solution using common solvers.

Estimating the fates of the control packets for Networked Control Systems with loss of control and measurement packets

In this work, we extend recently published work for the estimation of fates of the control packets in networked control systems (NCS). Unfortunately, the existing scheme is limited by the assumption that all measurement packets arrive. Moreover, a rank condition must hold, which is identical to a relative degree one assumption for SISO systems. In this work, we relax these assumptions and first extend the approach to NCS with measurement losses. Moreover, we present an approach to estimate the fates of the control packets for SISO systems with an arbitrary relative degree.

Robustification and optimization of a Kalman filter with measurement loss using linear precoding

This paper considers a linear precoding for a Kalman filter which receives the measurements over an erasure channel. We show how to design the precoding matrix that either robustifies the Kalman filter with respect to packet losses or reduces the estimation error. The effectiveness of the proposed method is demonstrated by two examples.

The performance of event-based control for scalar systems with packet losses

In this work, we analyze event-based control for scalar systems with an impulsive input that resets the state to the origin when applied. The event information is transmitted within packets, which are randomly dropped. We derive the cost as well as the expected times between events and discuss how to change the bounds, which generate the events, after a packet loss.

Optimal control over unreliable networks with uncertain loss rates

In this work, we consider an optimal control problem for Networked Control Systems (NCS), where the loop is closed via an unreliable network. In order to inform the controller about the fates of the control packets, the actuator sends acknowledgements to the controller. In this work, different types of acknowledgements are considered. Since all packets are send over a network, we consequently assume, that all packets can get lost, i.e. the measurement packets, the control packets as well as the acknowledgement packets. The loss of these packets is modeled by independent and identically distributed (iid) Bernoulli processes. Finally, we assume that the parameters of these loss processes are not known exactly. Using this setup, we derive bounds for the uncertainty in the loss rates. Moreover, we present an algorithm to design a sub-optimal controller which can tolerate a given amount of network uncertainties.