Salvatore Cavalieri

A novel learning algorithm which improves the partial fault tolerance of multilayer neural networks

The paper deals with the problem of fault tolerance in a multilayer perceptron network. Although it already possesses a reasonable fault tolerance capability, it may be insufficient in particularly critical applications. Studies carried out by the authors have shown that the traditional backpropagation learning algorithm may entail the presence of a certain number of weights with a much higher absolute value than the others. Further studies have shown that faults in these weights is the main cause of deterioration in the performance of the neural network. In other words, the main cause of incorrect network functioning on the occurrence of a fault is the non-uniform distribution of absolute values of weights in each layer. The paper proposes a learning algorithm which updates the weights, distributing their absolute values as uniformly as possible in each layer. Tests performed on benchmark test sets have shown the considerable increase in fault tolerance obtainable with the proposed approach as compared with the traditional backpropagation algorithm and with some of the most efficient fault tolerance approaches to be found in literature.

A neural network architecture for noise prediction

Abstract This paper presents application of neural networks (NNs) to the problem of prediction of noise caused by urban traffic. The most representative physical variable quantifying noise emissions is the equivalent sound pressure level. Up to now it has been identified on the basis of semi-empirical models, typically regression analysis, which generally do not provide very accurate approximations of the trend followed by sound pressure level. The authors have attempted to overcome this difficulty by adopting a neural approach based on a Backpropagation Network (BPN). Results obtained by the comparison of the BPN approach with those provided by selected relationships found in relevant literature, show how good is the approach proposed. The neural solution to the problem has shown the necessity, in certain phases, of a set of acoustic measurements which is as free as possible of error. The complexity of error identification by means of classical approaches has led the authors to explore the possibility of a neural solution to this problem as well. The authors therefore propose use of a neural architecture made up of two cascading levels. At the first level a supervised classifying network, the learning vector quantization (LVQ) network, filters the data discarding all the wrong measurements, while at the second level the BPN predicts the sound pressure level.

Optimization of acyclic bandwidth allocation exploiting the priority mechanism in the FieldBus data link layer

In process control applications, the bandwidth assigned to acyclic traffic is oversized in order to ensure that the time constraints of control traffic are met. In the FieldBus, a protocol designed to support the exchange of information between sensors/actuators and regulating elements in an automation environment, this means allocating an excessive portion of the macrocycle to acyclic traffic. The use of traffic management based on priority can allow more efficient management of the available bandwidth. The authors analyze the priority mechanism provided for in the FieldBus data link layer and assess various strategies for management of aperiodic traffic and different application scenarios. >

Pre-run-time scheduling to reduce schedule length in the FieldBus environment

The paper deals with the problem of scheduling the transmission of periodic processes in a distributed FieldBus system, defining the conditions guaranteeing correct transmission. The scheduling of periodic processes fixes the transmission times for each process in a table, whose length is equal to the Least Common Multiple (LCM) of all the periods. This involves great memorization problems when some periods are relatively prime. The authors identify the theoretical conditions which allow the length of the scheduling table to be drastically reduced, but still guarantee correct transmission. On the basis of the theoretical conditions given, the authors present a pre-run-time scheduling algorithm which determines a transmission sequence for each producing process within the desired scheduling interval. An online scheduling algorithm is also proposed to schedule new transmission requests which are made while the system is functioning. The reduction in the schedule length may increase the number of transmissions, thus reducing the effective bandwidth and increasing the communication overload. In order to make as complete an analysis as possible of the scheduling solution, the authors also present an analysis of both the computational complexity of the algorithms proposed and the communication overload introduced. >

A novel solution to interconnect FieldBus systems using IEEE wireless LAN technology

Abstract In various process control environments in which FieldBus communication systems are widely used it is often necessary to interconnect them. Unfortunately this interconnection raises several issues and the solutions found so far are not very satisfactory. In this paper, we propose a new solution which uses a wireless LAN based on the IEEE 802.11 draft standard as a backbone system. The problems of integrating the traffic generated by FieldBus systems, which notoriously features urgent time constraints, into IEEE 802.11 wireless LAN are dealt with and solved by proposing a simple modification to the 802.11 MAC protocol. Through simulation analysis, we then present a performance evaluation of the interconnection system and the cost of the modification proposed.

Optimal path determination in a graph by Hopfield neural network

Abstract Recurrent stable neural networks seems to represent an interesting alternative to classical algorithms for the search for optimal paths in a graph. In this paper a Hopfield neural network is adopted to solve the problem of finding the shortest path between two nodes of a graph. The results obtained point out the validity of the solution proposed and its capability to adapt itself dynamically to the variations in the costs of the graph, acquiring an “awareness” of its structure.

Meeting real-time constraints in CAN

This paper deals with the problem of information flow scheduling in a communication system based on CAN ISO IS-11898 physical medium access mechanism. It mainly features a bus access arbitration protocol based on a priority assigned to each message to be transmitted; if two or more messages are transmitted at the same time by different communication nodes, only the message with the highest priority continues to be transmitted, the other being stopped. In real-time applications, messages are relevant to process variables which must be transmitted within strict time constraints; according to the CAN ISO IS-11898 bus arbitration protocol, respect of real-time constraints depends on the priority assigned to each variable. The aim of the paper is to propose a procedure for dynamic assignment of priorities to variables to be transmitted, in such a way the relevant real-time requirements are fulfilled. Although many other approaches can be found in literature, the proposal is original as it is based on standard full CAN communication stacks.

WLAN-based Outdoor Localisation Using Pattern Matching Algorithm

Location detection is a very important issue in a wireless environment, allowing realisation of several useful services linked to the knowledge of the position of a mobile user (e.g. path optimisation, management of field crew in a large plant, providing a mobile robot with autonomous capabilities). Literature presents many approaches for location detection of a mobile device in a WLAN. One of these is based on the use of a pattern matching algorithm which provides for the position of the mobile device given the Radio Signal Strength (RSS) values received by it. Generally, localisation in WLAN environment is limited to indoor scenarios. Although GPS is commonly used for outdoor location detection, localisation in a WLAN communication infrastructure may be realised also for outdoor areas. The main advantage is a save in the hardware needed for the localisation, as it only requires the use of the WLAN communication card; further, the current limits of GPS, which may introduce not negligible errors in outdoor location detection, encourages the investigation of other techniques for outdoor localisation. On account of what said, the first aim of the paper is to investigate the feasibility of outdoor localisation of a mobile device in a WLAN environment; an IEEE 802.11b-based WLAN will be considered. Outdoor localisation has been realised using a pattern matching algorithm and the relevant performance, measured in terms of location errors, has been evaluated and compared with that commonly provided for by a GPS device. Then the paper will present a proposal for localisation, aimed to reduce the complexity typically featured by the pattern matching algorithm based approach when used for wide areas; the proposal is based on the use of a signal propagation model able to predict the RSS values available in each position of a IEEE 802.11b WLAN. Performance of the novel approach has been evaluated for outdoor localisation and the main results achieved will be shown in the paper. Finally, use of a Kalman filter for the location detection, will be presented in order to improve the performance of the pattern matching algorithm, making this technique very attractive for outdoor use.

Neural networks for process scheduling in real-time communication systems

This paper presents the use of Hopfield-type neural networks for process scheduling in the area of factory automation, where bus-based communication systems, called FieldBuses, are widely used to connect sensors and actuators to the control systems. We show how it overcomes the problem of the computational complexity of the algorithmic solution. The neural model proposed allows several processes to be scheduled simultaneously; the time required is polynomial with respect to the number of processes being scheduled. This feature allows real-time process scheduling and makes it possible for the scheduling table to adapt to changes in process control features. The paper presents the neural model for process scheduling and assesses its computational complexity, pointing out the drastic reduction in the time needed to generate a schedule as compared with the algorithmic scheduling solution. Finally, the authors propose an on-line scheduling strategy based on the neural model which can achieve real-time adaptation of the scheduling table to changes in the manufacturing environment.

Evaluating impact of security on OPC UA performance

The widespread use of OPC UA specification in client/server data exchange in industrial environment introduces many benefits as it allows to keep open the market of the industrial applications, due to the presence of a standard, worldwide and vendor-independent specification. On the other hand, the use of a software infrastructures (complex in the case of OPC UA) between industrial applications and devices (e.g. mainframes, PLC, microcontroller), may impact on the overall performance of the communication between industrial applications. The aim of this paper is to deal with the performance evaluation of OPC UA, pointing out all the main features which could influence performance in the client/server exchange of information. On the basis of the considerations about performance of OPC UA specification, outlined in the paper as said before, some results of a performance evaluation carried on by the authors will be presented; it will focus on the impact of OPC UA security mechanisms on the relevant performance.