Structural representation of the system architecture 

Contexts in source publication

Context 1

... in order to recognise the prediction criteria of the diagnoses of the emotional situations of disabled persons. The research results present further development of a multi-layered model of this framework, with integration of the evaluation of localization possibilities and decision support system constructions. The knowledge of decision support systems is represented by fuzzy neural control of the speed of two wheelchair-type robots working in real time providing movement support for disabled individuals. The method of fuzzy reasoning using fuzzy logical Petri nets [15] is described in order to define the physiological state of disabled individuals by recognition of their emotions. The proposed reinforcement framework is based on the interaction of intelligent remote bio-robots, localization services, embedded decision support systems and data stored in a data warehouse (Fig. 1). The data warehouse is based on distributed information systems with important personal data of the patients and sensor monitoring data. The framework includes the adaptive moving wheelchair-type robot which is remotely communicating with a wearable human affect sensing bio-robot. To record, for reasons of e-health care, relevant episodes based on humans affect stages [2], the context aware sensors are incorporated into the design of the Human Affect Sensing Bio Robot-x (HASBR- x ) for every disabled individual, and into the local Intelligent Decision Making Agent-x (IDMA- x ) for every intelligent support providing robot. This framework allows a multi-sensor data fusion before the transmission of the data to the Remote Control Server (RCS) to minimize the TCP/IP (UDP) bandwidth usage. Multi-agent based adaptive motion control of both robots is based on an adaptive Fuzzy Neural Network Control (FNNC) approach. The architecture of the FNNC controller represents an approach of Adaptive Neural Fuzzy Inference System (ANFIS) that combines the fields of fuzzy logic and neural networks [3] (Fig. ...

Context 2

... ability to learn about the nonlinear dynamics and external disturbances of the motor speed controller with a stable output, small steady error, and fast disturbance rejection is integrated into this framework. At the k th moment, the difference between motor speed reference value v(k) and motor speed output value v o (k) is split to speed error e(k) and speed error change Δ e(k). These values are used as proposed in [6] NN Learning Agent in Fig. 3 for learning the artificial neural network Artificial NN in Fig. 2 as well as the 2 nd order input vector of the Artificial NN . The output of the Artificial NN generates a percentage value of pulse width change Δ PW(k) to describe how much pulse width value PW(k) of the real motor speed control value at the moment k should be changed. This value then is generated in real time by the ATmega32 microcontroller to perform online calculating: PW ( k ) = PW ( k − 1 ) + Δ PW ( k ) . ...

Context 3

... case shown on Figure 2. It shows an entry ramp onto a freeway. Vehicle A and vehicle C cannot see each other and this may lead to a collision since vehicle A wants to enter the lane. But if the vehicles can exchange data among them, then this can be avoided. As soon as vehicle A nears the ramp it becomes a part of the ad hoc network and joins vehicle B and C. So now vehicle C knows that there is a vehicle entering from the ramp and it can adjust its speed to let vehicle A merge in between itself and vehicle B. The traffic patterns at street corners are unpredictable and can often lead to dangerous situations. On Figure 3, vehicle A wants to overtake vehicle B. Vehicle A cannot see vehicle C. If vehicle A changes the lane to overtake vehicle B it may collide with the approaching vehicle C. This could be avoided if the vehicles can exchange data. If vehicle B is equipped with vision sensors, it can sense vehicle C and send a signal to vehicle A that vehicle C is approaching. So vehicle A will not overtake and the collision will be avoided. The data flow from vehicle C to vehicle A has to pass through vehicle B. Even though the common ITS infrastructure based safety system may exist at certain intersections, still, a networked approach for communication among vehicles can increase the level of safety. On Figure 4, a traffic scenario at a road intersection is shown. Assume that the signal for the horizontal lanes has been turned green and consequently the signal for the vertical lanes is red. Vehicle B wants to make a right turn while vehicle A wants to proceed straight from left to right. Neither of the vehicles can see each other due to vehicle C. If the vehicles are able to communicate with each other, then vehicles A, B, D, E and F could form a sensor network and could communicate with each other. A and B can use their sensed data and data from other vehicles to avoid a collision even though vehicle C is blocking the direct line-of- sight view between them. Each vehicle would know the position and the velocity of the other vehicle using area coverage techniques and would take appropriate measures to avoid the accident and cross ...

Context 4

... into consideration everything mentioned above, the authors offer a filtering method, which helps to improve SPAM filtering indices. The offered method is based on the following postulates: • E-mail filtration algorithm modification; • Modified shingles algorithm; [1] • The use of the notions of “white”, “grey” and “black” lists; The block-scheme of offered algorithm of letter testing is given on Figure 2. This solution lets decrease the system recourses usage. As well gives a possibility to test new rules for SPAM blocking. For more effective filtration of SPAM it is suggested to use queues. When e-mail comes to server it will be queued while one of events will not happen – either the determined amount of letters is collected or it is passed the sufficient amount of time. Combining of these terms is not eliminated also. For the analysis of letters, which are in a queue, it is very convenient to use the shingle algorithm. This algorithm will help to expose similar letters and find a SPAM and even polymorphic SPAM. Realization of algorithm implies a few stages: • text canonization; • breaking up the text on shingles; • checksums finding; • search of identical sub lines; In the algorithm of shingles the comparison of the texts checksums is realized. As it is generally known, the checksums of the static functions are very sensitive to the changes. Therefore, at a search almost of doublet, all exclamation marks, points, commas and other will be deleted from text. Only words will be left. Now it is necessary to break up each of the received texts on sub lines – shingles. Further, checksums can be found from marked out shingles. Secondly, must be determinate how they must go – lapped (that is how exactly "shingles" are obtained), or butt [1]. To create a sample – the classical algorithm written by Broder suggests choosing either a fixed number of shingles with a minimal value or all shingles whose value can be divided into some small number (10-30). In the first case we get fixed size sample and large-sized set of shingles, but we cannot judge about the subdocument. In the second case the number of shingles is proportional to the size of the document, but it is possible to evaluate by the sample of shingles the attachment of documents to each other or the percentage of their intersection. Finally, the last algorithm generates a fixed size sample which size is determined by a given number (e.g. 85) of different independent random functions, each of which store exactly one shingle, which is the minimal by the value of the checksum. This approach combines the advantages of the previous two [2]. There is a problem with short documents, for which the algorithm of the shingle selection may not select a single suitable, or to choose too little. There are two solutions: one of them: to circle text of the document, that is virtually continuing its beginning after end. The second approach is to use a sample, size of which has a logarithmic dependence on the size of the document. If, for each letter to select more than one shingle, we are confronted with the task of identifying documents that have only a few matching shingles. If we reduced the number of shingles, it still would be a nontrivial amount of work. Therefore, over a set of shingles of the document, it is possible to calculate another checksum, the so-called "super shingle” ...

Context 5

... to the requirements, a system engineer has developed the following structural representation of the architecture based on Three-tier style (see figure 2). Note that the object domain remained the same that was used in the example with SOA, but in this sample the different architectural style is used. The Application tier contains the functionality for receiving and processing information regarding roads and traffic congestion status. The Data tier has the database which stores traffic congestion related information, also this tier contains map data which is provided by third-party supplier. Just as in the example with SOA, the system provides access to traffic congestion information with the help of Web applications and Web services. Web application is located in Presentation tier and serves for obtaining and adding information regarding some roads. Let’s calculate the metrics values for the Application tier. This tier contains the functionality for road status processing. Also it uses the underlying tier for storing data and provided processed data to the overlying tier. So this tier has: ● three external inputs with average complexity, which are necessary to obtain information regarding roads state; ● five external inputs with high complexity, which are used for the database and map data accessing; ● four external outputs with average complexity, which are used by the overlying tier and external systems for obtaining information regarding traffic congestions; ● two external outputs with a high degree of complexity, which are used for adding information into the database; ● one external interface file with a high degree of complexity, which is used by the overlying tier for obtaining a list of traffic congestions. According to the data listed above, let’s calculate the values of the ...

Context 6

... ( ) – pixel brightness value in point ( ) of the first frame; ( ) – pixel brightness value in point ( x, y ) of the second frame; ( y ) – the resulting pixel brightness value in point ( x, y ). Example of algorithm usage is shown on Figure ...

Context 7

... P(A) and P(B) – possibilities of event (B) occurrence on filtering entrance of both regular letters and SPAM; P(B/A), P(A/B) – conditional possibilities of false filter activation. The results obtained in formula (1) and on scale model have divergence of on more than 5-7%; Applying the notion of “white”, “grey” and “black” lists lets solve a number of problems while filtering. For example, creation of dispersed database of “white” lists simplifies the usage of “grey” lists. Decreases the load on server and improves “false negative” indices. Block-scheme of the offered generalized SPAM filtering algorithm is given on Figure ...

Context 8

... Figure 2 illustrates influence of communication overhead on the efficiency indexes of a computer system at distributed solution of the before considered divide and concur algorithm. Value on a vertical axis corresponds to absence of communication overhead, i.e. to the parallel system. The issue of the day in intelligent transportation systems development still is a task of effective selection and design of the used software and hardware tools. Within the framework of decision of this task, authors has analysed characteristics of given problem domain, singularities of structural organization and functioning for this family of computer systems, has explored their specificity, has offered a set of metrics for the estimation of efficiency of parallel and distributed modes of operation. These metrics allow estimation of integral efficiency of intelligent transportation systems, and taking into account the basic features of both parallel and distributed ...