Convexified Graph Neural Networks for Distributed Control in Robotic Swarms

A network of robots can be viewed as a signal graph, describing the underlying network topology with naturally distributed architectures, whose nodes are assigned to data values associated with each robot. Graph neural networks (GNNs) learn representations from signal graphs, thus making them well-suited candidates for learning distributed controllers. Oftentimes, existing GNN architectures assume ideal scenarios, while ignoring the possibility that this distributed graph may change along time due to link failures or topology variations, which can be found in dynamic settings. A mismatch between the graphs on which GNNs were trained and the ones on which they are tested is thus formed. Utilizing online learning, GNNs can be retrained at testing time, overcoming this issue. However, most online algorithms are centralized and work on convex problems (which GNNs scarcely lead to). This paper introduces novel architectures which solve the convexity restriction and can be easily updated in a distributed, online manner. Finally, we provide experiments, showing how these models can be applied to optimizing formation control in a swarm of flocking robots.

Pitch Transcription of Solo Instrument Tones Using the Autocorrelation Method

Pitch transcription is basically identifying and copying pitch on an audio or music. In this case, the pitch of a solo instrument music is processed to find the composition of the music tones, the method used is autocorrelation. After processing, the system will produce pitch transcription results from the audio that has been processed. this research done digitalization of an old method of identifying pitches into an application that is able to transcribe pitches in an audio. This Pitch Transcription application was created using the Python programming language with Librosa library as a library for audio processing. The purpose of making this system is to facilitate the identification of pitch from a solo instrument music. This Application Feature itself besides being able to show the results of pitch transcription can also display the Onset Graph, Signal Graph play the results of the synthesis transcribed audio sounds. Testing in this study uses audio sourced from 4 single instruments, there are flute, piano, violin and acoustic guitar. The test results show that the implementation of the autocorrelation method in the solo instrument tone transcription application has an accuracy of 92.85%.

Construction of open-loop electromechanical system fundamental matrix and its application for calculation of state variables transients

The article considers the methods of calculating the transition matrix of a dynamic system, which is based on the transient matrix representation by the matrix exponent and on the use of the system signal graph. The advantages of the transition matrix calculating using a signal graph are shown. The application of these methods to find the transition matrix demonstrated on the simple electromechanical system example. It is shown that the expression for the transition matrix as a matrix exponent completely corresponds to the expression found by means of the inverse matrix and based on the use of the signal graph. The transient matrix of a dynamical system thus found as a matrix exponent can be used to analyze processes in a system that is described by differential equations with integer derivatives. The formation of a transient matrix for the analysis of system processes, which is described by equations with fractional derivatives, is also considered. It is shown that the description of processes in systems with fractional derivatives based on the transient matrix and the representation of the fractional derivative in the form of Caputo-Fabrizio makes it possible to study coordinate transients without approximations in the description of the fractional derivative.

Resilience of Urban Technical Networks

The need to overcome the insulated treatment of urban technical infrastructures according to the nature of the transferred flows is argued. The operation of urban technical networks is affected by endogenous and exogenous random events with consequences for users. By identifying these operational risks and the difficulties of estimating the impact on the performance of the urban technical networks, the authors chose to study the risk management through a concise expression—in relation to the engineering resilience and its connections with vulnerability. Further, the research is confined to the case of urban traffic networks for which the resilience is expressed by the capabilities of these networks of resistance and risk absorption (both motivated by the redundancy in design and execution). The dynamics of the network, in correlation with the resistance and absorption capacities, is introduced by three states for which the signal graph is built. In a stationary regime, the probability of each state is computed. These probabilities allow the calculation of the entropy of the network, relevant for assessing the preservation of the network functionality.

Design and Evaluation of a MEMS Magnetic Field Sensor-Based Respiratory Monitoring and Training System for Radiotherapy

The patient’s respiratory pattern and reproducibility are important factors affecting the accuracy of radiotherapy for lung cancer or liver cancer cases. Therefore, respiration training is required to induce respiration regularity before radiotherapy. However, the need for specialized personnel, space, and time-consuming training represent limitations. To solve these problems, we have developed a respiratory monitoring and training system based on a micro-electro-mechanical-system (MEMS) magnetic sensor. This system consists of a small attaching magnet, a sensor, and a breathing pattern output device. In this study, we evaluated the performance of the signal measurement in the developed system based on the various respiratory cycles, the amplitudes, and the position angles of the magnet and the sensor. The system can provide a more accurate breathing signal graph with lower measurement error and higher spatial resolution than conventional sensor methods by using additional magnet. In addition, it is possible the patient to monitor and train breathing himself by making it easy to carry and use without restriction of time and space.

Investigation of the Indirect Hypercube as Natural Architecture for Parallel Algorithms of a Transpose Type for FFT and Other Fourier-Related Transforms

The natural architectures are architectures, derived from the signal graph of the corresponding algorithm. That is why they are considered to be the most appropriate architectures for parallel realization of this algorithm. For Fast Fourier Transform algorithm (FFT) two types of natural architectures are known – the direct and the indirect hypercube. The direct hypercube has been investigated and analyzed a long time ago. The development of the concept of Indirect Hypercube, although quite old, is too difficult, controversal and still unfinished. Fast Hartley transform (FHT)/Real-valued Fast Fourier transform (RFFT) algorithms are important Fourier-related transforms, because they lower twice the operational and memory requirements when the input data is real-valued. These types of algorithms, however, have an irregular computational structure, which makes their parallel implementation a very difficult task. The aim of this paper is, based on the results achieved so far, to present further development of the concept Indirect Hypercube. A method of parametric synthesis of an indirect hypercube is described as a model of parallel FFT algorithms of a transpose type with different granularity/radix. This method is generalized for relevant RFFT/FHT and FCT algorithms. Two types of SIMD array architectures are described (radix-2 and radix-4), based on the indirect hypercube concept. These architectures are implemented as fast FFT/RFFT/FHT processors for real time applications. The performance estimation, as well as the estimation of resource utilization is carried out.