section-59d2af4

STAGE OF APPLICATIONS 9SOL - SOFTGALAXY

  • Module M1.1 for response planning, establishing measures and post-incident action scenarios using multi-agent artificial intelligence algorithms. The following applications were made within this module:

 

1. Panic simulator:

The application realizes the optimization of multimer flows, in the case of disasters at events with high population densities in the open air and in closed spaces. Using this application, we follow the execution of evacuation scenarios by decreasing the population density in order to make efficient evacuations. The mathematical models used are: multi-agent Artificial Intelligence algorithms, discretization in time and space, realization of autonomous cell type models, mathematical modeling of each agent to act individually, the simulation is performed on static tactical field with probabilistic dynamics. The basic characteristics of the simultaneous panic are: settings - performs the property management for agents (simulation of natural persons), the simulated environment, the equation of motion (the behavior model of pedestrians) and the simulation speed; arena editor - allows the adjustment and manipulation of the environment by creating, eliminating or modifying the properties of the agents and the arena; automatic - a statistical tool that allows to change variables and visualize data.

The stage: the prototype was realized, the work scenarios established with the beneficiary and the integration in the final system will be implemented.

 

2. Application "escape route"

The application allows the optimal planning of an "escape route", avoiding obstacles. Optimal road planning applications are used in robotics. The implemented technique brings improvements in terms of traditional route planning and obstacle avoidance techniques, offering the mobility to change the course if new obstacles are discovered. The method used uses modern algorithms such as Hamiltonian logic with minimum weights, the A * algorithm for avoiding obstacles and global points of opportunity to update the optimal path globally using the Voronoi polygons.
Stage: 2 prototype applications have been developed, to be implemented on the working scenarios established with the beneficiary and integration into the final system.

 

3. Predictive application based on machine learning

The predictive application includes a variety of statistical techniques (data extraction, predictive modeling) that analyze current data (acquired in real time) and historical data to make predictions about future events. The defining advantage of this technical approach is that the predictive analysis offers a predictive score (probability) for each object, to determine, or influence the organizational processes.
The stage: a prototype has been developed, is to be implemented on the working scenarios established with the beneficiary (historical data and real-time data) and integration into the final system.

 

  • Subsystem Sb2.2hardware and sofware for capturing video streams from mobile operating sources, located on the ground, or taken over by different systems (smartphone, augmented reality glasses, body-worn cameras, cameras installed on vehicles), their integration and operative transmission of video images. In this module the following application was made:

4. Real-time object recognition application:

The application identifies real-time person-type objects, cars, furniture objects, animals, etc. The identification is done by using the deep learning algorithm of convolutional neural network type. Convolutional neural networks (Convolutional Neural Networks, short for CNN) are classifiers specially designed for working with images. Their structure facilitates the classification of data with a large number of parameters, because these types of networks allow the efficient processing of these categories of data. CNN does not have a predetermined architecture or structure, it adapts to the type of image being processed or to the nature of the objects to be classified.

The stage: the prototype was realized, the work scenarios established with the beneficiary and the integration in the final system will be implemented.

  • Module M2.1. accessing operational video resources virtual / augmented reality and blockchain. In this module the following application was made:

VR application: for the topography of the Casa Poporului area, with the possibility of monitoring the area in real time by an operator from the command center, purchasing real-time images from drones / Electro-optical systems. The stage: the prototype was realized, the working scenarios established with the beneficiary and the finalization of the VR system of analysis will be implemented.

  • Module M3.1. of integration with the digital table and of access and remote monitoring of the physiological data. The following applications were made within this module:

5. Applications of ECG and EEG biosignals analysis

or the analysis of ECG and EEG biosignals, we have developed a set of applications.

  • Pulse Rate application - this application monitors in real time the values of the human pulse, analyzes it by extracting values for the number of pulses, the average pulse of the time interval (s) and the average pulse rate.

  • ECG Feature Extraction application - is an application that extracts and analyzes the QRS signals from the ECG, so that we can determine the patient's normal clinical condition, tachycardia and hyperkalemia (metabolic disorders).

  • Fatigue / drowsiness analysis - The application is based on the analysis of brain waves of alpha, beta, delta and theta type and extracts the alternative eye movement so that we can predict whether the patient is cognitively alert, tired or drowsy.
    The stage: the prototype applications were realized, to be implemented on the ECG and EEG (real-time operation) sensors and the integration in the communications system.

  • Subsystem Sb2.3 for interconnection and data transmission between all participating entities take action and the command center using Blockchain technology. Within this module, the development of secure intra-organizational mobile communications (mobile and web) applications is underway.
  • Other developments - depending on the implementation of the hardware project.

6. Machine learning application for searching people / objects in real time from a set (reserve)

Is a machine learning type application, which allows the user to create a training template, which is searched in real time in a scene.

  • Pulse Rate application - this application monitors in real time the values of the human pulse, analyzes it by extracting values for the number of pulses, the average pulse of the time interval (s) and the average pulse rate.

  • ECG Feature Extraction application - is an application that extracts and analyzes the QRS signals from the ECG, so that we can determine the patient's normal clinical condition, tachycardia and hyperkalemia (metabolic disorders).

  • Fatigue / drowsiness analysis - The application is based on the analysis of brain waves of alpha, beta, delta and theta type and extracts the alternative eye movement so that we can predict whether the patient is cognitively alert, tired or drowsy.
    The stage: the prototype applications were realized, to be implemented on the ECG and EEG (real-time operation) sensors and the integration in the communications system.

  • Subsystem Sb2.3 for interconnection and data transmission between all participating entities take action and the command center using Blockchain technology. Within this module, the development of secure intra-organizational mobile communications (mobile and web) applications is underway.
  • Other developments - depending on the implementation of the hardware project.

7. Imaging application for drones (reserve)

Iis a set of specialized applications dedicated to the processing of images and video streams in real time, purchased from EO sensors (visible, thermal, IR, multispectrum) with applicability in Critical Infrastructure applications.

  • Pulse Rate application - this application monitors in real time the values of the human pulse, analyzes it by extracting values for the number of pulses, the average pulse of the time interval (s) and the average pulse rate.

  • ECG Feature Extraction application - is an application that extracts and analyzes the QRS signals from the ECG, so that we can determine the patient's normal clinical condition, tachycardia and hyperkalemia (metabolic disorders).

  • Fatigue / drowsiness analysis - The application is based on the analysis of brain waves of alpha, beta, delta and theta type and extracts the alternative eye movement so that we can predict whether the patient is cognitively alert, tired or drowsy.
    The stage: the prototype applications were realized, to be implemented on the ECG and EEG (real-time operation) sensors and the integration in the communications system.

  • Subsystem Sb2.3 for interconnection and data transmission between all participating entities take action and the command center using Blockchain technology. Within this module, the development of secure intra-organizational mobile communications (mobile and web) applications is underway.
  • Other developments - depending on the implementation of the hardware project.

Contact Us

Bulevardul George Co┼čbuc nr. 39-49, Sector 5, BUCHAREST, ROMANIA

(+40) 21 335 46 65 (+40) 21 335 46 68

Leadership

The entire project is runned on a close observation of Military Technical Academy