Electronics and Artificial Intelligence

Electronics and Artificial Intelligence

The ITCL’s Electronics and Artificial Intelligence department has wide background on working with companies. This area uses emerging technologies as Machine Learning, Big Data, Data Science, Blockchain, Neural Networks and Artificial Intelligence, which allow handling massive amount of data coming from heterogeneous origins. Thus, our Electronics and Artificial Intelligence department extracts value, in terms of information, from huge amounts of information coming from different environments, facilitating its use in innovative business models.

We are experts in the design and prototyping of advanced electronic boards and devices for its integration into capital goods and telemedicine. Likewise, we develop personalized algorithms and design experiments for characteristic selection, process optimization, dynamic series clustering and output prediction models for data analysis.

Additionally, we develop smart systems for high precision agriculture, smart cities and different types of infrastructure.

This area has a wide working experience with national enterprises, having been part of several R&D and process improvement projects, highlighting the development of smart systems, advanced communication and microelectronics in environments where these technologies are key factors. (Smart Cities, Smart Energy, Industrial Internet of Things –IIoT, Industry 4.0, Factories of the Future – FoF, Machine to Machine Technology – M2M).

Some strands of work the Electronics and Artificial Intelligence department are the following

  • Design and prototyping of electronic boards and advanced electronic devices for its integration in capital goods and telemedicine
  • ARM Solution Design
  • Programming of embedded systems under Linux
  • Programming of microprocessors for data acquisition, control and advanced communication.
  • Data analysis in Smart Systems: Personalized algorithmic designs and design of experiments. Characteristic selection, process optimization based on characteristics, clustering of dynamic series of data and output prediction models.
  • Specific developments for environmentally friendly mobility: In-house and collective charging devices, analysis of charge demand, energy distribution according to restrictions, localization and control systems, carpooling, car sharing, route optimization algorithms.
  • Smart systems’ design for high-precision agriculture, infrastructures and cities.

MEMBERS OF THE R&D GROUP

The Electronics and Artificial Ingelligence team consists of young researchers, led by Dr. Javier Sedano, and it is one of the most dynamic and prestigious groups of ITCL, with several patents developed.

Ph.D. Javier Sedano: is an expert in the development of artificial intelligence, electronic systems (hardware) and industrial control systems, as well as in the design of connectionless models for the identification and modelling of dynamic systems in Big Data and Blockchain. He is the head of R&D of the ITCL. Javier Sedano also participates in other research groups and has spent more than 25 years working on projects and publications related to artificial intelligence and system modelling. He collaborates in the organization of international scientific conferences, program committees and organizations, etc. Also, he is an active member of the IEEE Systems, Man & Cybernetics Chapter Society Spanish Chapter and collaborates in the organisation of international congresses, programme committees and organisations such as SOCO, HAIS, IDEAL, NABIC and ISD. Over his career, he has published more than 100 international publications and holds records of software and industrial patents.

 

R&D in electronic design and artificial intelligence

Apart from its activity of conducting applied research and industrial developments under technical collaboration or subcontracting, the Electronics and Artificial Intelligence department offers the design of prototypes as well as its manufacturing and the management of the product’s industrialization, including tests, validation and CE certification.

R&D SERVICES IN APPLIED ELECTRONICS AND ARTIFICIAL INTELLIGENCE

Services for Industry 4.0

One of the objectives of industry 4.0 is to manufacture in a different, flexible and efficient way, so that the integration of intelligent systems and data analysis can improve productivity and operations management, as well as the development of new lines of business.

To this end, the ITCL Electronics and Artificial Intelligence unit has launched a portfolio of specific services in Industry 4.0 for productive companies and manufacturers of capital goods, which allow the start of innovation projects adapted to the needs of each of them.

Consult here the services Industry 4.0 that we offer you Services industry 4.0

Electronic prototype factory

Design, programming and manufacture of electronic prototypes

ITCL offers its customers an electronic design service both hardware and software level, which provides engineers at your service, with the ability to design and program electronic solutions tailored to their needs.

The design group is formed by the Research Group of Applied Electronics and Artificial Intelligence ITCL, composed of a qualified technical staff of professionals.

ITCL has in its facilities the necessary equipment to carry out all the necessary stages in the manufacture of electronic prototypes, from the initial design phase, through the milling and gluing of the components, to the encapsulation or resining of the final component.

We offer

Manufacture of prototypes and short series according to the specifications established by the client:

  • Component selection
  • Schematic design
  • Routing of the PCB and generation of Gerber files for later manufacture
  • Selection and programming of the most suitable microcontroller for each application
  • Manufacture of the prototype
  • Tests and trials in certified laboratories

Electronic systems

Microcontrollers, Wireless Communication, Communication Protocols, Positioning, Sensors and HMI(Human Machine Interface)

 

The ITCL Applied Electronics Group has the technical, human and laboratory capacity for the development of electronic systems. The group is capable of tackling any of the phases involved in the design of a new electronic system, ranging from the selection of components, electronic design, prototyping and the testing of prototypes according to requirements.

In addition, the systems are studied to be implemented according to the needs of each client, being able for example to design microelectronics (PCB’s of reduced dimensions), use various types of communication (by cable or wireless), design the electronics to be integrated within a superior system, embed Soft Computing or develop sensorics according to the needs, etc.

Systems we work with

  • Microcontrollers: different architectures and manufacturers (Microchip, Atmel, ARM (freescale, Raspberry).
  • Wireless communication: Zigbee, Bluetooth v2.0 and v4.0, radio frequency (434 Mhz), Wi-Fi, GPRS, RFID Mifare.
  • Communication protocols: RS-485, RS-232, I2C, SPI, CAN bus, PLC, UART, TCP/IP.
  • Positioning: GPS, GPRS.
  • Sensors: temperature (thermocouple, PT100, RTD), light sensor, load cell for weights, motion (accelerometers, gyroscopes), biosensors, potential-free contacts, tactile sensors, skin conductivity, infrared, etc.
  • HMI (Human Machine Interface): light and acoustic indicators, displays (LCD, graphics, alphanumeric, etc.), keyboards.

Software based on intelligent algorithms

Soft Computing Solutions

ITCL offers different Soft Computing solutions, which focus on the incorporation of automatic learning in hybrid systems, data analysis, feature selection, system modelling, classification and optimisation of systems. It also includes solutions for bioinformatics systems processing, clustering, classification and dynamic series modelling applied to gene expression profiles.

Soft Computing for the classification of dynamic series applied to gene expression profiles.

Sot Computing of our own design that includes different proprietary algorithms for the clustering of co-expressed genes in microarray data analysis (MDA). Suitable for use by researchers trying to determine the important genes and co-expressed relationships between them for large dynamic datasets to optimise an output feature.

  • Shape Index (SC). Clustering without taking into account the output of each sample.
  • Output Shape Index (SOC). Clustering taking into account the correlation of the gene with the output.
  • Dynamic Shape Index (DSC). Dynamic version of the SC method.
  • Output Dynamic Shape Index (DSOC). Dynamic version of the SOC method.
  • Relaxed Shape Index (RSC). Extension of the SOC method

The software integrates fusion methods that combine into a single cluster the clusters created from each of the independently performed time series of microarray data. The detection of the most important clusters within the possible clusters is performed using different measures on the genes, including the Information Correlation Coefficient (ICC), Pearson Correlation Coefficient (PCC) and Shape Increase measures.

Optimisation of machine variables based on the model.

Soft Computing systems of our own design, for the optimisation of machine parameters in manufacturing processes including the development of models to evaluate the behaviour of the machine variables in the process and to find the suitability function to be optimised.

This process is carried out firstly by obtaining the parametric or black box objective function of the system to be optimised. Subsequently the optimisation of the machine parameters is carried out by means of single-objective and multi-objective Genetic Algorithms, simulated annealing, or ant colonies.

An example of objective searches on a manufacturing system with three objectives is presented below.

Vehicle routing optimisation systems

Design of different hybrid evolutionary algorithms for route planning for several vehicles, minimising the costs associated with transport, such as: minimising the total transport time, minimising the total distance travelled, minimising the waiting time, minimising the number of vehicles to be used, etc. All this while satisfying a number of multiple constraints and limitations.

 

Main elements: 

  • Customers/delivery or collection points
  • Warehouse/Warehouses
  • Vehicle fleet (capacity of each vehicle)
  • Transport network
  • Function to minimise

Modelling:

Modelling the optimisation problems as different types of VPR (Vehicle Routing Problem):

  • VRP with Pickup and Delivery (VRPPD): some customers represent pickup locations and others represent delivery locations.

     

  • VRP with Time Windows (VRPTW): customers have time windows in which they have to be served.

     

  • Capacitated VRP (CVRP): vehicles have limited capacity.

     

  • Distance-Constrained VRP (DVRP): Maximum length allowed for a route.

     

  • Multiple Depot VRP (MDVRP): multiple depots to supply customers.

Since VRP is a complex combinatorial optimisation problem (NP-hard), different heuristic techniques have been developed to solve the different VRPs: Genetic Algorithms, Swarm Intelligence and Ant Colony Systems.

PRODUCTS DEVELOPED IN APPLIED ELECTRONICS AND ARTIFICIAL INTELLIGENCE

VE recharging station in domestic environments

ITCL has developed EV charging stations for indoor use.

Frontal del punto de carga para vehículo eléctrico

Charging point front for electric vehicle

The system offers two versions according to the user’s needs:

  • Multipoint: intended for systems where you want to monitor, control a network of chargers under the international standard of communication with OCPP v1.5 server (standard and open protocol). The compatibility of OCPP guarantees its operation with OCPP servers. User identification is done by means of a Mifare card (RFID).
  • Autonomous version: more economical and simplified version that allows the vehicle to be recharged without the station having to be connected to any server. The user can start recharging either by pressing a button or by using a key.

This product can be adapted with other forms of recognition by the user, implementing RFID or Bluetooth functionality, so that the user is validated either with RFID card or with mobile APP.

The recharge points can be connected to the network via an RJ45 connector and all access data (user databases, points, history, etc.) are stored on an OCPP server.

The WEB application that manages the system is accessible from different users and through different devices; personal computer, mobile, Tablet, etc.

Arquitectura Multipunto

Multipoint Architecture

Characteristics:

  • The stations are compatible with charging mode 3 according to UNE-EN 61851.
  • Charging socket type 2 according to UNE-EN 62196.
  • Socket lock management that prevents the connector from being removed during the charging process.
  • LED status indicators located on the front: FREE, CHARGING, CONNECT, OFF, FAULT…
  • User identification through MIFARE reader located on the front.
  • Voltage and max output current: 230VAC/32A.
  • Possibility of output with hose (electrolinear format).

Functionalities of the management application

  • Compatible with OCPP v1.5 points
  • User management: additions, deletions, modifications
  • Monitoring of the state of the points.
  • Registration of load consumptions.
  • Consultation of histories.
  • Recording of incidents.
  • Visualization in map of the location of points.
  • Reservation of points by hours.

Potentioztatic devices for biosensors

Autonomous potenciostatic device based on biosensors and signal conditioning and treatment electronics for the detection of acids, sulfanamides, benzoles, etc.

This electronics consists of a proprietary design of mixed analog and digital type that consists of a signal conditioning part provided by the electrodes of the biosensor, an amplification stage, and a microcontroller with its embedded software that is responsible for the digitization of the signal and its analysis.

biosensor

An example of use is made as a compact and autonomous solution for the detection of Benzoylmethylecgonine (cocaine in aqueous solution).  It consists of a biosensor that reacts to the presence of the substance in the saliva by generating a signal that is treated and analyzed through the associated electronics. The detection time depends on the biosensor used.

Characteristics

  • Low cost
    placa potenciostato
  • Compact and small size
  • Stand-alone operation
  • Illuminated indication of the presence of cocaine in the sample
  • Microcontroller 8 bits
  • Integrated Software
  • 10-bit analog-to-digital converter
  • Processing of very small signals (nanoamperes)
  • Low consumption. Powered by 3.6V battery.
  • Touch sensor ignition
  • Automatic device shutdown

Application

Security, Public health, Environment, Legal Medicine

Software for gene clustering

SOFT COMPUTING FOR DYNAMIC SERIES CLASSIFICATION APPLIED TO GENE EXPRESSION PROFILING

Proprietary soft computing that encompasses different proprietary algorithms for clustering co-expressed genes on microarray data analysis (MDA). Suitable for use by researchers trying to determine the important genes and co-expressed relationships between them for large dynamic data sets in order to optimise an output feature.

Presentation of the laboratory software
Soft Computing application manual for gene clustering
Some of the designed algorithms integrated in the software are:

Shape Index(SC). Clustering without taking into account the output of each sample.
Output Shape Index (SOC). Clustering taking into account the correlation of the gene with the output.
Dynamic Shape Index (DSC) Dynamic version of the SC method.
Output Dynamic Shape Index (DSOC). Dynamic version of the SOC method.
Relaxed Shape Index (RSC). SOC method extension
The software integrates fusion methods that combine into a single cluster the clusters created from each of the independently performed time series of microarray data. The detection of the most important clusters within the possible clusters is performed using different measures on the genes, among which the Information Correlation Coefficient (ICC), Pearson Correlation Coefficient (PCC) and Shape Increase measures.

Wrapper system modeling software

Feature selection, modelling and classification of data

SOFTWARE FOR FEATURE SELECTION ON MODEL BY MEANS OF A WRAPPER SYSTEM

Sistema wrapper de diseño

Design wrapper scheme

Advanced software for the selection of own design features on models by means of a wrapper system.

The tool allows for dimensionality reduction of large data sets, based on modelling, classification or pattern recognition. A genetic algorithm is used to select the features that produce the best models or classification systems; namely, where the highest representation percentages -best predictive models- are obtained.

The software also allows modelling and classification without taking into account the features, which is suitable for a small number of features. The application allows the use of 5 different ways to validate the models.

Applications

Data analysis for Industry 4.0, Agrotech, FinTech, Biotech, BigData, Research. Suitable for researchers who, in large datasets, seek to determine important features and relationships between them, and output models with respect to the inputs.

Software for feature selection in the design of experiments and generation of new products

Selección de características del software

Soft Computing software of our own design for the selection of characteristics and the reduction of the dimensionality of the data for the conception of new products, under the development of a design of experiments.

The software allows the selection of characteristics by means of a genetic algorithm that obtains a limited set, by means of the following measures:

Population Correlation Coefficient (PCC).
Information Correlation Coefficient (ICC).
Mutual Information (MI).

The input variables selected are those with the highest ICC, PCC and MI values with respect to the output.

In order to reduce the redundancy of the information of the selected features, the mRMR algorithm is used, which allows the selection of the features with the highest relevance with respect to the output and the minimum redundancy between the input variables.

Applications

Experimental research, development of new products in areas such as pharmaceuticals, chemistry, new materials…