Projects

Fernando Herranz has collaborated in several national projects as well as industrial projects in the field of robotics and intelligent transportation systems.

Industrial Projects

  • VISUALISE

    VISUALISE

    Development of semi-automatic detection tool of Signs and Pannels

    Visual Inspection of Signs and Pannels (VISUALISE) project. In collaboration with the multinational companies 3M corporation and EUROCONSULT S.A.

    VISUALISE project was focused on the development of an automatic inspection system of traffic signs. The purpose of this system was to analyze the retrorreflection curves of traffic signs in order to decide if they fulfill the Spanish regulations related to traffic signaling. Traditional traffic sings inspection techniques were very expensive and time consuming, however VISUALISE approach was based on an automatic inspection system, mounted onboard a vehicle, which performed inspection tasks at conventional driving speeds. VISUALISE allowed for an improvement in the awareness of the road signaling state, supporting planning and decision making on the administration’s and infrastructure operators’ side.

    Computer vision techniques were applied to detect and track traffic signs and calculate their
    retrorreflection curves automatically. The detection algorithm consisted of a shape recognition technique based on the Hough transform, while the tracking was carried out by using Kalman filtering.

    – VISUALISE project has been developed in C++, Matlab
    – The system was based on the integration of independent modules such as Detection module, Trackin moudule, Retrorreflection module.
    – Real tests have been performed with satisfactory results. VISUALISE was tested in Spanish roads (about 5000 km.).

    Video

National Projects

  • ROBO SHOP

    ROBO SHOP

    Robot assistant for shopping

    ROBO SHOP project continues PROPINA project and aims at using PROPINA platform to assist people. The project focuses on the development of a robotic middleware that provides services to high-level applications. Those services are categorized into low level and high level services. Low level services implement functionalities such as obstacle avoidance, environment reconstruction, etc. while high level services implements localization, path planning, object recognition, etc. The design of the project is based on a modular architecture where services interconnect with each other by means of messages.

  • ABSYNTHE

    ABSYNTHE

    Abstraction, Synthesis, and Integration of Information for Human-Robot Teams

    ABSYNTHE is an interdisciplinary project coordinated by the European Centre for Soft Computing (ECSC), in cooperation with the Robotics and e-Safety (RobeSafe) research group of the University of Alcalá (UAH).

    The main aim of this project is the development of concepts, tools, and approaches to facilitate the collaboration of humans and robots team by taking advantage of the joint intelligent processing and exploitation of knowledge. For this purpose ABSYNTHE is focused in the application of knowledge extraction and data mining techniques to abstract, from complex data objects, information elements essential to the safe and effective operation of the human-robot team.

    ABSYNTHE project has been developed in C++, Matlab, Javascript
    The system is a decentralized architecture based on the integration of different systems such as data-mining algorithms, robots management, etc.
    Real tests have been performed with satisfactory results. The tests consisted in a team of robots in charge of guiding a person around a building. The experiment considered a building with multiple floors and real conditions.

    Video

  • Robocity2030

    Robocity2030

    Innovative integration of applications of Service Robots

    The objective of Robocity2030-II is to develop an innovative integration of applications of Service Robots, in an effort to increase the quality of living of citizens in metropolitan areas. This means that the human is now the center of things and the Service Robots are developed from, for and to the benefit of humans.

    The applications of Service Robots that this project proposes, focuses on metropolitan areas of the Community of Madrid, and can be divided into four major groups:
    – Field robotics: construction (building, maintenance, civil building), agriculture (harvest, dust), inspection of infrastructures (aerial and submarine robots) and intelligent transport system (autonomous vehicles in metropolitan environments)
    – Applications in unfriendly environments such as mining (rescue), power plants (through ”remote handling”), humanitarian landmine detection, intervention in natural disasters (avalanches, earthquakes), surveillance and security systems
    – Applications in domestic environments such as personal and social robots, education, entertainment and robots for cultural guide purposes
    – Applications in health care system such as disable people or elders’ aid, surgical robots (laparoscopy and traumatic injuries) and robots for rehabilitation purposes (prosthesis and orthosis).

  • PROPINA

    PROPINA

    Reseach Robotic Hardware Platform

    Propina project focuses on building a robotic research platform and its integration with ROS to develop high-level applications. The platform is designed to work in indoors with a differential traction configuration. It is equipped with odometry and distance sensors such as ultrasonic and infrared. An Arduino-based architecture is used to run ROS modules that control actuators (motors) and to acquire information from sensors. Hence, the perception is completely transparent to the user. The modular design has been chosen to increase the functionality and autonomy. In addition, it has been designed a 3D model in Gazebo simulator that can be used as prior before designing the actual application.

    – Propina project has been developed in C++.
    – The integration of the architecture has been done by a modular configuration of Arduino components that take care of specific tasks.
    – Propina platform has been tested in real indoor scenarios

  • VISETRAF

    VISETRAF

    Visual detection system of Signs with Radio Frequency

    VISETRAF project aimed at developing a complete traffic sign recognition system based on vision sensor onboard a moving vehicle. A restricted Hough transform was used as detection method from the information extracted in contour images, while the recognition system was based on Support Vector Machines (SVM). The project also implemented a novel solution to the problem of discarding detected signs that does not pertain to the host road. For that purpose infrastructure-to-vehicle (I2V) communication and a stereo vision sensor were used. Furthermore, the outputs provided by the vision sensor and the data supplied by the CAN Bus and a localization system based on RF were combined to obtain the global position of the detected traffic signs, which was used to identify a traffic sign in the I2V communication.

    – VISETRAF project has been developed in C++.
    – Tests were performed under real driving conditions, both day and night. An average detection rate over 95% and an average recognition rate around 93% were obtained with an average runtime of 35 ms that allowed the system to run in real-time.

  • SISLOPEWI

    SISLOPEWI

    WiFi People localization system

    The main objective of SISLOPEWI project was the development of localization system for people using WiFi signal strength. The system was based on a priori radio maps that were generated automatically by a robot system during a training stage. Hence, the calibration effort was reduced. The system was applied to localize interesting points for people such as meeting rooms, offices, etc.

    – SISLOPEWI project has been developed in C++, Matlab
    – The localization system was deployed in a small portable device with a WiFi interface such as a smartphone
    – It was not necessary to add any additional hardware to the building. The project used the available wireless network of the building
    – The system used robots and ROS to automatize the training stage
    – Real test were performed obtaining a localization rate of 90% and demonstrating that the system was able to guide people around the building