INDUSTRIALES RESEARCH MEETING

8/17/2019 INDUSTRIALES RESEARCH MEETING

1/224

! "# $%&

'

( !)* +,* "# $%&

-

"# $%& .# / &

' 0

1& 2 & 13

- 4 !

"# $%&

0 5 "# #$% $

"# $%& 6 732 /

4 & ' "#

$%& & 1 ! !2 "

5 &(#8*)

92% "# $%& !

1


2/2242


3/224

AbstractAcoustic signals are present almost in every process, every machine, every action. In many cases, we can only consider noise as anunwanted result, an unwanted sound that we would like to remove from a product or a process. But, we shouldn’t forget that, as result of a

process, noise and vibration are acoustic signals that contain information about it. Is it possible to extract that information from the

acoustic signals? Signal processing and pattern recognition can be applied to classify sounds, inferring the information they provide. This is

a research line that the I2A2 research group initiated in 2008, and that has achieved quite a good number publications in scientific journals,

an three patents. Pattern recognition has been applied to the detection and classification of environmental noise sources, the detection of

threats in a security system, the monitoring of water boiling stages, and the road estate detection in a vehicle.

APPLICATIONS OF PATTERN RECOGNITION

IN ACOUSTICS

C. Asensio

Instrumentation and Applied Acoustics Research Group (I2A2)

Universidad Politécnica de Madrid - [email protected]

Applications of sounds

classification

PAPERS IN JOURNALS

§ Tarabini, M.; Moschini, G.; Asensio, C.; Unattended acoustic events classification at the vicinity of airports. Applied Acoustics (2014)§ Alonso, J.; López, J.M.; Pavón, I.; Recuero, M.; Asensio, C.; Arcas, G.; Bravo, A. On-board wet road surface identification using tyre/road

noise and Support Vector Machines. Applied Acoustics (2014)

§ Asensio, C.; Moschioni, G.; Ruiz, M.; Tarabini, M.; Recuero, M.Implementation of a thrust reverse noise detection system for airports.

Transportation research Part D (2013).

§ Tabacchi, M.; Asensio, C.; Pavón, I.; Recuero, M.; Mir, J.; Artal, M.C. A statistical pattern recognition approach for the classification of

cooking stages. The boiling water case. Applied Acoustics (2013).

§ Asensio, C; Ruiz, M; Recuero, M. Real-time aircraft noise likeness detector. Applied Acoustics (2010)

WATER BOILING STAGES – The aim in this project was to use sound andvibration signals for monitoring the stages of boiling. An almost perfect

classification was achieved on the pilot case, using a Mixture of Gaussians

classifier. The following step towards an acoustic cooking assistant is under

research.

A SOUND BASED SECURITY SYSTEM - The system is able to detect in real

time individual sound events, and classify them of the basis of a previously

defined threats (glass break, steps, door slam, shot …). Impulsiveness of

sounds (instead of their intensity) is used to capture the events, then, a

GMM classifier is used for the classification stage.

Anomalous

sounds detection

ROAD STATE – A pilot case was implemented to detect wet asphalt using

the sound that tyre-road interaction generates. The final goal in this projectis to extract further information to infer information from road

maintenance status, grip of the vehicle, type of asphalts… This approach

can be also applied to railway.

signal

TSL

ONE-CLASS CLASSIFIERS are trained when only one of the classes can

be described properly. For instance, we have used this approach todetect aircraft sounds in an environmental noise monitoring system.

One of the classes is aircraft sound, but the other class is composed

by all the rest of sounds, and therefore we will always find sounds for

which the system has not been previously trained. In this case, the

detection model fits the “target class”, and the system classifies as

outliers those sounds that are not close enough to the target.

This is an approach that can be also applied for the identification of

alerts, or ouliers in a system (failures in a machine or anomalous

conditions).

3


4/2244


5/224

AbstractMost of airborne sound insulation test are performed according to certification standards, and only deal with

sound pressure as a variable to estimate insulation. Besides, standards are aimed to provide direct path

insulation, so one gets no idea regarding flanking paths. Without data from every path involved in sound

transmission, a proper insulation designing stage can became mistaken and unfinished.

ROOM FLANKING SOUND TRANSMISSION

ASSESSMENT USING A VIBRATION PROBEa San Millán-Castillo, Roberto; b Pavón, Ignacio.

a, b Instrumentation and Applied Acoustics Research Group (I2A2), Technical University of Madrid (UPM)

[email protected] E.T.S.I. Telecomunicación – Universidad Rey Juan Carlos, Camino del Molino, s/n - 28943 Fuenlabrada

(Madrid, España). [email protected]

Goals• Room surfaces vibration signal acquisiton method

set-up and detailed description.

• Simplified airborne sound insulation model

development.

First tasks going on…Focus is on vibration measurement on room surfaces to

get information about flanking sound transmission.

Working with accelerometers is usual, its mounting

techniques are less obvious. The suggested method

consists of performing vibration velocity level

measurements, with a FFT analyzer, at the same time as

a regular airborne sound insulation test is carried out.

Then a surface weighted contribution model would let us

know about the key room parts on sound proofing.

Different room materials and probes are analyzed. It isprioritized a solution that lets one a fast, comfortable,

and barely intrusive data acquisition.

Figure 2. Concept sound transmission sketch.

…First conclusions coming out• OPERATIONAL USE: Ease of use and time saving

with probes; Better Intrusion and acces to difficult

points.

• SIGNAL QUALITY: Velocity underestimation and

SNR decrease; Reasonable standard deviations on

tests.

Figure 3. Measurements with different vibration probes

!"

# $%

tasks o ng on…s on ratii on measurement on oom surfaces t

onformati about flanking sound transmissio

ng wit a celerometers is usual, ts ountin

ques re ess obvious. he suggested etho

ts o erforming vibration elocity e

nts,reme ith a FF analyzer, at the same time a

lar rboai ne sound insulation test s carried ou

cesurf eighted contribution model would let u

re 2.Fig Concept soun transmission s etc .Figure 1. Tests layout: Airborne Emission, vibration recepetion

Figure 4. Different probes averaged standard deviation ( (f)) .

5


6/2246


7/224

AbstractNoise is one of the most complex problems to manage in our society, as the environmental policies must be put in common with othermportant factors, such as mobility, employment and economic growth. The irruption of information and communication technologies

ICTS) an social networks has enabled novel opportunities to decision and policy making processes.

CTS give voice to wider audiences, allowing to extend the debates from local to international scales. In this research we are investigating

the basis for a future implementation of a noise management model based in Policy Making 3.0. This is a participatory and evidence based

model that tries to involve authorities and the rest of stakeholders to define policies, on the basis of scientific evidence and collective

aspirations of citizens.

EXPLOITATION OF ICTS FOR COMMUNITY

NOISE MANAGEMENT

C. Asensio; L. Gascó; G. de Arcas



Transport acceptance, noise and ICTS

Unlike other pollutants, most of the noise effects have

an important subjective component, beyond the

physiological effects that the physical phenomenon

produces. The community response against noise affects

the degree of acceptance of transport infrastructures

and city activities.

It is not so evident the effect of non-acoustic factors that

increase the societal rejection, like: lack of sensitivity and

empathy from the authorities and noise managers, the

lack of trust in them, the lack of information andtransparency, the perception of being excluded from the

decision making… Much of these non-acoustic factors

can be modified through a correct use of information

and communication technologies (ICTS), being this issue

addressed in our research.

I2A2 proposes a parallel strategy, complementary to the

traditional ones (based on the decrease of decibels),

which aims to minimize the impact that noise has on the

rejection of transport and other city activities. Through

the exploitation of ICTS in the noise monitoring systems,

we will face the following issues:

§ Provide transparent information, easy to understand,

close to people and reliable, through the use of thenoise indicators and reporting mechanisms better

aligned to citizens’ expectations.

§ Implement engagement tools that allow citizens to

express their feelings about noise, and to establish a

bi-directional communication with the noise

managers.

This research will directly improve the efficacy of present

communication tools, and will start a roadmap to the

implantation of a participatory noise management

model that, based in Policy Making 3.0, will urge the

interaction between community and other stakeholders,

having in mind either their interests and aspirations,

besides the scientific evidence.The research in this novel approach is focused in two

different paths: transport noise, and noise in smart

cities.

§

§

i

i

Policy Making 3.0 applied to community noise

management Noise management in smart

societies

7


8/224

AbstractThe perceived quality of any electrical appliance, vehicle, or either interior or exterior spaces is influenced by the quality of its sound, andthis includes the reduction of noise, but also a fine tuning of its sound features. Sound quality is growing field of study that transversally

nfluences the design of products and spaces, combining objective measures with subjective perception of people. Closed to mechanical

and acoustics engineering, sound quality and soundscapes bring people views into the scene, trying to translate their subjective response

nto objective parameters that can be implemented in a product or a space. I2A2 has been involved in several research projects closely

related to the optimization of sound emissions of products, such electrical appliances or vehicle components, and spaces, such as vehicle

cabins, building components or rooms in an industry.

IMPROVEMENT OF SOUND QUALITY IN

SPACES AND PRODUCTS

C. Asensio, I. Pavón



Sound quality in electrical appliancesTHE REFRIGERATOR CASE - The aim in this Project was to reduce sound emissions in a

refrigerator, in order to improve the product acceptance and its commercial outcomes.

The first objective was to determine if it was the compressor or the fan the responsible for

the noise levels. We made sound intensity, sound pressure and vibration measurements, to

finally come to the conclusion that it was necessary to prioritize the mitigation measures in

the fan. Cold air flow directed from the electric fan to upper compartment of the

refrigerator produces turbulences that generate noise. This is a tonal sound with a main

frequency closely related to the RPM and the number of blades in the fan. A FEM analysis

was performed, in order to analyse the frequency response of mass of air that is enclosed

by the duct and the upper compartment. We observed that the tonal noise was tuned with

the resonances of the compartment, and this amplified the sound produced. Therefore it

was suggested to adjust smoothly the RPM to break this coupling.

THE WASHING MACHINE CASE – A similar approach was

followed, but vibrations were the main source of noise in this

case. The aim of the project was to fin a model to find the

propagation paths and the influence of damping, when inserted

in different parts of the structure.

Indoor an outdoors spaces

§ Asensio,C.; Recuero, M.; Pavón, I.; Citizens' perception of the efficacy of airport noise insulation programmes in Spain. Applied Acoustics(2014)

§ Tabacchi, M.; Pavón, I.; Ausejo, M.; Asensio, C. Assessment of noise exposure during commuting in the Madrid Subway. Journal of

Occupational & Environmental Hygiene (2011)

§ Ausejo, M.; Tabacchi, M.; Recuero, M.; Asensio, C.; Pagán, R.; Pavón, I. Design of a noise action plan based on a road traffic noise map.

Acta Acustica united with Acustica (2011)

§ Asensio, C; Pagán, R; Trujillo, J; Recuero, M. Intelligibility of Speech in the Spanish Congress of Deputies. Journal of Acoustical

Engineering society (2010)

The acoustics of a theatre, a cinema, or a conference room have been widely studied in the past,

as room acoustics are important to improve the speech intelligibility, or the quality of a musical

program. But, its basis can be also applied to the design of either interior or exterior spaces.

Acoustic in the cabin of an aircraft can be crucial, in order to improve communications between

the pilots and the air traffic control. And it can be also critical to improve the performance of an

operator in machinery, or to reduce the accidents’ risk. But the acoustic performance can be also

important for the perceived overall quality of a product, like in the cabin of a car or a train.

Our most recent research in sound quality and soundscapes is closely related to model subjective

response from users and citizens, in order to customize the acoustic environment.

PAPERS IN JOURNALS

8


9/224

!"#!$%! & &' !

() '& )*+ , - . / 0 123 "4!5

6 7& 86 6 9 '' ) 2 )): ) ' & '- ' '' &

1"4!53

;.2 86 6 & 86 )


10/224

AbstractThe irruption of information and communication technologies (ICTS) and social networks has enabled novel opportunities to manage noisen Smart Cities. For instance, noise assessment using networks of low-cost sensors and dynamic noise maps provides a certain opportunity

to apply the polluter pays principle to road traffic management in a city, using noise quotas that can be dynamically adjusted to individual

vehicles, depending on factors such as the acoustic pollution conditions in an area, the people living in that area, the state of maintenance

of a vehicle, the driver’s behavior… Additionally, it is necessary to establish the minimum requirements that noise sensors must comply so

that noise assessment can be accurate and precise applying a cost-benefit criteria.

NOISE MITIGATION STRATEGIES IN SMART

CITIES

C. Asensio; I. Pavón; G. de Arcas; J. Alonso; J.M. López; P. MéndezInstrumentation and Applied Acoustics Research Group (I2A2)


Low acoustic emissions zones

§ To assess noise environment, having objective indicators describing the

acoustic situation

§ To inform the residents about their noise exposure, using measures that

improve trust among the general public

§ To rate the situation in comparison to regulations

§ To compare different locations and areas, in order to establish priorities

for action plans

§ To determine the reduction required in each area

§ To determine which are the stages of noise at each location, along a

working day, during weekends…

§ To raise noise awareness at all society levels (bar, pub and restaurant

owners and users, residents, and even authorities).

§ To give support to any mitigation action, which could have effect for

owners, users or residents.

§ To evaluate the efficacy of any mitigation measure

Requirements for a low-cost noise monitoring network

SINGLE VEHICLE’SNOISE EMISSIONS

NOISE PROPAGATIONMODEL

ON-BOARDOBSERVATIONS

EMISSIONS MODEL

NOISE MAPPING

VEHICLE BASEDPEOPLE’S EXPOSURE

GPSCoordinatesdBKm/hm/s2

Driver’s behabiourehicle type / maintenanceheels type / maintenance

SINGLE VEHICLE

DYNAMICNOISE QUOTAS

LOW NOISE EMISSIONS ZONE APP Awareness

Drivers

informationRoutes definitionRates applicationPayment method

POLLUTER PAYSPRINCIPLE

ROAD TRAFFICNOISE COSTS

LOW NOISE

EMISSION ZONES

ROAD TRAFFICDYNAMIC NOISE

MAPS

NOISE PROPAGATIONMODEL

CITY SENSORS AwarenessDrivers informationTraffic flow Traffic speedTraffic conditionsNoise pollution

Integration intoSmart City

Communication tothe public

Road trafficmanagement

Road traffic restrictions and vehicles’ speed reduction are typical measures that can

be implemented to manage noise in a city. But these are quite rough measures that

could be customized attending to more complex criteria: number of residents

affected, overall noise levels… But, even these criteria can be specifically fit forpurpose on the vehicle emissions basis, in order to implement a polluter pays policy,

distributing noise costs among the separate vehicles, through the use of noise quotas

that can be dynamically adjusted depending on the acoustic conditions of the area or

the vehicle.

Monitoring acoustic pollution in a city is necessary because of the following reasons:

There is an international standard describing the requirements of sound level meters, and traditional noise monitoring units must conform to these

requirements. But these type certified monitors are quite expensive, and their implementation cannot be as extended as Smart Cities applications would

require. This Project aims to identify the requirements that can be relaxed on view of the specific characteristics of urban noise, so that cost-benefit can be

optimized in the low-cost Acoustic sensors.

10


11/224

AbstractThe attention processes in people is affected by background noise produced by many different sources. Beta (13-30 Hz) and Theta (4-7 Hz) waves aredirectly related to attention and memory processes. Volunteers were asked to perform an attention test with and without background noise and their

cerebral activity was recorded through electroencephalography (EEG). Results shows significant decreases in both beta and theta frequency bands (beta

13-30 Hz and theta 4-7 Hz) under background noise exposure. The attentional improvement is related to increases of the beta and theta waves, and we

have observed that those decreases are directly related to a lack of attention caused by the exposure to background noise.

Objective

NOISY ENVIRONMENTS: THE INFLUENCE

ON BASIC COGNITIVE PROCESSES

E. Tristán-Hernández, I. Pavón, J. M. López.



Backgroundnoise

Cognitive

processes

Methods (II)

1.1 SURVEY: perception about noise

1. Acoustic characterization

1.2 Noise annoyance1. Acoustic characterization

ICBEN

Numeric scale range

Loudness

Roughness

Sharpness

Fluctuation Strenght

Obj j. Subjective

Noise

annoyance

Numeric scale range

a) 0 – 3.60 b) 3.61 – 23.55 c) 23.56 – 58.62,

d) 58.63 – 89.77, e) 89.78 – 100

Results

Methods (I)

1.1 Survey

1.2 Noise exposure

1.3 Noise annoyance

1. Acoustic characterization

2. Psychological evaluation

2.1 Attention tests

2.2 Memory tests +Background

noise

Electroencephalogram

+

Noise exposure

1.3 Noise Exposure1. Acoustic characterization

UNE-EN ISO9612:2000

Measurementstrategy design

Dataloggers

Measurementcampaign

Analysis of the results

LAeq, TLEX

8hANSI S12.60-2010

40 dBA interior

60 dBA exterior

2. Psychological

characterization

Attention tests

- Toulouse Pieròn

- Own proposal

Memory test

Own proposal

3.1 Electroencephalogram + Noise exposure

3. Electrophysiological

assesment2.1 Attention tests

2.2 Memory tests

33

66.97V

54.22 V

29.19V

21.67 V

p=0.015

p=0.010

# Questions* Participants

Mode Educational level

OnlineFace-

to-face Bachelor Master PhD

8 454 50% 50% 85% 10% 5%

Conclusions

Noise exposure Noise Annoyance Psychological evaluation

E l e c t r o p h y s i o l o g i c a l

a s s e s m

e n t

- Survey: Sound environments studied are aggressive.

- Levels of noise: Excessives and not appropriate for concentration

activities and the environmentes are considerated very annoying.

- Psychometric evaluation: Unreliable data.

- EEG: ! y " decrease with influence of backround noise. We can relate it

with attention and memory reductions.

Papers at Journals & Conferences Tristán E., Pavón G., López, J.M (2015). “Characterization of sound environments of university students” . Int. J. Occupational Safety and

Ergonomics (JOSE).

Tristán E., Pavón G., López, J.M. (2016) “Evaluation of psychoacoustic annoyance and perception of noise annoyance inside university

facilities” . International Journal of Acoustics and Vibration (IJAV).

3. Electrophysiological

assesment

The main objective of this research

s to prove the negative influence

of background noise on the brain

waves related to basic cognitive

processes as attention and

memory.

11


12/224

12


13/224

9+ ' ## )* #$ +:;

.**

9+ - +,&*'

(&; 3


14/22414


15/224

ABSTRACTDrones have become very popular in our society because of the low cost, high performance and the multitude of activities that they can be used for. From

single copter to octacopter relies on high energy density Lithium Polymer (LiPo) batteries to provide enough power to lift high weigh but with limited flight

time. To increase the batteries lifespan, increase flight time, charge onboard batteries when landed, increase the available payload, reduce batteries

capacity and lighten the drone itself, an hybridization system must be added. It means that apart from an high grade LiPo battery, an “Auxiliary Power

Unit” must be applied as an upgrade.

ARDUINO-BASED LIGHTWEIGHT APU FOR DRONE

HYBRIDIZATION

Juan Bautista Ruiz-Yherla García

[email protected] / [email protected]

POWERPLANTRC glow engine

mall, light and powerful. These engines are

ased on a two stroke cycle with HCCI

ombustion process. The combustion starts

with a red hot platinum coated glowplug as

atalyzer. The fuel consist of a mixture

etween methanol, nitromethane (for

ncrease power output), synthetic oil and

astor oil.

he output power vary from 30W to more

han 1000W, up to 30000 rpm and a weight

ower than 600g.he shape of the performance curve is

milar to:

ELECTRIC GENERATORRC DC Brushless motor

Without energy loss by brushes friction, ball

bearings, three phase coil, smooth running,

high power density, extremely light, can work

as high frequency AC generators and low cost.

High grade brushless motors have an efficency

up to 95%. A great variety of models and

especifications.

A performance curve example:

CURRENT RECTIFICATIONSix diode rectifier

The easiest way to rectify the three phase AC

current consist of using a six high current

diode bridge rectifier. The system can feed a

boost converter or a voltage forced output

system such as a LiPo battery.

REFERENCES[1]Intelligent Energy fuel cells significantly boost drone flight time. Fuel Cells Bulletin Volume 2016, Issue 1, January 2016, Pages 4

[2]Snorri Gudmundsson, Chapter 7 – Selecting the Power Plant . General Aviation Aircraft Design, Applied Methods and Procedures 2014, Pages 181–234

LOAD REGULATION

Arduino based power output control

To manage the engine power output, an Arduino

board with a microcontroller is used to check

battery voltage status, current feeding and

control the throttle %, as well as engine rpm. The

throttle is controlled by a RC servo.

THE IDEAThe APU is based on a small lightweight hobby “two stroke nitro engine”, a paired hobby brushless motor with voltage boost-regulator and a small

“Arduino Nano” board with an Atmega328 microcontroller to manage the engine throttle (to adequate the power delivery on demand), check battery and

APU voltage, current demand, engine speed and also to add telemetry. The parallel hybridation system is suggested.

Glow

Engine

Brushless

Generator

Boost-

rectifier

LiPo

BatteryLoad control

and system

check (Arduino)

Drone speed

controllers

T h r o t t l e %

E n g i n e

s p e e d

Voltage and

current feedback

RESULTS AND CONCLUSIONSThe status of the research it’s in progress.The early tests go as expected .

Need to try the APU in a drone to check vibration damping .

15


16/224

AbstractTo simplify the usage of the Unmanned Aerial Systems (UAS), extending their use to a great number of applications, their fully autonomous operation is

needed. There are many open-source architecture frameworks for UAS that claim the autonomous operation of UAS, but they still have two main open

ssues: (1) limited level of autonomy and (2) versatility, being most of them designed specifically for some applications or aerial platforms.

As a response to these needs and issues, we developed Aerostack (http://www.aerostack.or g/ ), a system architecture and open-source multipurpose

software framework for autonomous multi-UAS operation. To provide higher degrees of autonomy, Aerostack's system architecture integrates state of the

art concepts of intelligent, cognitive and social robotics, based on five layers: reactive, executive, deliberative, reflective, and social. To be a high versatile

practical solution, Aerostack's open-source software framework includes the main components to execute the architecture for fully autonomous missions

of swarms of UAS; a collection of ready-to-use and flight proven modular components that can be reused by the users and developers; and compatibility

with five well known aerial platforms, as well as a high number of sensors. Aerostack has been validated during three years by its successful use on many

research projects, international competitions and exhibitions.

Autonomous Visually-Guided Navigation of

Unmanned Aerial Systems

aJose Luis Sanchez-Lopez, aPascual Campoy (Advisor)

aComputer Vision Group, Centre for Automation and Robotics, CSIC-UPM

([email protected])

References:

[1] J.L. Sanchez-Lopez, J. Pestana, P. de la Puente, R. Suarez-Fernandez, and P. Campoy. A system for the design and development of vision-based multi-

robot quadrotor swarms. In Unmanned Aircraft Systems (ICUAS), 2014 International Conference on, pages 640–648, May 2014.

[2] J.L. Sanchez-Lopez, J. Pestana, P. de la Puente, and P. Campoy. A reliable open-source system architecture for the fast designing and prototyping of

autonomous multi-uav systems: Simulation and experimentation. Journal of Intelligent & Robotic Systems, pages 1–19, 2015.

[3] J. Pestana, J. L. Sanchez-Lopez, P. de la Puente, A. Carrio, and P. Campoy. A vision-based quadrotor multi-robot solution for the indoor autonomy

challenge of the 2013 international micro air vehicle competition. Journal of Intelligent & Robotic Systems, pages 1–20, 2015.

[4] J. Pestana, J.L. Sanchez-Lopez, S. Saripalli, and P. Campoy. Computer vision based general object following for GPS-denied multirotor unmanned

vehicles. In American Control Conference (ACC), 2014, pages 1886–1891, June 2014.

[5] J.L. Sanchez-Lopez, J. Pestana, J.-F. Collumeau, R. Suarez-Fernandez, P. Campoy, and M. Molina. A vision based aerial robot solution for the mission 7

of the international aerial robotics competition. In Unmanned Aircraft Systems (ICUAS), 2015 International Conference on, pages 1391–1400, June 2015.

Aerostack Architecture

Person following by a UAS [4].

Research on multi-UAS [1] and [2]

IMAV 2013 Indoors Autonomy Competition [3].

CVG won the first Prize.

IARC 2014 Mission 7 [5]. CVG won

the two special prizes.

www.vision4uav.com

www.aerostack.org

16


17/224

AbstractUnmanned Aerial Vehicles (UAVs) have recently become a versatile platform for many civilian applications including inspection, surveillance and mapping.

Current regulations require these vehicles to be capable of maintaining self separation for conflict avoidance, providing an Equivalent Level of Safety

ELOS) to that of human pilots. Sense & Avoid (SAA) systems provide solutions to this problem with the final objective of safely integrating UAVs in our

airspace, which is currently one of the main technological challenges for using these platforms in civil applications. Electro-optical systems, in particular

RGB and thermal cameras, are suitable for onboard obstacle detection, given their low cost, reduced size and light weight. In our research, we propose

solutions for obstacle detection and tracking onboard UAVs using Computer Vision.

COMPUTER VISION FOR SENSE & AVOID

ONBOARD UNMANNED AERIAL VEHICLES

Adrian Carrio and Pascual Campoy

Computer Vision Group, Center for Automation and Robotics (UPM-CSIC),

Calle José Gutiérrez Abascal 2, Madrid, Spain ([email protected])

Synthetic image generationObtaining real ight imagery of potential collision scenarios is hard and dangerous, which

complicates the development of Vision-based detection and tracking algorithms. We have

developed software for synthetic imagery generation, allowing to blend user de!ned ight

imagery of a simulated aircraft with real ight scenario images to produce realistic images with

ground truth annotations. These are extremely useful for the development and benchmarking of Vision-based detection and tracking algorithms at a much lower cost and risk.

Related publications

• Carrio, A.; Changhong Fu; Pestana, J.; Campoy, P., " A ground-truth video dataset for the

development and evaluation of vision-based Sense-and-Avoid systems," in International

Conference on Unmanned Aircraft Systems (ICUAS), 2014, pp.441-446, 27-30 May 2014

• Carrio, A.; Changhong Fu; Collumeau, J.F.; Campoy, P., “SIGS: Synthetic Imagery Generating

Software for the Development and Evaluation of Vision-based Sense-And-Avoid Systems," in

Journal of Intelligent & Robotic Systems, 2015, pp.1-16, Springer Netherlands

Sky segmentation and horizon line detectionVision-based sky segmentation can be extremely useful to perform important

tasks onboard Unmanned Aerial Vehicles (UAVs), such as pose estimation and

collision avoidance. We investigate how to perform sky segmentation in RGB

images using a supervised Machine Learning approach by !rst splitting the image

into !xed-size patches, extracting and classifying color descriptors for each patch

and performing a !nal post-processing stage to improve the quality of the

segmentation. A method for automatic horizon line detection is also proposed.

The performance of our approach was evaluated on ight images captured

onboard UAVs, achieving performance accuracies above 93% at real-time frame

rates.


• Carrio, A.; Sampedro, C.; Changhong Fu; Collumeau, J.F.; Campoy, P., " A Real-

time Supervised Learning Approach for Sky Segmentation Onboard Unmanned

Aerial Vehicles," in International Conference on Unmanned Aircraft Systems

(ICUAS), 2016, under revision

Aircraft detection and trackingWe present a lightweight obstacle detection system for small Unmanned Aerial Vehicles (UAVs) combining a Thermal Infrared (TIR) camera and an

Automatic Dependent Surveillance-Broadcast (ADS-B) receiver. Additionally, synchronized visible spectrum (VIS) images have been captured in order to

ompare the capabilities of TIR and VIS imaging technologies for aircraft detection in Sense-and-Avoid (SAA) under different lighting conditions. TIR

maging is proved to be an effective technology for vision-based SAA systems, allowing for operations under extreme illumination conditions such as direct

un exposure or during the night, outperfoming current vision-based SAA systems.


Carrio, A.; Saripalli, S.; Campoy, P., "Obstacle Detection System for Small UAVs using ADS-B and Thermal Imaging," in 2016 IEEE International Conference

on Intelligent Robots and Systems (IROS), under revision

Fig 1. Real (above) and simulated (below)

aircraft intruders in real footage captured

onboard a UAV

Fig 2. Sky segmentation (left) and horizon line detection (right)

Fig 3. Aircraft detection in a RGB image (left), thermal image (center) and UAV equipped with our obstacle detection system (right)17


18/224

AbstractThe interest in missions with multiple Unmanned Aerial Vehicles (UAVs) has increased significantly in last years. These missions take advantage of the use

of fleets instead of single UAVs to ensure the success, reduce the duration or increase the goals of the mission. In addition, they allow performing tasks

that require multiple agents and certain coordination (e.g. surveillance of large areas or transport of heavy loads). Nevertheless, these missions suppose a

challenge in terms of control and monitoring. In fact, the workload of the operators rises with the utilization of multiple UAVs and payloads, since they

have to analyze more information, make more decisions and generate more commands during the mission.

This work addresses the operator workload problem in multi-UAV missions by reducing and selecting the information. Two approaches are considered: a

first one that selects the information according to the mission state, and a second one that selects it according to the operator preferences. The result is an

nterface that is able to control the amount of information and show what is relevant for mission and operator at the time.

Multi-UAV Coordination and Control Interface

Juan Jesús Roldán, Jaime del Cerro and Antonio Barrientos.

Centre for Automation and Robotics (CAR, UPM-CSIC), [email protected]

Mission modelingA mission model is required for monitoring the mission state and its possible evolution. This

model allows selecting the information that is relevant according to the mission state.

n this work, the models are generated automatically from the experience of previous missions.

The discovery algorithms of process mining are used to obtain Petri nets that represent themissions.

Currently… Future…

Figure 1: Current and future schemes for control and monitoring of multi-UAV missions.

Multi-UAV mission

Operator behaviorOn the other hand, this work is focused on the

operator. Therefore, the selection of the

information must be performed not only

according to the mission but also according tothe operator.

In this way, the information should be adapted

to the physical and psychological state of

operator. For instance, when the operator is

stressed, tired or bored, the system should

reduce the amount of data, in order to prevent

saturation and errors.

In addition, the information should be adapted

to the preferences of operator. For example, if

the operator usually asks for a specific

information in a certain state of the mission, the

interface should anticipate the operator request

and provide this information.

Multi-UAV missionA multi-UAV mission consists of a set of resources (both UAVs and payloads) that are allocated to some tasks (e.g. surveillance, capture and release,

maintenance…) in order to achieve a series of objectives (fire surveillance, fire extinguishing…).

These missions generate a huge amount of data about the UAVs, payloads and scenario. Each UAV may generate a telemetry of around 10-100 messages

and 100-1,000 variables depending on its size. Meanwhile, each payload may provide its own data: e.g. the images of cameras or the measures of sensors.

Finally, the scenario may also generate data related to the tasks, their objectives, their status, etc.

InterfaceThe expected outcome from this work is an Intelligent Adaptive Interface (IAI), which is able to change dynamically according to the context of the

mission. This interface will have two parts: a static one with the information that must be shown permanently (e.g. mission map, current targets, available

UAVs…) and a dynamic one with the information that is adapted to the context (e.g. system alarms, critical UAV data, important payload measurements…).

Figure 2: Monitoring of fire surveillance mission by using Petri net.

18


19/224

AbstractNowadays there has been an important increase of electric equipment in modern aircraft which leads to a rising demand of electrical power. This

highlights the necessity of efficient, reliable and high energy density power converters which rectify three-phase 115Vac to 270Vdc for power distribution.

This work presents an isolated single-stage PWM rectifier with the integration of a current injection path and a full-bridge topology (IS2FBCIP PFC

Rectifier). The operating principle and modulation method are presented. Simulation results are shown to validate the functionality of the topology. The

advantages of the proposed rectifier are the relatively low number of switching components, low implementation effort and its ZVS feature which leads to

a high overall efficiency.

Isolated Single-Stage Three-Phase Full-Bridge

with Current Injection Path PFC Rectifier for

Aircraft Application

Sisi Zhao, Marcelo Silva, Pedro Alou, Jesús A. Oliver

Centro de Electrónica Industrial ( [email protected] )

Proposed Topology

Conclusions• Isolated three-phase single-stage PFC Rectifier: Compact,

without bulky DC link capacitor

• Relatively low number of semiconductors, ZVS in all switches

• Modulation method verified in simulation

• Detailed design guidelines

Operating Principle & Modulation Method

Design Specifications

Modulation Principles:

• Transformer volt-second balance

• Ohmic behavior for each phase

• !" = #

$%&' () ' *+,

• !- = "

-.

#

$% &' (/ 0

(12()

(12(/'

#

$%&' () 3 ' *+,

• !4 = "

-.

#

$% &' (/ 5

(12()

(12(/'

#

$%&' () 3 ' *+,

• !6 = *+, 5 !" 5 !- 5 !4

Buck-Type

Rectifier

# =-

4'

&"

&-'

(7(!

$% &

Modulation Index:

3D CAD Model

Future work• Demonstrator prototype construction

• Experimental verification

• Comparison with other single-stage or two-stage topologies in

terms of efficiency and power density

19


20/224

AbstractThis work presents a new isolated single-stage PWM rectifier system based on the recently presented Swiss rectifier topology providing the isolation by

replacing a buck with a forward converter. The principle of operation and a new modulation technique which compensates the reactive power generated

by the input filter at light load maximizing the power factor are discussed. Furthermore, analytical equations for stresses in semiconductor useful for the

system optimization are derived. The proposed topology

Isolated Swiss-Forward Rectifier

for Aircraft Applications

Marcelo SilvaCentro de electrónica Industrial ([email protected])

Principle of operation

Demonstrator DesignControl scheme

Experimental Result

ConclusionsThis work presents an isolated single-stage three phase rectifier based on the Swiss Rectifier suitable in application with a tight input/output voltage range.

Besides the benefit of the isolation itself, the transformer allows scaling up or down the output voltage of the rectifier maximizing the applicability of the

concept. This work presents a new isolated single-stage PWM rectifier system based on the recently presented Swiss

20


21/224

AbstractMain features to be optimized in aerospace applications are weigth, size and cost. According to

this, Forward with Active Clamp topology could be a good candidate for DC/DC converters in this

ow power applications.

POWER DISTRIBUTION SYSTEM FOR

AEROSPACE APPLICATION (I)

G. Salinas, B. Stevanovic, P. Alou, J.A. Oliver, M. Vasic

Centro de Electrónica Industrial ([email protected])

Topology

Forward with active clamp

Design process1. Analytical modelling of the converter.

2. Semiconductor selection, according to losses.

3. Magnetic design.

4. Output capacitor selection.

5. Selection of the optimum design (losses/volume)

Operation principle

Specifications

Results

Main featuresUp

ci

f

U e

Uds

Qaux

t '

U U e c+

ON OFF ON OFF

• Active clamp technique allows the

core demagnetization during t’ :

• Vds is independent of input

voltage and does not have spikes.

• ZVS can be achieved with the

leakage inductance or an external

inductor.

Vin · D · T = Vclamp · (1 – D) · T

FORWARD WITH ACTIVE

CLAMP

100 kHz 200 kHz

RANGE 1 (50V) RANGE 2 (28V) RANGE 1 (50V) RANGE 2 (28V)

Mosfet Switching (W) 2.169 1.114 4.339 2.228

Mosfet Conduction (W) 0.327 0.605 0.327 0.605

Diode Switching (W) 0.205 0 0.410 0

Diode Conduction (W) 0.902 0.658 0.902 0.658

Output capacitor (mW) /

(mm3)< 1 / 756


22/224

STRACT The design difficulty of DC/DC converters is increased in aerospace applications due to the severe requirements and specifications that must be

ended. The launch cost is very high and there is a permanent interest in efficiency improvement and in the reduction of the mass and volume of the satellites

ipment. High efficiency is important since the primary power source of a satellite is a solar array and batteries, which contribute significantly to the total

ss, volume and cost of the satellite. High reliability is also very important requirement and therefore, a simple power stage with very high efficiency would be

best choice for this kind of applications. Among others, the Flyback with Active Clamp is chosen as the most suitable topology to be analized in detail.

ailed theoretical analysis predicts suitable solutions which can reach 91.81 % efficiency.

Power Distribution System for Aerospace

Application II

B. Stevanovi, G. Salinas, P. Alou, J. A. Oliver, M. Vasi

Centro de Electrónica Industrial ([email protected])

BACK WITH ACTIVE CLAMP

L1

V OUT

L2 /n2

V C LMAG

n:1

I ( Ll e ak ) I ( D2 0 )/ 4 - I (L m ag 1 )

-2.0A

0A

2.0A

SEL>>

/DC CONVERTER FOR AEROSPACE

PLICATION

Output power: 50W

nput voltage (two ranges):

Range 1: 28V (18V 38V)

Range 2: 50V (45V 55V)

Output voltage: 28V ± 5%

Minimum load: 10mA

PRINCIPLE OF OPERATION

IMAG

IL1IDIODE/n

Choice of appropriatesemiconductors in order to

get minimal losses.

DESIGN PROCESS

Design of magnetics obtainingappropriate losses /volume ratio.

Obtain enoughleakage inductance

necessary for ZVS.

CONCLUSIONS The Flyback with Active Clamp represents a trade off between simple, but low

performance, Standard Flyback topology and good performance, but complex, Flyback with Synchronous

Rectification. This topology has a lot of advantages for this kind of application:

Keeps wide regulation capability with medium complexity. Active clamp circuit allows magnetizing

current to be negative, hence DCM is avoided and the output voltage is regulated even under no load

conditions.

ZVS can be achieved with a proper design.

Achieved ZVS reduces the size of the common mode EMI filter at the input of the converter.

× Main drawback of this topology is the high RMS value of the secondary diode current, which directly

influence the size of the output capacitors and, hence the size and weight of the converter.

Flyback with

Active Clamp

100 kHz 200 kHz

RANGE 1

(28V)

RANGE 2

(50V)

RANGE 1

(28V)

RANGE 2

(50V)

MICONDUCTORS

LOSSES [W]2.53 3 3.51 4.58

MAGNETICS

LOSSES [W]1.78 0.93 0.81 0.68

MAGNETICS

OLUME [mm3]2419 2419 2419 2419

OUTPUT

CAPACITORLOSSES [W]

0.17 0.16 0.17 0.16

OUTPUT

CAPACITOR

OLUME [mm3]

4058 2706 4058 2706

TAL LOSSES [W] 4.48 4.09 4.49 5.42

OTAL VOLUME

[mm3]6477 5125 6477 5125

FFICIENCY [%] 91.05 91.81 91.03 89.16

Proper selection of the output capacitor is essential toobtain low losses and volume.

Simulation and verification of the obtained

design.

High efficiency isobtained for both

input voltage

ranges.Calculation of the losses contribution per

component to the total losses of the converter

and looking for the posible improvement of

the efficiency/volume ratio.

22


23/224

AbstractModern aircraft design have a strong tendency towards a More Electrical Aircraft (MEA). Airborne converters

need to have high efficiency and power density (range). This posters shows the design of a 10kW isolated

converter with a series resonant tank in order to have soft transitions in all of the devices. This project is a

collaboration with Airbus and Indra in the AIR (Aircraft Isolated Rectifier) project.

Series Input-Parallel Output 10kW Converter based

on Series Resonant Dual Active Bridge Topology for

High Efficiency Aircraft Application

Yann E. Bouvier, Nicolás Alonso, Pedro AlouCEI-UPM ([email protected])

Topology overview Soft switching is achieved in all transition with the

help of the resonant current shape, which has zero

current at switching.

The series input parallel output double transformer

configuration allows for more transistors in parallel

without driving problems. Also each transformer

turns ratio is decreased by half from 14:1 to 7:1

making the design easier.

Duty cycle is fixed at 50% and output voltage is notregulated.

T1

T3

T2

T4T

VIN

T5

T7

T6

T8

R CLR CR

T5

T7

T6

T8

C

T

VR

t

Ires

Vp

Vs

Vr

Input

Voltage

Output

Voltage

Output

PowerFrequency

400 V 28 V 10 kW 10 kHz

Resonant capacitor voltage is also sinusoidal with minimal noise

at the switching times. The resonant tank current (green) is sinusoidal with some offset

at the zero crossing, caused by the magnetic inductance of the

transformer.

Input bridge voltage (blue) is free of high frequency spikes

because ZVS is achieved in all transitions.

Gate voltage of input transistors (magenta) is also free of noise

because of ZVS

Primary Bridge

Resonant Tank

Secondary Bridges

LR CR R Eq

V1 V2

IRES

Devices Part number Type

Input Capacitors 5x B32656S Metallized Polypropylene Film

Primary Transistors 8x IPW65R037C6

Coolmos Si MOSFET

Resonant Capacitors 13x B32656S2473 Ceramic capacitors

Transformer Cores 2x E80/38/20 3C92 Magnetic Ferrite core

econdary Transistors

32x IPP100N04S2L

Optimos Si MOSFET

Output Capacitors 12x B32774 Metallized Polypropylene Film

Prototype

Equivalent resonant circuit

Experimental results

-50

-40

-30

-20-10

0

10

20

30

40

50

0 2 4 6 8 10

R e s o n a n

t C u r e e n t ( A )

Time (µs)

-60

-50

-40

-30

-20

-10

0

10

1 10 100 1000 10000 100000

A d m i t a n c e ( d B )

Frequency (kHz)

-50

-40

-30

-20

-10

0

10

1 10 100 1000 10000

A d m i t a n c e ( d B )

Frequency (kHz)

Q = 10

Q = 1

Q = 0,1

Design of Resonant TankSimulation of resonant tank quality factor behaviour

Measurement of the real resonant tank The quality factor of the resonant

tank is going to influence in a critical

way the operation of the converter

High quality factor is going to be

achieved by decreasing the total

resistance path of the converter

(transistors included) or increasing

the inductance while decreasing the

capacitance in the resonant tank.

Resonant current and input voltagesResonant current and voltage in capacitor

Efficiency is close to 97% and aluminum

heatsink is place under the converter to

manage all of the components heat

23


24/224

AbstractClean Sky is the most ambitious aeronautical research programme ever launched in Europe. Its mission is to develop

breakthrough technologies to significantly increase the environmental performances of airplanes and air transport. The

goal of this project is to develop a 10kW three-phase buck type rectifier with variable line frequency for applications in

future More Electric Aircrafts.

THREE-PHASE BUCK RECTIFIER FOR AIRCRAFT

APPLICATION WITH VARIABLE LINE FREQUENCY

aU. , bN. Alonso

aCEI-UPM ([email protected])bCEI-UPM ([email protected])

ArchitectureThree Phase Buck-Type Rectifier

Specification

Input Voltage: 230V RMS phase to neutral point

Main frequency: 400Hz

Nominal Bus Voltage (Vo): 400V

Rated output power: 10kW

Military derating 30%

Frequency range 360-800Hz

PF: 0,95 as minimum

THD: Each harmonic has an individual limit

to neutra point

ivi ua imit

EMI Filter

New damping network

Based on the R-C parallel damping with an additional

inductor to reduce the losses

The inductor only conducts the low frequency

harmonics and the damping resistor conducts thehigh frequency harmonics.

Simulation ResultsPower steps Steady state Line frequency steps

5%

5kW


25/224

AbstractPiezoelectric sensors and actuators are the bridge between electronic and mechanical systems in structures. This type of sensor is a key element in the

ntegrity monitoring of aeronautic structures, bridges, pressure vessels, wind turbine blades, and gas pipelines. In this paper, an all inone system for

Structural Health Monitoring (SHM) based on ultrasonic waves is presented, called Phased Array Monitoring for Enhanced Life Assessment. This integrated

nstrument is able to generate excitation signals that are sent through piezoelectric actuators, acquire the received signals in the piezoelectric sensors, and

carry out signal processing to check the health of structures. To accomplish this task, the instrument uses a piezoelectric phasedarray transducer that

performs the actuation and sensing of the signals. The flexibility and strength of the instrument allow the user to develop and implement a substantial

part of the SHM technique using Lamb waves. The entire system is controlled using configuration software and has been validated through functional,

electrical loading, mechanical loading, and thermal loading resistance tests.

Ultrasonicbased Structural Health Monitoring

for aeronautic structures

aEduardo Barrera, aMariano RuizaInstrumentation and Applied Acoustics Research Group (I2A2),

Technical University of Madrid (UPM). Email: [email protected]

ACKNOWLEDGEMENTS

This work was supported in part by the by the Centre for the Development of Industrial Technology (CDTI) under CENIT projects

ICARO, TARGET and PROSAVE2 25


26/224

! " ! # $%

&" "! ' ( )*+,-.

/

+

0

!

12 3

"

4 !

5 $6 7 8! $69 "! 76 9! :

; 5 *:?

!

+"> :=;@" ) .

)ABC?.

D .

=CF

5 G9? )$7.! G8? ).

$75 F;


27/224

AbstractSatelites and other space aplication devices are subject to requirements far more straining than those found in regular applications. Such requirement

usually include temperature ranges that go from -40ºC to 125ºC, as well as vacuum and radiation resistance. In this project the team designed a series of

converters that feed seven different , isolated loads with a very stable voltage level with almost negligible ripple. These coverters are design to face the

harshness of a low orbit environment (800Km) while maintaining correct funcion for all of the mission duration (7 years)

Space application high efficiency and compact

design multiple output DC/DC Converter

Jesús Maañón, Javier López, Miroslav Vasic, Branislav Stevanovic

Constrains and challenges

Isolation between outputs

Every output must be independently grounded. This makes a topology with transformer

mandatory

Optical feedback

In order to keep the voltages in regulation range, the regulation must be made in the output,

with

The information sent to the primary through an optical coupler. This devices are very sentitive

to

Radiation and temperature .

Voltage regulation

Output voltages must be tightly regulated. This makes the main magnetic component a key

device in the converter, as its design will affect heavily the global peformance of the conevrter

Monte Carlo SimulationsSeveral Monte Carlo analysis are done to verify specifications (PM, BW, Zout) in all possible operating conditions. Radiation and temperature are specially

tough in space and thus drifting in nominal values of the components are wider than in any other application.

O u t p u t I m p e d a n c

e [ m O h m ]

Frequency [Hz]

10 100 1k 10k 100k

0

20

40

54

60

40

20

0

20

40

60

35

90

45

0

45

90

Bode Plot Output Impedance

Frequency [Hz]

10 100 1k 10k 100k27


28/224

28


29/224

/8#

.$ #( #

1)# 12 +& ! 1&

.$ # ): 3 1)# 12

+& !

.$ 3 3

(*3B .* 7)2 !% !# ( /&&

# E +F #B ># % !G / 7%2 !%

.$ ;#

3 #

.$ ' ' 4$#3132

8# % !/ 7%2 !%

.$ 4 4 +( 2 2 8

922 +% 8

.$ 5 5 / ' ! ) .), J

+3 $% ", ! # 2 8 922 +%

8.$ 8#


30/224

.$ "!

I% 1&B ! +%

30


31/224

AbstractHundreds of millions of people worldwide are affected by neurological disorders. These include Alzheimer disease, Parkinson's disease or Autistics

disorders among others. Neurological disorders represent a public health problem, with a extraordinary economic importance. Mainly, treatments consist

n drug therapies and surgical interventions but complementary therapies are sought to manage symptoms since drugs or surgery are neither definitive

nor long-term solution. We are exploring acoustical engineering as effective method in order to improve both diagnosis and symptomatology to

complement current treatments, evaluating its influence with typical scales according to disorder and electroencephalography (EEG).

ACOUSTIC ENGINEERING APPLIED TO

NEUROPATHOLOGY: ACOUSTIC THERAPIES

aG. Gálvez, aM. Recuero, aL. Gascó

aInstrumentation and Applied Acoustics Research Group (I2A2-UPM)


Background

Neurological disorders, a big problem

Sound as therapy?

Methodology

Sound design and psychoacustic evaluation

1. Audiometric tests determines a subject's hearing levels

2. Digital sound is synthesized using Matlab with different sound properties

(binaural beats auralization, rhythm, etc.) depending on a specific disease.3. Specific acoustics instrumentation is used to evaluate acoustical properties

of synthesized sound which depends on: player, headphones and sound

signal properties.

HATS Pulse Analyzer

Sound influence measurements in humans

1. Electroencephalography (EEG) data is collected to compute:

1. EEG Power2. Functional brain connectivity

2. Specific scales adapted to every disease (cognition and motor performance

in Parkinson Disease, attention in Autistic Disorders, etc.)

Interventions

Sound analysis

Healthy population

1. Adults didn’t show improvementson cognitive performance under

20 min stimulation. Most

interestingly any subject showed

problems under stimulation.

2. Healthy children showed a better

cognitive performance after 12

weeks of listening program.

Patients

1. Parkinson Disease patiens

reported improvements on

symptomatology.

2. Now we’re designing interventionsfor:

1. Fibromyalgic patients.

2. Autistic patiens

All interventions followed placebo-controlled design

Neurological disorders correspond50% of all disabling disease

Complementary therapies are

sought to manage symptoms

Music has been used

from ancient times

Music Therapy unknows

neurological basisIncreasing interest of

music/sound effects

Approaches without

scientific basis

ContributionsCRESPO, A.; RECUERO, M.; GALVEZ, G.;

BEGOÑA, A.: “Effect of BinauralStimulation on Attention and EEG”, Archives of Acoustics, 38(4): 517-528,

2013.

AcknowledgementsWe thank “Center for BiomedicalTechnology” of UPM and Associations.

31


32/224

AbstractParkinson disease (PD) is 2nd most common neurological disorder implying a high cost to society. Although there is no cure, drug therapy -with important

side effects- controls the physical symptoms successfully, but it loss efficacy over time. For this reason complementary therapies are sought to help when

effect medication decays (OFF-ON phenomena). In this context, we hypothesized that sound stimulation controlled by acoustical engineering methods

could help medication to control symptoms, since there are similar interventions –music or vibrational therapy – known by helping in Parkinson Diseaseexperimentally. We worked with healthy population to improve methodology and later we carried out a study with 12 patients where we found an

mprovement in several symptoms as well as brain activity normalization. Additionally we are working to find biomarkers of PD based on acoustics, due to

there is no lab test for PD, so it is difficult to diagnose, depending of observational procedures.

ACOUSTIC TECHNOLOGIES FOR PARKINSON

DISEASE

aG. Gálvez, aM. Recuero

aInstrumentation and Applied Acoustics Research Group (I2A2-UPM)


Background

Parkinson Disease (PD)

Acoustic engineering in Parkinson Disease

Contributions

Methodology

Participants profiles

HATS

Pulse Analyzer

Challenges & future

CRESPO, A.; RECUERO, M.; GALVEZ, G.;

BEGOÑA, A.: “Effect of BinauralStimulation on Attention and EEG”, Archives of Acoustics, 38(4): 517-528,

2013.

Challenges Future

• Sound could be and

effective co-assistant to

pharmacological

treatment.

• Besides sound as

treatment, diagnosis is a

prospective field using

sound.

• Translation to other

neurological diseases.

TremorRigidity InstabilityBradykinesia

Sound

Music

Melody

Harmony

Rhythm

[…]Noise

Acoustic

Technologies

Speech

[…]

• Good response tomedication

• ! 3 Hoen & Yahr scale

• Normal Hearing

• DBS not implanted

Sound design

1. Digital design sound with Matlab2. Acoustical properties evaluation

HiFi HeadphonesPlayer

Data collection

• Cognition scales

• EEG recording (“FunctionalConnectivity” and EEGPower)

• Gait

• Tremor

EEG

recording

Tremor

Globes Gait

Ethics and analysis

Double blind and placebo-controlled

experimental design following “Principlesof the Declaration of Helsinki” is required.

Statistical pre-post analysis with SPSS and

Matlab.

Symptomatology

PD is 2nd most prevalent

neurological disease

Current treatments don’t stopprogression and side effects are

significant and varied.

Acoustic

engineering

Treatment Diagnosis

Effective sound on PD

Objetives

Medical validation

through longitudinal

studies.

Understanding of

neural processing

mechanisms of

sound.

Acknowledgements

We thank “Asociación de Parkinson deMadrid” and “Center for BiomedicalTechnology” of UPM.

32


33/224

AbstractDry eye disease is an ocular disorder that affects to the eye surface and causes not only view distortion and surface damage , but also changes the

chemical composition of the tear.

A lot of people are affected by dry eye diseases . However, in the current market, there are only few test that detect this pathology in its severe stage.

The aim of this project is, thanks to the label free optical biosensor technology, to be capable to detect the dry eye disorder in its primary stage.

Thanks to the optical biosensor, the changes of the tear composition can be detect. More precisely, through a label free immunoassay, the concentration

of tree specific biomarkers related to this pathology can be know. Moreover, the read-out method that is used for the detection reduces the cost of this

test and the sizes of the devices, becoming a cost-efficient, compact and reliable technology.

DEVELOPMENT OF OPTICAL BIOSENSOR FOR

DRY EYE DISEASE DETECTION

B. Santamaríaa*, A. L. Hernándeza, M. Maiglerb, F. J. Sanzaa, R. Casquela,

A. Lavina, M. F. Lagunaa, M. Holgadoa.

a Center for Biomedical Technology UPM (Campus de Montegancedo), bBio Optical Detection

E-mail adress: [email protected]

Label free optical biosensor

Conclusion:• The compact point of care read-out device is able to detect antigens

concentrations of ng/mL.

• The SU8+NC surface is good performance to be used as a sensing surface.

Future research lines:• Detection of S100A6 and CST4 biomarkers

• Launch of the TearMonitor into de market.

• Future compact point of care devices for other disease detection.

• More information:

Direct immunoassay:

ample collection:

Fabrication of BICELL biosensor: Biomarkers for dry eye

(antigens to detect):

vMMP9

v S100A6

v CST4

v P9

S100A6

v CST4

Tear analysis and results

Optical Read-out based on Increased Relative Optical Power (IROP):

me process for nitrocellulose

esults/ MMP9 recognition:

33


34/224

AbstractThe subharmonic component in the signal scattered by contrast micro-bubbles immerged in a liquid and subject to ultrasound fields is investigated, in

order to provide quantitative estimations of local blood pressure. The problem is formulated by assuming a gas bubble encapsulated by a shell of finite

hickness whose dynamical behavior is modeled by the nonlinear Neo-Hooke constitutive equation. The corresponding problem for free bubbles and two

nterfacial models has also been considered. Pressure thresholds for the onset of the subharmonic and limit values for the transition to a chaotic

response have been determined for different ambient pressures. Thus, for any driving frequency, the interval of acoustic pressures where a subharmonic

component may be expected has been identified. Subsequently, subharmonic resonance curves for different ambient pressures have been performed.

The results obtained show that as the overpressure increases, the resonance curves are displaced to higher frequencies so that, subharmonic amplitudes

ncrease with overpressure for frequencies above the corresponding to subharmonic resonance and decreases for frequencies below resonance. The

opposite behavior is observed when the overpressure is decreased.

Local blood pressure estimations by subharmonic component

analysis of microbubble-scattered signals

aJ. Jiménez-Fernándeza Grupo de Mecánica de Fluidos aplicada a la Ingeniería Industrial. Dpto. Ingeniería Energética,

([email protected])

IntroductionThe knowledge of the ambient blood pressure provide important information regarding the diagnose of pulmonary hypertension, cardiac abnormalities and

other severe diseases. In recent past, research effort has been devoted to the development of the clinical techniques for noninvasive local blood pressure

measures by means of encapsulated microbubbles driven into oscillatory motions by the action of ultrasound fields. Analysis of the subharmonic

component in the scattered signal, has been proposed as a feasible method to local pressure estimation. Indeed, the quantitative dependence of the ½

subharmonic amplitude on the ambient pressure has been recognized by in vitro and in vivo experiments: Shi et al. (Ultrasound Med. Biol. 25 (1999) 275-

283)], Forsberg et al. (IEEE Trans. Ultrason. Ferroelect. Freq. Control 52 (2005) 581-583), Dave et al. (IEEE Trans. Ultrason. Ferroelect. Freq. Control 58 (2011)

2056-2066), as well as, by numerical simulations: Andersen and Jennsen (J. Acoust. Soc. Am. 126 (2009) 3350-3358), Katiyar et al. (J. Acoust. Soc. Am. 129

2011) 2325-2335). Recently, in vitro experiments have shown the feasibility of this procedure to estimate the interstitial fluid pressure in tumors,

Halldorsdottir et al. (Ultrasonics 54 (2014) 1938-1944). In contrast to harmonic components, subharmonic response is only detected for exciting pressures

above a threshold value which depend on the driving frequency, the equilibrium bubble radius and the physical properties of the host fluid and coating. On

he other hand, it should be highlighted that, as the acoustic forcing is increased, the nonlinear response may become chaotic. Thus, the existence of the

subharmonic is restricted to a bounded interval of acoustic pressure. For overpressures in the physiological range 0-190 mm Hg, threshold values of the

ncident acoustic pressure for the onset of the subharmonic have been determined from the power spectrum. In addition, the limit values for the

ransition to a chaotic response have been evaluated by using methods from dynamic systems theory, in particular, by means of a systematic calculation of

he corresponding Lyapunov exponents. Thus, for any driving frequency, the interval of acoustic pressures where a subharmonic component may be

expected has been identified. Once this parameter region of existence has been established, the influence of ambient pressure on the subharmonicntensity has been analyzed. For fixed driving pressure, subharmonic amplitudes has been evaluated in terms of the excitation frequency and ambient

overpressure.

Results:The main results obtained have been summarized in Fig. 1-3 for bubbles with an equilibrium radius R=3 µm . Fig.1 bifurcation diagram (upper plot) and

Lyapunov exponents (lower plot) with the pressure amplitude as control parameter at a driving frequency f=1.897MHz . Fig.2 Pressure thresholds for the

onset of the subharmonic signal (blue) and for chaotic behavior (red) as function of the frequency. Fig.3 Subharmonic amplitude vs frequency and

overpressure for acoustic pressure amplitudes : left 0.375MPa, center 0.35MPa, right: 0.25MPa$ .

Fig. 1Fig.2

Fig.3

34


35/224


36/224

36


37/224

!"#$%&'("#¥ !"#$%&'(#&)*"+)(#& -.#/#'#* .01%*/02 1%''011$%* $#. /30 456 "77#8"*"+)(#& #& 9:;

¥ )1 #-(')* 8"#10&1#.1

¥ ?30 @"7"/ #$ A0/0'(#& "1 )20B%)/0 /# 70)1%.0 17)** 1"+0 7#*0'%*01C :0D0./30*011 "/ "1 1(** &0'011).> /# "&'.0)10 /30 10&1"(D"/> #$ 9:; /# 70)1%.0

0E/.070*> 3).7$%* 7)."&0 -#**%/)&/1F G3"'3 ).0 -.010&/ "& ) D0.> *#G '#&'0&/.)(#&C

H"5%.0 IJ KC 477#8"*"+)(#& )&2 .0'#5&"(#& #$ 8"#7#*0'%*01C !C L3"M #$ /30.01#&)&'0 -0)N "& .0'#5&"(#& -.#'011

)1

8"#*#5"')* 10&1#.1 /3.#%53 /30

20D0*#-70&/ #$ )& 1/)&2).2

"77%)11)>C

90:7,;"2"%" -.0 /.0)/02 "& #.20. /# )**#G )&2 0&3)&'0 /30

)Z)'370&/ #$ /30 8"#..0'-/#.C ?3"1 -.#'011 "1 ')**02

8"#$%&'(#&)*"+)(#& )&2 "&'*%201 2"O0.0&/ 1/0-1 0&'*#102

#& -"'/%.0 WC

H"5%.0 ]J 9:;

H"5%.0 WJ !"#$%&'(#&)*"+)(#& -.#/#'#* )--*"02

H"5%.0 _J ̀ 0.(')* "&/0..#5)(#& *)>#%/

>0?7'&%,'

H"5%.0 [J 901#&)&'0 13"M )1 ) $%&'(#& #$ 8"#7#*0'%*01 )''%7%*)(#&

*??*@A ?BA5C*CD C*C5E1FF*?A *A

G15F5H1!*F ABCA5?A

.*0F0 I7-#J#27KL +?0 !.'M&7%L G0 A.#,.N.-O.L :0 :.(


38/224

AbstractMain idea and mathematical support by JP Gauthier (LSIS – Université de Toulon) – U Boscain (CNRS - France) – M Sigalotti (INRIA - Paris) – D Prandi

Université Paris – Dauphine) – F Rossi (Aix – Marseille Université).

Summary. The goal of the project is to develop a new software-hardware architecture inspired by the structure and the functionalities of the primary

visual cortex in mammals. The model is based on a redundant 3D representation of images, which adds to the spatial data of the image a further local

angular variable. The final objective of the project will be the construction of a prototype for the proposed software-hardware architecture. By its own

nature, such a prototype will be characterized by a high level of parallelism and will be well adapted for tackling several problems that the visual cortex

solves extremely efficiently : inpainting, shape recognition, contour enhancement. . .

Artificial primary Visual

Cortex designLeonardo Suriano, Eduardo de la Torre, Teresa Riesgo

[email protected]

THE HYPOELLIPTIC DIFFUSION

AND HUMAN VISION

THE ALGORITMS ARE INSPIRED BY THE HUMANBRAIN (GeCo Methods)

•Recall the idea of the Citti-Petitot-Sarti sub-Riemannian model of

V1.

•Lift of a curve: how to bring an image from 2D to 3D domain space.

•Reconstruction of the level sets via geodesics (Sub-Riemannian

Geometry).

•Reconstruction of complex images: the hypoelliptic diffusion model.

•The semi-discrete version of the model: reconstruction of mild

corrupted images.

•Dynamic restoration: reconstruction of highly corrupted images.

Thanks to the massive mathematical parallelization

demonstrated, we developed algorithms starting from

standard serial code to CUDA – OpenCL (with an

improvement of 1000x of the elaboration speed on GPUs).

We are going to realize an FPGA hardware version using

the potential of the ARTICo3 and the HLS. The mission is to

build a device that is able to reproduce, in hardware, the

functionalities of the primary Visual Cortex V1.

HIGHLY CORRUPTED IMAGE

INPAINTING THROUGH HYPOELLIPTIC

DIFFUSIONINPAINTING RESULTS

85% CORRUPTED RECONTRUCTION

CAN YOU FIGURE OUT WHAT THERE IS

BEYOND THE CORRUPTED IMAGES?

ARTICO3FPGA - VERSION

ARTICo3 is a bus-based reconfigurable architecture for SRAM based

FPGAs that enables multithreaded hardware acceleration. Being able to

dynamically trade off between computing performance, energy

consumption and dependability at run time makes ARTICO3 really

suitable for real time demos on drones and autonomous robots in which

we can implement and simulate the behavior of the Visual Cortex.

CONCLUSION AND FUTURE WORK

38


39/224

AbstractFPGA architectures are well known to be an efficient option for digital signal processing, especially when a considerable amount of data has to be

processed. In this scenario Brain Computer Interface (BCI) appears as a very good candidate to benefit from this. A BCI is a direct communication pathway

between a person’s brain and an external device. BCIs are often directed at researching, mapping, assisting, augmenting, or repairing human cognitive or

sensory-motor functions.

Bioelectric Signal Acquisition and Processing

on an FPGA

Ramón Conejo, Pablo Iglesias, Javier Mora,

Eduardo de la Torre, Teresa Riesgo

Centro de Electrónica Industrial – Universidad Politécnica de Madrid{ramon.claguna; pablo.iglesias.lopez}@alumnos.upm.es{javier.morad; eduardo.delatorre; teresa.riesgo}@upm.es

System design

• Data acquisition – ADS1299

Brainwave data is acquired using Texas Instruments’

ADS1299, an affordable 24-bit 8-channel analog-to-digital

converter (ADC) designed for electroencephalogram (EEG)

applications.Communication between FPGA and ADS1299 has been

optimized in order to achieve sampling rates up to 16000 Hz

in streaming.

Moreover, the microprocessor embedded on the FPGA is

used to control the configuration of the electrodes

connected to the ADS1299.

• Signal processing

Data processing is performed on a dedicated hardware on

the FPGA. In this research a Virtex-5 FPGA is being used. The

main block of the developed system is an FFT core, which

carries out the frequency decomposition with a high

precision in a matter of microseconds.

In addition, the block allows the application of a Hann

window to the signal before applying the FFT.

Detection of Alpha brainwaves

Many evaluation experiments can be performed. Early experiments

have shown that the system is able to detect alpha waves. These

brainwaves are dominant during resting states of the brain, such as

relax or meditation.

Some other responses that could be tracked are evoked potentials,

which reflect the processing of physical stimuli, or event-related

potential , caused by the “higher” processes, which might involve

memory, anticipation, attention, or changes in the mental state,

among others.

ADS1299 (ADC)

ELECTRODES

SUBJECT STATE

LOGIC MODEL

FEATURES

FAST FOURIERTRANSFORM

(FFT)

HANN WINDOW

SPI

Future stages

• Evolvable hardware – Machine LearningCombining the previous technology with evolvable hardware results inmany possible applications in the field of neurology. One of them is the

development of a self-adaptive collective BCI where the responses

retrieved from multiple users are used to obtain a common response.

Interferences frommains power

Alpha Waves

Signal Processing

BCI - SYSTEM

ADS1299

Signal Processing

FPGA – EVOLVABLE HARDWARE

39


40/224

AbstractLaser Induced forward Transfer (LIFT) is a direct writing technique in which very small quantities of material can be transferred into a substrate. This

material can be either a solid material, a paste or even a liquid. There are a type of viscous liquids called hydrogels or bio-inks that are being used as

medium to transfer or print living cells with several methods. The aim of this investigation is to find a suitable hydrogel or bio-ink that is able to maintain

cell viability during all the printing process. In this process included the culture of the cells, preparation of the hydrogel or bio-ink, LIFT experiment set up,

processing and processing of deposited material, cell proliferation and data analysis.

High accuracy 3D bioprinting using laser

induce forward transfer techniques

aA. Márquez, aS. Lauzurica, aM. Morales, aC. Molpeceres

aCentro Láser, Universidad Politécnica de Madrid, C/ Alan Turing 1, 28038, Madrid, Spain

Email: [email protected]

1. Cell type

Cell line

Primary cell

Stem cell

Natural• Alginate

• Gelatin

• Collagen

• Hyaluronic Acid

• Fibrin

Synthetic• PEG

• Polaxamers

• PLGA

• PCL

2. Hydrogel

Characteristics• Biocompatibility

• Degradation

•

Cross-Linking• Biological integration

• Bio-Replication

• Viscosity

• Hidration

Donor set up

• Radiation

transparent

• Absorption layer

• Liquid deposition

Acceptor set up

• Cushion layer

• Wet medium

Process parameters

• Gap

• Liquid layer

thickness

• Materials

Laser parameters

• Wavelength

• Power

• Spot size

3. Experiment Set Up

Culture

LIFT

2D Construct

3D Construct

4. Processing

Laser

Donor

Acceptor

1. Immediately after printing

• Crosslink of the hydrogel• Insert construct in

appropriate culture

medium

• First cell viability test

2. After cell culture and

crosslinking

• Preparation of new layer

for 3D construct

• Data analysis

5. Post-Processing

• Cell viability (Live/Dead Assay)

• Structural analysis

• Geometry

• Statistics

• Degradation of the hydrogel

• Cell proliferation

• Cell reproduction and dissociation

• Thermal stress

• Genotoxic damage

6. Data analysis

1. M. Xu, X. Wang, Y. Yan, R. Yao, and Y. Ge, “An cell-assembly derived physiological 3D model of the metabolic syndrome, based on adipose-derived

stromal cells and a gelatin/alginate/fibrinogen matrix,”Biomaterials, vol. 31, no. 14, pp. 3868 –3877, 201040


41/224

!"#$%&'$

!"# %&'()*+,&-#. /0+1#2+ 345 64+ 7-#41#. &* 899: 6&1" 1"# 4&2 '; ,#-;'-2&*< -#+#4-7" '* *#6 471=4>'*

2#7"4*&+2+ ;'- 2'1'-(?#++ 4*. H7&4? -&./$+ 4*.

0#&./$+@ 1"# %41%'! 74* ;'?. 4*. =*;'?. &1+ 6&*H7&4? 2=+7?#+ 24.# '; /"4,# M#2'-0 N??'0

E/MN+OA

!"#&- ;=*7>'*&*< &+ &*+,&-#. 50 1"# H+" -#. 2=+7?#+@

1"41 5#*. 1"# H+" 5'.0 '-.#- 1' ,-'.=7# 4 B4-

'; '*.=?41'-0 +6&22&*< 2'.#+A

'-?.G+ H-+1 541(?&D#

!"4*D+ 1' 1"#&- 7'2,?#L

6&*


42/224

AbstractThe characteristics and capacities of sintered stainless steels have been investigated in order to evaluate their ability to retain microorganisms. Three

sintered stainless steels were tested in different culture media using four types of bacteria. In general, the austenitic stainless steels was the most

favourable for this type of study. Regarding the microorganisms, the best results for observation and counting were obtained with Staphylococcus

aureus. The results suggest that PM stainless steel could be a useful surface where planktonic microorganisms can attach and develop biofilms, which

could have environmental applications such as decontamination of polluted water.

BACTERIAL COLONIZATION ON PM STAINLESS STEELS

aJosé M. Ranninger, aManuel Cisneros, aJosé M. Ruiz-Román,aLuis E. Cambronero, bAna M. García, bAndrés Nuñez, bDiego A. Moreno

ETSI Minas y Energía, Universidad Politécnica de Madrid

ETSI Industriales, Universidad Politécnica de Madrid ([email protected])

IntroductionPowder metallurgy (PM) consists in obtaining pieces of powder metal

that are processed at high temperatures and pressure. Due to its

characteristic manufacturing process, the materials can have a specific

and controlled porosity, which makes it possible to obtain porous parts

such as ball bearings, gears, roller bearings, etc. This porosity is what

made us think about how easy biofouling would be on these materialsand its possible environmental applications.

Materials and MethodsThree different PM stainless steels (SS) were studied: a 316L austenitic

SS, a 430L ferritic SS and a 50/50 duplex SS, the latter resulted from

mixing 430L ferritic SS and 316L austenitic SS powders (Table 1). The

mixtures were obtained by dry mixing and die pressing at three

different pressures (300, 500 and 700 MPa) using an uniaxial press, as

well as die lubrication with zinc stearate. Green samples were sintered

in vacuum at 1250 oC for 30 minutes. The densities of the materials

obtained were in the range of 5.4-6.9 g·cm-3. Coupons were polished to

1 m.

ConclusionsOur preliminary results indicated that Staphylococcus aureus grwth has a

direct relationship with the porosity of PM stainless steels. However,

additional investigations are still needed for a thorough understanding

of the biofouling mechanism.

The results suggest that PM stainless steel could be a useful surface

where planktonic microorganisms can attach and develop biofilms. This

could have environmental app

INDUSTRIALES RESEARCH MEETING

Documents

Transcript of INDUSTRIALES RESEARCH MEETING