IC1004  /  Newcom#  School:  

"Beyond  4G  Networks  in  Cities:  From  Theory  to  Experimentation  and  Back”    

November  25  -­‐  28,  2013  CTTC,  Castelldefels,  Spain  

                                             

Abstracts  of  Lectures  and  Description  of  Lab  Sessions  

 

 

Werner  Mohr  (NSN),  T1:  Research  Challenges  for  B4G  Mobile  and  Wireless  Communications  Systems  

Mobile  and  wireless   communication  are  growing   further  globally  and  will  offer  a  plethora  of   services   to  users.  Broadband  mobile  communication  systems  like  LTE  are  currently  being  deployed.  Communication  networks  are  also  increasingly  be  used  for   vertical   sectors   to   manage   utility   services   like   energy,   gas   and   water   systems   and   are   critical   infrastructures   for   our  societies   and   economies.   In   particular   these   applications   are   based   on   sensor-­‐,   IoT-­‐   or  M2M-­‐based   systems  with   a   huge  number  of  connected  devices,  which  will  be  part  of  future  networks.  These  developments  are  resulting  in  exponential  traffic  growth  in  the  coming  years.  Future  networks  will   increasingly  be  based  on  heterogeneous  radio  access  systems  from  short  range  to  wide  range  communications,  from  low  to  broadband  throughput  rates  and  from  low  to  higher  possible  latencies  in  order  to  support  all  envisaged  application  scenarios.  

The  expected  traffic  growth  and  the  huge  number  of  connected  devices  results   in  challenging  requirements  on  the  energy  efficiency   of   future   systems   in   order   to   keep   the   energy   and   CO2   footprint   of   future   systems   as   low   as   possible.   The  availability   of   sufficient   frequency   spectrum   is   a   major   concern   to   support   the   necessary   system   capacity.   Additional  challenges  are  significantly  reduced  latency  for  improved  user  experience,  heterogeneous  systems,  self-­‐organizing  networks  to  cope  with  the  system  complexity,  a  more  flexible  use  of  the  frequency  spectrum  and  cloud-­‐based  architectures.  Cognitive  networks  are   complementing   future   systems   in  order   to   improve   reliability,  quality  and   to   reduce  operational   and  energy  cost.  This  is  creating  huge  challenges  for  research,  design  and  deployment  of  future  communication  networks.  The  network  architecture  will  have  to  undergo  significant  changes  to  meet  requirements  also  to  support  economic  service  provision.  It  will  become  more  flat,  virtualization  of  network  resources  and  cloud  concepts  reduce  capital  (CAPEX)  investment  and  operational  expenditures  (OPEX)  and  will  allow  faster  and  easier  introduction  of  new  services  and  thereby  innovation  in  service  offering.  Such  Software-­‐Defined  Networking  (SDN)  concepts  in  combination  with  cognitive  systems  will  support  self-­‐organized  systems  in  order  to  manage  the  overall  network  complexity.  

Future   communication   networks   require   a   holistic   approach   in   research,   standardization,   regulations   and   system  development.   Research   initiatives   have   started   in   different   regions   of   the  world.   In   Europe   the   new   Framework   Program  Horizon  2020  will  address  this  area  in  coordinated  research  activities  and  in  particular  in  a  5G  Public  Private  Partnership  on  communication  network  infrastructure.  

The  presentation  will   present   the   trends   towards   future   systems,   the  expected   research   challenges  and   the  upcoming  5G  Public-­‐Private-­‐Partnership.  

 

Fabio  Dovis  (PoliTO),  T2:  Localisation  by  Satellite  Systems  The  class  will  provide  the  basic  principles  of  Satellite  Based  localisation,  describing  the  signals  and  the  main  stages  of  the  receivers.  The  presentation  will  address  also  the  main  factors  limiting  the  performance  and  it  will  discuss  the  opportunities  for  the  hybridisation  with  terrestrial  communication  networks  to  extend  a  reliable  positioning  service  to  harsh  environments.    

 

Chiara  Buratti  (UniBO),  T3:  On  the  Impact  of  Medium  Access  Control  Protocols  on  Multi-­‐Hop  Networks  

The  lecture  will  be  split  into  two  parts:  the  first  part  will  provide  some  basics  on  Medium  Access  Control  (MAC)  and  routing  protocols   for  wireless   sensor  networks   (WSNs),   and  a   second  part  which  will   be   focused  on   the   IEEE  802.15.4  and  Zigbee  standard  and  on  smart  city  applications.  

In   particular,   in   the   first   part   some   of   the   most   suitable   contention-­‐based   MAC   protocols   (e.g.,   CSMA/CA)   and   routing  protocols  (e.g.,  AODV),  for  WSNs  will  be  described  in  details.  The  second  part  of  the  lecture  will  focus  on  protocols  for  smart  city  applications,  having  the  IEEE  802.15.4  and  Zigbee  standard  as  reference.  Some  case  studies,  providing  some  real-­‐world  implementations,  will  be  described  and  numerical  results  achieved  through  experimentation  on  the  field  will  be  shown.  The  latter  will  provide  some  insight  on  the  interplay  between  theory  and  experimentations  and  the  impact  of  MAC  on  the  routing  protocols  and  on  the  topologies  generated  in  the  networks.    

Reiner  Thomae  (Ilmenau),  T4:  Experimental  Propagation  Studies  Using  Wideband  Real-­‐Time  MIMO  Channel  Sounding  

Multidimensional  sounder  system  architecture,  choice  of  transmit  signals,  estimation  of  delay  Doppler  response,  sounding  antennas  and  antenna  arrays,  application  of  sounding  data,  high  resolution  parameter  estimation,  multidimensional  data  model,  iterative  ML  parameter  estimation,  antenna  data  model.  

 

Nicola  Baldo  (CTTC),  T1:  Simulation  of  LTE  Systems  in  ns-­‐3  

This  lecture  is  aimed  at  researchers  who  are  interested  in  the  simulation  of  LTE  systems  for  their  research  activities,  and  who  have  no  prior  experience  with  ns-­‐3.  The  lecture  will  first  provide  a  generic  overview  of  the  ns-­‐3  simulator,  discussing  its  architecture,  its  main  features,  and  its  models  library.  After  this  introduction,  the  LTE  module  of  ns-­‐3  will  be  presented,  discussing  its  modeling  assumptions  first,  and  then  examining  its  components  (radio  protocol  stack,  PHY  model,  MAC  Scheduler  API  &  models,  EPC  model,  buildings  &  propagation  models).  

 

Nikolaos  Bartzoudis  (CTTC),  T1  Seminar:  "Prototyping  the  Physical  Layer  of  Modern  Wireless  Communication  Systems:  Development  Flows,  Challenges  and  Pitfalls"  

The  processing  complexity  of  a  PHY-­‐layer  prototype  may  vary  depending  on  the  target  scenario  that  needs  to  be  developed  and  experimentally  validated.  A  critical  factor  towards  this  end  is  if  the  prototype  should  feature  offline,  semi  real-­‐time  or  full  real-­‐time  operation.  Other  parameters  that  play  an  important  role  in  the  low-­‐level  design  and  implementation  of  the  prototype  include  the  signal  bandwidth,  the  use  of  multi-­‐antenna  schemes,  the  receiver  mobility  and  the  degree  (or  rate)  that  a  signal  needs  to  be  re-­‐formatted/re-­‐constructed  in  order  to  yield  adaptive/flexible  communications.  The  influence  of  the  mentioned  factors  and  parameters  needs  to  be  thoroughly  analyzed  in  order  to  choose  the  most  adequate  prototyping  methodology  which  could  be  either  based  on  a  software  (e.g.  SDR),  hardware-­‐software,  or  hardware  (e.g.,  FPGA  or  ASIC  design)  approach.  Moreover,  the  specifications,  operating  conditions  (including  assumptions  and  simplifications)  and  configuration  of  the  selected  experimental  setup  greatly  affect  the  validity  of  the  conducted  benchmarking  for  each  prototype.  Apart  from  the  low-­‐level  design,  implementation  and  performance  objectives  the  researcher/engineer/student  needs  to  deal  with  other  heterogeneous  challenges  such  as  the  learning  curve  of  the  selected  methodology,  the  development  time/cost  and  the  modularity/reusability/scalability/portability  of  the  prototype.  Additionally,  industry  trends  (market  manipulation)  or  popular  academic  initiatives  usually  impair  the  correct  selection  of  a  suitable  methodology  tailored  for  the  specific  needs  of  a  research  and  development  project.  Thus,  the  goal  of  this  seminar  is  to  give  an  overview  of  the  different  approaches  used  across  the  academic  community  to  prototype  the  Physical  Layer  of  wireless  communication  systems  trying  to  highlight  the  associated  challenges  and  limitations  that  apply  in  each  case.  Likewise  it  is  intended  to  convey  the  message  that  each  PHY-­‐layer  prototype  requires  a  certain  development  flow  and  thus  a  specific  research/engineering  modus  operandi.    

Jesus  Alonso-­‐Zarate  (CTTC),  T1  Seminar:  "Machine-­‐to-­‐Machine  Technologies  and  Smart  Cities"  

The  unprecedented  communication  paradigm  of  machine-­‐to-­‐machine  (M2M),  facilitating  24/7  ultra-­‐reliable  connectivity  between  a  prior  unseen  number  of  automated  devices,  is  currently  gripping  both  industrial  as  well  as  academic  communities.  The  aim  of  this  tutorial  is  to  provide  an  academic,  technical  and  industrial  insight  into  latest  key  aspects  of  wireless  M2M  networks,  with  particular  application  to  smart  cities  and  smart  grids.  We  will  provide  an  in-­‐depth  introduction  to  the  particularities  of  M2M  systems,  and  then  dwell  in  great  depths  on  the  capillary  and  cellular  embodiments  of  M2M.  The  focus  of  capillary  M2M  will  be  on  IEEE  (.15.4e)  and  IETF  (6LoWPAN,  ROLL,  COAP)  standards  compliant  low-­‐power  multihop  networking  designs;  furthermore,  for  the  first  time,  low  power  Wifi  will  be  dealt  with  and  positioned  in  the  eco-­‐system  of  capillary  M2M.  The  focus  of  cellular  M2M  will  be  on  latest  activities,  status  and  trends  in  leading  M2M  standardization  

bodies  with  technical  focus  on  ETSI  M2M  and  3GPP  LTE-­‐M;  furthermore,  we  will  discuss  analytical  and  simulation  works  quantifying  the  performance  and  impact  of  M2M  in  legacy  cellular  networks.    

Along  the  entire  tutorial,  challenges  and  open  issues  will  be  identified,  thus  making  the  material  presented  in  this  tutorial  useful  for  industry  and  inspiring  for  researchers  and  academics  alike.  

 

Claude  Oestges  (UCL),  T4:  Reference  and  Standardized  Channel  Models  

This  lecture  will  propose  an  overview  of  recent  standardized  and  reference  models  for  beyond  4G  networks  in  urban  environments.  In  particular,  the  lecture  will  highlight  how  most  models  essentially  rely  on  a  combination  of  experimentation  and  smart  data  processing.  Covering  and  comparing  the  recent  modeling  efforts  in  WINNER  and  COST  2100,  the  lecture  will  also  point  out  at  the  various  modeling  challenges.  

 

Florian  Kaltenberger  (Eurecom),  T4:  Physical  Layer  Abstraction  for  LTE  

Physical  layer  abstraction  is  the  process  of  modeling  the  performance  of  the  physical  layer  (in  terms  of  block  error  rates  or  throughput)  as  a  function  of  the  radio  channel  without  running  the  time  consuming  MODEM  and  the  channel  convolution.  Such  models  are  useful  for  two  different  purposes:  Firstly  they  can  be  used  in  the  implementation  of  a  user  equipment  (UE)  to  compute  the  feedback  (channel  quality  information  0  CQI)  and  secondly  they  can  be  used  in  large-­‐scale  system  level  simulations  to  speed  up  simulation  time.  

In  this  1.5h  course  we  will  study  the  basics  of  physical  layer  abstraction  with  the  application  example  of  turbo  codes  in  LTE.  We  will  cover  both  Exponential  effective  SINR  mapping  (EESM)  and  Mutual  Information  based  SINR  mapping  (MIESM).  We  will  describe  the  calibration  process  in  detail  and  give  some  practical  examples  using  the  OpenAirInterface  simulation  tools.  We  will  conclude  the  course  by  applying  the  learned  techniques  to  the  LTE  feedback  calculation  and  fast  link  adaptation  and  system  level  simulations.  

 

Ronald  Raulefs  (DLR),  T2:  Network  Based  Localisation  

Recent  years  have  seen  an  explosion  of  location  based  services.  On  the  other  hand,  the  limitations  of  GPS  for  indoor,  urban  canyons  etc.  have  led  to  an  evolution  of  existing  networks  (e.g.  LTE)  to  provide  network  based  location  estimation.  Today’s  and  future  mobile  radio  devices  provide  a  heterogeneous  portfolio  of  radio  access  technologies.  We  provide  an  overview  of  recent  advances  in  cooperative  communication  based  location  estimation,  fingerprinting  techniques  that  also  work  in  non  line-­‐of-­‐sight,  the  exploitation  of  inertial  modalities  and  self-­‐learning  techniques  to  enable  context-­‐aware  network  adaptions  based  on  inferred  radio  maps.  

 

Roberto  Verdone  (UniBO),  T3:  B4G  Networks  and  the  Smart  2020  City  

Smart  City  applications  need  to  rely  on  flexible  and  high  density  radio  access  networks  deployed  in  the  urban  context.  Proper  integration  of  short  range,  B4G  cellular  and  multi-­‐hop  wireless  techniques,  is  required.  The  first  part  of  the  lecture  will  discuss  the  technical  and  economic/practical  aspects  of  the  deployment  of  radio  network  infrastructures  for  Smart  City  applications.  In  the  second  part,  an  interactive/brainstorming  session  will  be  dedicated  to  the  definition  of  the  possible  technological  options  and  the  open  research  problems  that  need  to  be  addressed  within  Horizon2020.  The  lecture  will  serve  somehow  as  a  recap  of  the  various  topics  addressed  during  the  school.  

 

Carles  Fernandez  (CTTC),  Javier  Arribas  (CTTC)  ,  T2:  Hybrid  localisation:  from  satellite  to  heterogeneous  localisation  techniques  for  Beyond  4G  Networks  

Attendees  to  Track  T2  will  be  exposed  to  a  hands-­‐on  approach  to  the  signal  processing  involved  in  satellite-­‐based  navigation  receivers,  working  with  real-­‐life  signals  and  replacing  the  traditional,  'black-­‐box'  integrated  circuit  receiver  by  GNSS-­‐SDR,  an  open  source  project  that  implements  a  global  navigation  satellite  system  software  defined  receiver  in  C++.  With  GNSS-­‐SDR,  users  can  build  a  GNSS  software  receiver  by  creating  a  graph  where  the  nodes  are  signal  processing  blocks  and  the  lines  represent  the  data  flow  between  them.  The  software  provides  an  interface  to  different  suitable  RF  front-­‐ends  and  implements  the  entire  receiver’s  chain  up  to  the  navigation  solution.  Its  design  allows  any  kind  of  customization,  including  interchangeability  of  signal  sources,  signal  processing  algorithms,  interoperability  with  other  systems,  output  formats,  and  offers  interfaces  to  all  the  intermediate  signals,  parameters  and  variables.  This  allows  a  scrutinization  of  all  the  stages  

involved  in  GNSS  signal  processing,  provides  a  clear  view  of  how  a  GNSS  receiver  works  internally  and  constitutes  a  platform  for  further  research  and  experimentation.  

Attendees  are  supposed  to  be  familiar  with  basic  digital  signal  processing  concepts  and  techniques,  as  well  as  having  some  previous  exposure  to  C++  and  software  design  patterns.  Programming  will  not  be  required,  but  familiarity  with  the  language  could  make  the  course  more  profitable.  A  background  in  localisation  systems  is  desirable  but  not  a  requirement,  since  all  the  key  aspects  will  be  explained  in  detail  before  the  experimental  activity.  

 

Chiara  Buratti  (UniBo),  T3:  IoT  component  in  the  2020  City:  From  MAC  to  Routing  Aspects

During  the  laboratory  experience  students  will  use  the  EuWIn  facilities  and  in  particular  the  FLEXTOP  platform,  composed  of  100  IEEE  802.15.4  /  Zigbee  devices  deployed  in  a  corridor  at  the  University  of  Bologna.  The  platform,  in  fact,  is  remotely  accessibly  through  a  VPN  connection.  Students  will  have  the  opportunity  to  modify  a  software  implementing  a  smart  city  application,  download  the  modified  software  on  the  devices  in  the  FLEXTOP  platform  and  wait  for  the  results,  to  process  them.  

The  throughput  in  a  multi-­‐hop  point-­‐to-­‐point  network  will  be  first  computed  through  experimentations.  While  as  a  second  experiment,  a  smart  city  application  will  run  on  28  devices.  Results  in  terms  of  topologies  generated  in  the  network  and  packet  loss  rate  will  be  analyzed  and  processed.The  first  experiment  will  mainly  aim  at  demonstrating  the  importance  of  the  interplay  between  theory  and  experimentations,  while  the  second  experiment  will  be  much  more  focused  on  the  interplay  between  MAC  and  routing.  

UniBo  will  provide  a  virtual  machine  to  be   installed  on  the  PCs   for   the  LAB  session.The  virtual  machine  could  be  based  on  Oracle  Virtual  Box   (preferred)  or  VMware  VM  player  depends  on  your  requirements   (see  below).  The  PCs  were  the  virtual  machines  (VM)  will  be  installed  (for  a  maximum  number  of  5,  assuming  to  have  5  groups  composed  of  3  people  each),  should  be   able   to   access   to   Internet,   in   particular   a   VPN   session  will   be   started   on   each  VM.  One   possible   solution   is   that   CTTC  provides  5  PCs  with   the  VM  environment   installed  and   the   Internet   connection.  Alternatively,   students   can  use   their  own  laptops,  however  CTTC  should  provide  to  students  a  proper  Internet  connection  which  ensures  to  establish  a  VPN  session.  

 

Christian  Schneider  (TU  Ilmenau),  T4:  Derivation  of  Large  Scale  Parameter  for  WINNER  channel  models  based  on  Data  Sets  from  the  Ilmenau  Reference  Scenario  (Urban  Macro  Cell)  

So  called  large  scale  parameters  (LSPs)  are  the  scenario  dependent  control  parameters  for  stochastic  geometry  based  channel  models  as  from  the  SCM  or  WINNER  family.    To  provide  the  models  with  reliable  statistical  distributions  of  these  parameters    is  important  to  derive  them  based  on  channel  sounding  data  sets  from  appropriate  scenarios.  Within  the  lab  session  the  theoretical  background  as  well  as  measurement  data  related  aspects  will  be  highlighted.  The  experimental  part  will  cover  the  implementation  and  derivation  of  the  required  power  und  delay  domain  parameters  as  path  loss,  shadow  fading,  K-­‐factor,  delay  spread  and  XPR  as  well  as  their  de-­‐correlation  und  cross-­‐correlation  statistics.