Wireless Communication in Industrial Networks

1
Wireless Communication in Industrial Networks
Kavitha Balasubramanian Teaching Assistant, CprE
458/558
2
Agenda

• Introduction
• Existing Wireless Standards
• Non real time applications
• Soft real-time applications
• Hard Real-time applications
• Techniques for improving reliability
• The Future

3
Introduction

• Immense development of wireless communication
  technology for consumer electronics
• Also finding its way into industrial setup
• Salient features of communication system in
  industrial application
• Part of production facility
• Loss of production costly in comparison to the
  communication system
• Focus on reliability, predictability and fault
  tolerance

4
Introduction

• How wireless is accepted?
• Technology should have clear benefits and is
  reliable so as to justify investment
• Globally accepted standards has led to mass chip
  production
• Low prices for complex products
• Some of these are used in the industry

5
Motivation for wireless industrial networks

• Reduced installation, reconfiguration and
  maintenance costs
• Easy access to machines for diagnostic or
  programming purposes
• Improved coverage of the factory floor
• Eliminates adverse effects due to damage of
  cabling
• Desire to save on cabling
• Globally accepted standards leading to mass
  production and reduced prices

6
Industrial Applications
7
Workload
Application Requirements
Meeting Deadlines Predictability and Reliability Guaranteed packet delivery Guaranteed delivery times Prioritizing messages
Message Characteristics
Periodic traffic with deadlines Acyclic packets (alarms) with bounded latency Short packets (order of kilobytes)
8
Industrial wireless network market
TEXT
GRAPHICS
INTERNET
HI-FI AUDIO
STREAMING VIDEO
DIGITAL VIDEO
MULTI-CHANNEL VIDEO
LAN
802.11b
802.11a/HL2 802.11g
SHORT lt RANGE gt LONG
Bluetooth 2
ZigBee
PAN
Bluetooth1
LOW lt DATA RATE gt HIGH
9
Existing protocols- comparison
Feature 802.11 Bluetooth Zigbee
Interference from other devices -- Avoided using frequency hopping Dynamic channel selection possible
Optimized for Multimedia, TCP/IP and high data rate applications Cable replacement technology for portable and fixed electronic devices. Low power low cost networking in residential and industrial environment.
Energy Consumption High Low (Large packets over small networks) Least (Small packets over large networks)
Voice support/Security Yes/Yes Yes/Yes No/Yes
Type of Network / Channel Access Mobile / CSMA/CA and polling Mobile Static / Polling Mostly static with infrequently used devices / CSMA and slotted CSMA/CA
Bit error rate High Low Low
Real Time deadlines ??? ??? ???
10
Challenges and Spectrum of Solutions
Wireless Challenges
Attenuation Fading Multipath dispersion Interference High Bit Error rate Burst channel errors
Existing Solutions
Application Requirements
Reliable delivery Meet deadlines Support message priority
11
Existing Wireless Standards

• 802.11
• Oldest and most mature
• Spread spectrum techniques for PHY layer
• Direct Sequence Spread spectrum with Differential
  Binary Phase Shift Keying or Differential
  Quadrature Phase Shift Keying
• Frequency Hopping Spread Spectrum with Gaussian
  Frequency Shift Keying
• Split into
• 802.11a 5GHz ISM band, Up to 2 Mbps
• 802.11b 2.45GHz ISM band, Up to 11 Mbps
• 802.11e Supports QoS
• Uses CSMA/CA MAC (contention based) with dynamic
  packet length up to 4096 bytes long

12
Existing Wireless Standards

• HiperLAN/2
• Support for soft real-time applications like
  media
• Uses Time Division Duplex scheme (Contention
  free)
• Static packet length of 54 bytes
• Better suited for real-time applications because
  of time deterministic MAC
• OFDM modulation
• Multicarrier modulation scheme
• Adapts bit-rate based on channel conditions
• Handles performance reduction due to multipath
  fading radio channels

13
Existing Wireless Standards

• Bluetooth
• Low-complexity low-cost solution for short-range
  wireless communications
• Intended for cable replacement and adhoc
  connections of consumer devices
• Instead of creating protocols in all layers of
  the protocol stack for each application, profiles
  exists
• Subset of protocols serving a certain application
• Uses FHSS with hopping frequency of 1600 hops/sec
• Also supports soft real-time applications like
  multimedia

14
Non real-time applications

• Remote Control
• Used for remote control of overhead cranes
• Dependent on security
• Uses long code words to initiate remote control
  action
• Machine health monitoring
• Accurate information about the status of a
  process
• Locally on demand or over a wireless network to a
  control room
• For local information, use PDA or laptop that
  connects to sensors or actuators
• Plant equipped with access point for
  communication with the control room at the same
  time
• Using Bluetooth is a cheap way of achieving local
  health monitoring

15
Non real-time applications

• System Configuration and Information Exchange
• Information is downloaded to a target device
• Higher demand for fault tolerance
• Use File Transfer Protocol with secure delivery
• Profiles in BT
• Object Exchange protocol includes a file transfer
  protocol
• TCP/IP over Bluetooth transport protocols
• Fault Tolerance
• With Error detection/correction, there is a
  probability that the errors go undetected
  (depends on type of coding method and the code
  rate)
• Repeat message a number of times and use majority
  voting
• For small configuration changes, value is
  uploaded into the device and echoed back. A
  mismatch will initiate a re-transmission
• Configured device can echo the changed value a
  number of times if correct value is received
  all the times, there is a very little chance that
  the configured value is wrong

16
Non real-time applications

• Internet Connectivity
• Used for surveillance, supervision and health
  monitoring of devices
• Demands soft real-time guarantees if applications
  like voice and video need to be supported
• E.g. Distributed supervision of heat and
  ventilation system
• Heat/ventilation system has an embedded web
  servers that runs scripts
• Personnel can supervise and configure the system
  using a PC with a web browser
• Can use 802.11 or BT depending on the bandwidth
  requirements

17
Soft real-time applications

• QoS Parameters include
• deadline delay constraints of the application
• probability of correct delivery within the
  deadline bit error rate of the communication
  channel
• Jitter of the delay

18
Soft real-time applications

• Event Registration
• Time-stamp an event with high time resolution
• Transmission of the event from device that
  registered the event is not so critical i.e.
  requires only a reasonable response time.
• E.g. Events dont control anything but should be
  logged at the right time
• Requires clock synchronization

19
Soft real-time applications

• Measurement
• Measures a physical process, timestamp the sample
  values and transmit sequence of values to the
  user
• Present course of events in a correct way so that
  the time stamps can reconstruct the process
• Notion of global correct time is important
• Requires clock synchronization precision demands
  is dictated by the granularity of the measured
  values
• For e.g. Geological or industrial wireless sensor
  system with sensors collecting data and
  transmitting them to base station or control room
• No delay constraints in transmission but measures
  and timestamps done in real-time to reconstruct
  course of events of the physical process
• High accuracy and synchronized clocks
• 802.11 150ms BT 20us

20
Soft real-time applications

• Media
• Applications like voice and video transfer
• Delay and loss rate constraints based on comfort
  provided by application to the user
• BT
• Supports voice channels
• Range of 10m
• Establishes adhoc connections As soon as 2 or 3
  people are within the range, a communication
  channel is established
• HiperLAN/2
• Supports guaranteed media streams
• Slow control loop
• Control loops used in process control of slow or
  non-critical operations
• Low sample rate which are not affected by a few
  samples being lost
• Delay constraint based on comfort demands
• E.g. heat control and ventilation system
• No time stamping necessary but low jitter
  requirement
• HiperLAN/2 or IEEE 802.11e is suitable

21
Hard real-time applications

• Result presented too late is disastrous to the
  system
• One missed deadline will led to a disaster and
  cannot be tolerated
• E.g. control loops
• Most distributed real-time systems assume
  fault-free operation of the processing unit and
  communication channel that is not true
• Error probability cannot be neglected for
  wireless channel and its function cannot be
  guaranteed at all times
• Errors are sporadic in nature for wireless while
  for wired, permanent errors caused by wire or
  connector damage dominates

22
Hard real-time applications

• Parameters
• Deadline
• Probability to deliver within the deadline
• If we quantify the probability for communication
  channels, we can analyze the error probability of
  the entire system and then employ risk analysis
  methods for calculating failure characteristics
  of the system
• Problem Quantify Probability to deliver within
  the deadline for certain system conditions

23
Techniques for improving reliability of wireless
channel

• Based on radio signals
• Transmitted wave interferes with the surrounding
  environment creating multiple waves hitting the
  receiver antenna
• Waves delayed with respect to each other
• For some conditions, there is destructive
  interference at the receiver antenna causing
  signal attenuation. This is called fading
• Fading causes bursts of errors in wireless
  channel
• Fast fading occurs when the transmitter or the
  receiver move fast. Slow fading occurs in
  stationery situations.
• FEC Add redundant information to the bit stream
  that helps receiver correct channel induced bit
  errors. Used for fast fading channels
• ARQ Retransmit entire packets of data when the
  receiver cannot decode the packet. Used for slow
  fading channels

24
Techniques for improving reliability of wireless
channel

• Soft decision decoding
• Efficiency of decoding depends on chosen codeword
• At receiver, the decoder tries to decode the code
  words into information symbols
• Separate symbols as much as possible to minimize
  probability that decoder interprets one symbol
  for another. These are called maximum distance
  codes. E.g. Reed Solomon codes.
• In soft decision decoding, the distance between
  the code word and symbols is calculated as the
  Euclidian distance rather than the hamming
  distance. No information is discarded during the
  coding process.
• Turbo coding is a recursive scheme that iterates
  a number of cycles to enhance the probability of
  successful decoding

25
Techniques for improving reliability of wireless
channel

• Deadline dependent coding
• Uses FEC and ARQ to improve Bit Error Rate
• Have a number of re-transmissions before the
  deadline retransmissions have different coding
  rate depending on the remaining time to the
  deadline
• Tradeoff between throughput and how much
  redundant information is needed to secure the
  communication link

26
Techniques for improving reliability of wireless
channel

• From release time to deadline, there is a
  transmission time window. The aim is to transmit
  the information within this window and succeed
  with a certain probability
• Send predefined number of transmissions.
  Therefore no need to wait for ACK. Decoder keeps
  packet even if decoding fails and performs
  additional processing such as majority voting.
  Use different codes for different packets to
  conserve energy.
• Fading nature of channel will give error bursts.
  ARQ isolates the error bursts
• Decoder keeps information for future use
  enhancing the decoder efficiency.

27
Ultra wide-Band radio

• Transmitter sends Gaussian pulses with duration
  less than a nanosecond. This is called a Gaussian
  monocycle
• Monocycle is a wide bandwidth signal with center
  frequency and bandwidth dependent on the width of
  the pulse
• Modulation done in time domain by introducing
  different delays between pulses for different
  symbols
• Signal not vulnerable to multipath fading to the
  same extent as signals with carrier frequency
• The bandwidth allows the transmitter to send at a
  very low output power. Hence it does not
  interfere with already allocated bands
• Fear of interference with GPS signals because of
  their relatively low power

28
The Future

• Wireless Technologies will play an important role
  even in safety critical industrial applications
• First step is to implement non-critical
  applications and get wireless technology accepted
  in the industrial sector
• For safety critical applications, more research
  is required
• New wireless technologies with higher throughput
  makes it possible to use complex coding schemes
• These coding schemes handle the insecure wireless
  media in a predictable and secure way