Intra-Vehicular Wireless Sensor Networks
Sinem Coleri Ergen
Wireless Networks Laboratory,
Electrical and Electronics Engineering,
Koc University
Outline
 Motivation for Intra-Vehicular Wireless Sensor Networks
 Physical Layer Design
 Medium Access Control Layer
 Conclusion
 Current Projects at WNL
Outline
 Motivation for Intra-Vehicular Wireless Sensor Networks
 Physical Layer Design
 Medium Access Control Layer
 Conclusion
 Current Projects at WNL
History of In-Vehicle Networking
 Early days of automotive electronics
 Each new function implemented as a stand-alone ECU, subsystem
containing a microcontroller and a set of sensors and actuators
 Data exchanged between point-to-point links
Body Control
Module
ECU
ECU
sensor
sensor
History of In-Vehicle Networking
 In the 1990s
 Increase in the number of wires and connectors caused weight, cost,
complexity and reliability problems
 Developments in the wired communication networks
Body Control
Module
ECU
sensor
ECU
sensor
ECU
actuator
sensor
History of In-Vehicle Networking
 In the 1990s
 Increase in the number of wires and connectors caused weight, cost,
complexity and reliability problems
 Developments in the wired communication networks
 Multiplexing communication of ECUs over a shared link called bus
Body Control
Module
ECU
sensor
ECU
sensor
ECU
actuator
sensor
History of In-Vehicle Networking
 Today
 Increases in number of sensors as electronic systems in vehicles are
replacing purely mechanical and hydraulic systems causes weight, cost,
complexity and reliability problems due to wiring
 Advances in low power wireless networks and local computing
Body Control
Module
sensor
sensor
sensor
ECU
ECU
sensor
ECU
ECU
sensor
sensor
sensor
sensor
actuator
sensor
sensor
History of In-Vehicle Networking
 Today
 Increases in number of sensors as electronic systems in vehicles are
replacing purely mechanical and hydraulic systems causes weight, cost,
complexity and reliability problems due to wiring
 Advances in low power wireless networks and local computing
 Intra-Vehicular Wireless Sensor Networks (IVWSN)
Body Control
Module
sensor
sensor
sensor
ECU
sensor
ECU
ECU
sensor
sensor
sensor
sensor
actuator
sensor
sensor
First IVWSN Example: Intelligent Tire
Active Safety Systems
•Change the behavior of vehicle in pre-crash
time or during the crash event to avoid the
crash altogether
•Examples: Anti-lock Braking System (ABS),
Traction Control System (TCS), Electronic
Stability Program (ESP), Active Suspension
System
Requires accurate and fast estimation of
vehicle dynamics variables
Enable a wide range of new applications
•Forces, load transfer, actual tire-road friction,
maximum tire-road friction available
Intelligent Tire
•More accurate estimation
On-board sensors + indirect estimation
•Even identify road surface condition in
real-time
IVWSN: Distinguishing Characteristics
 Tight interaction with control systems
 Sensor data used in the real-time control of mechanical parts in different
domains of the vehicles
 Very high reliability
 Same level of reliability as the wired equivalent
 Energy efficiency
 Remove wiring harnesses for both power and data
 Heterogeneity
 Wide spectrum for data generation rate of sensors in different domains
 Harsh environment
 Large number of metal reflectors, a lot of vibrations, extreme temperatures
 Short distance
 Maximum distance in the range 5m-25m
Outline
 Motivation for Intra-Vehicular Wireless Sensor Networks
 Physical Layer Design
 Medium Access Control Layer
 Conclusion
 Current Projects at WNL
Wireless Channel Measurements
 Building a detailed model for IVWSN requires
 Classifying the vehicle into different parts of similar propagation
characteristics
 Collecting multiple measurements at various locations belonging
to the same class
passenger compartment
engine
trunk
beneath chassis
Wireless Channel Model: Beneath Chassis
 81*18 measurement points at
 Two different vehicles: Fiat Linea and Peugeot Bipper
 Different scenarios: engine off, engine on, moving on the road
Channel Model: Large Scale Statistics
 Path loss model
Channel Model: Large Scale Statistics
 General shape of impulse response: Saleh-Valenzuela Model
inter-arrival time
of clusters
cluster amplitude
ray decay rate
Channel Model: Small Scale Statistics
 Characterized by fitting 81 amplitude values to
alternative distributions
Channel Model: Simulation Results
 Qualitative comparison
experimental power delay profile
simulated power delay profile
 Quantitative comparison
Outline
 Motivation for Intra-Vehicular Wireless Sensor Networks
 Physical Layer Design
 Medium Access Control Layer
 Conclusion
 Current Projects at WNL
Medium Access Control Layer: System Requirements
 Packet generation period, transmission delay and
reliability requirements: (Tl ,dl ,rl )
 Network Control Systems: sensor data -> real-time control of
mechanical parts:
 Automotive system, no automatic way of validating the performance of
control systems for different (Tl ,dl ,rl )
-> use extensive simulations of closed loop models for a given (Tl ,dl ,rl )
 Main characteristics:
 Fixed determinism better than bounded determinism in control systems
 As
Tl
increases, the upper bound on
dl decreases down to zero
Medium Access Control Layer: System Requirements
 Adaptivity requirement
 Nodes should be scheduled as uniformly as possible
EDF
Uniform
Medium Access Control Layer: System Requirements
 Adaptivity requirement
 Nodes should be scheduled as uniformly as possible
1
EDF
Uniform
Medium Access Control Layer: System Requirements
 Adaptivity requirement
 Nodes should be scheduled as uniformly as possible
2
EDF
Uniform
Medium Access Control Layer: System Requirements
 Adaptivity requirement
 Nodes should be scheduled as uniformly as possible
3
EDF
Uniform
Medium Access Control Layer: System Model
(Tl ,dl ,rl ) given for each link l
 T1 £ T2 £ ... £ TL

 Choose subframe length as T1 for uniform allocation
 Assume Ti /T1 = si is an integer: Allocate every si subframes
 Uniform distribution
º
minimize max subframe active time
max active time=0.9ms
EDF
max active time=0.6ms
Uniform
✓
Medium Access Control Layer: One ECU
Maximum active time of subframes
Periodic packet generation
Delay requirement
Energy requirement
Maximum allowed power by UWB regulations
Transmission time
Transmission rate of UWB for no
concurrent transmission case
Medium Access Control Layer: One ECU
 Optimal power and rate allocation is independent of optimal
scheduling
 One link is active at a time
 Given transmit power, both time slot length and energy
minimized at maximum rate
 Maximum rate and minimum energy at
and
Medium Access Control Layer: One ECU
 Optimal scheduling problem decomposed from optimal power
and rate allocation: Mixed Integer Programming Problem
Periodic packet
generation
Maximum active time of subframes
 NP-hard: Reduce the NP-hard Minimum Makespan Scheduling
Problem on identical machines to our problem.
Medium Access Control Layer: One ECU
 Smallest Period into Shortest Subframe First (SSF) Scheduling
 2-approximation algorithm
Medium Access Control Layer: One ECU
 SSF Scheduling:
Medium Access Control Layer: One ECU Simulations
 Use intra-vehicle UWB channel model
 Ten different random selection out of
predetermined locations
Medium Access Control Layer: Multiple ECU
 How to exploit concurrent transmission to multiple ECUs to
decrease the maximum active time of subframes?
 Allow concurrent transmission of sensors with the same packet
generation period -> fixed length slot over all frame assignment
What is the power, rate allocation and resulting length of time slot
if they are combined?
How to decide which nodes are combined?
Medium Access Control Layer: Multiple ECU
 Optimal power allocation for the concurrent transmission of n
links: Geometric Programming Problem
-> Power control needed in UWB Packet based networks
Delay requirement
Energy requirement
Transmission time=
packet length/
rate of UWB for
concurrent transmission
Medium Access Control Layer: Multiple ECU
 Which slots to combine?
-> Mixed Integer Linear Programming problem
 Propose Maximum Utility based Concurrency Allowance
Scheduling Algorithm
 Define utility of a set: decrease in transmission time when
concurrent
 In each iteration, add the node that maximized utility
 Until no more node can be added to increase utility
Medium Access Control Layer: Multiple ECU
Outline
 Motivation for Intra-Vehicular Wireless Sensor Networks
 Physical Layer Design
 Medium Access Control Layer
 Conclusion
 Current Projects at WNL
Conclusion
 Intra-vehicular wireless sensor networks
 Increases in number of sensors causes weight, cost, complexity and
reliability problems due to wiring
 Advances in low power wireless networks and local computing
 Physical layer
 Large scale-statistics: path loss, power variation
 General shape of impulse response: Modified Saleh-Valenzuela model
 Small-scale statistics
 Medium access control layer
 Adaptivity requirement: Minimize maximum active of subframes
 Tight interaction with vehicle control systems
 Delay, energy and reliability requirements
 One ECU: 2-approximation algorithm
 Multiple ECU: Utility based algorithm to decrease subframe length
Publications
 Y. Sadi and S. C. Ergen, “Optimal Power Control, Rate Adaptation and
Scheduling for UWB-Based Intra-Vehicular Wireless Sensor Networks”, IEEE
Transactions on Vehicular Technology, vol. 62, no. 1, pp. 219-234, January 2013.
[pdf | link]
 C. U. Bas and S. C. Ergen, “Ultra-Wideband Channel Model for Intra-Vehicular
Wireless Sensor Networks Beneath the Chassis: From Statistical Model to
Simulations”, IEEE Transactions on Vehicular Technology, vol. 62, no. 1, pp. 14-25,
January 2013. [pdf | link]
 U. Demir, C. U. Bas and S. C. Ergen, "Engine Compartment UWB Channel Model
for Intra-Vehicular Wireless Sensor Networks", IEEE Transactions on Vehicular
Technology, vol. 63, no. 6, pp. 2497-2505, July 2014. [pdf | link]
Outline
 Motivation for Intra-Vehicular Wireless Sensor Networks
 Physical Layer Design
 Medium Access Control Layer
 Conclusion
 Current Projects at WNL
Current Projects
Intra-Vehicular Wireless Sensor
Networks
•Supported by Marie Curie Reintegration Grant
Energy Efficient Robust
Communication Network Design for
Wireless Networked Control Systems
•Supported by TUBITAK (The Scientific and
Technological Research Council of Turkey)
Energy Efficient Machine-to-Machine
Communications
•Supported by Turk Telekom
Cross-layer Epidemic Protocols for
Inter-vehicular Communication
Networks
•Supported by Turk Telekom
RSSI Fingerprinting based Mobile
Phone Localization with Route
Constraints
•Supported by UC Berkeley
Intra Vehicular Sensor Networks
•Supported by TOFAŞ
People
• Sinem Coleri Ergen
Director
Ph.D.
Students
M.S.
Students
• Yalcin Sadi | Seyhan Ucar | Merve Saimler | Elif Dilek Salik | Utku Demir | Ali
Vosoughi | Bugra Turan | Melih Karaman
• Anique Akhtar | Bakhtiyar Farayev
• Umit Bas | Nabeel Akhtar | Irem Nizamoglu | Mehmet Kontik
Alumni
Thank You!
Sinem Coleri Ergen: sergen@ku.edu.tr
Personal webpage: http://home.ku.edu.tr/~sergen
Wireless Networks Laboratory: http://wnl.ku.edu.tr