Secure Cooperative MIMO Communications Under Active Compromised Nodes

1
Dr. Liang Hong LHONG_at_TNstate.edu (615)
963-5364
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure Cooperative MIMO Communications Under
Active Compromised Nodes
Liang Hong, McKenzie McNeal III, Wei
Chen College of Engineering, Technology, and
Computer Science Tennessee State University
2
Outline
College of Engineering, Technology, and Computer
Science Tennessee State University

• Introduction
• System Model
• Compromised Nodes Detection and Symbol Recovery
• Cooperative Transmission and Security Key
  Management Schemes
• Simulations
• Conclusions

3
Introduction
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• Wireless sensor networks (WSNs) have been widely
  deployed in
• Military sensing and tracking, environment
  monitoring, smart home appliances management,
  health care, etc
• WSNs are expected to be the basic building block
  of pervasive computing
• Cooperative MIMO can achieve the benefits of MIMO
  technique without the need of multiple antennas
  at each sensor node
• MIMO can provide significant increases in data
  rate and link range without additional bandwidth
  or transmission power
• Physical implementation of multiple antenna at a
  small node is not feasible
• Distributed individual single-antenna nodes
  cooperating on data transmission and reception as
  a multi-antenna MIMO node

4
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• Many WSNs have mission-critical tasks, however,
  the involvement of multiple nodes for
  transmission and/or receiving poses a challenge
  to the reliability of the information
• Node compromise is one of the most detrimental
  attacks
• Active attack the compromised nodes maliciously
  modify the relay information and inject falsified
  information (more severe than passive attacks)
• Previous security schemes either did not detect
  and defend against node compromise or needed
  extra MIMO antennas to achieve data assurance.

5
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• Our previous work proposed a cross-layer security
  scheme that combined a cryptographic technique
  implemented in higher layer with data assurance
  analysis at the physical layer
• Has much less bit errors
• Receiving cluster periodically detects
  compromised nodes
• The transmitting and receiving cluster have equal
  numbers of nodes
• Only one compromised node is present in the
  transmitting cluster
• No compromised node is present in the receiving
  cluster

Research objectives
Develop compromised node detection algorithm for
more general scenarios.
6
System Model
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes
Multi-hop cooperative MIMO system
Cooperative strategy decode-and-forward
7
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes
Received signal
(1)
received signals at the receiving cluster
transmitted signals at the transmitting cluster
channel matrix (mR?mT)
additive Gaussian noise components, they are
identically distributed and mutually
statistically independent, each with a zero mean
and a power spectral density
8
Compromised Nodes Detection and Symbol Recovery
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• Methodology
• The CH h of B will perform compromised nodes
  detection at random time t in each time interval
  with adaptive security level

9
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• Compromised nodes detection algorithm
• The CH h of B requests each of the primary nodes
  in B to send the received symbols to h. Then h
  obtained the combined signal
• Node h uses Inverse Channel Detector to estimate
  the transmitted symbols s, i.e., ,
  where W is an mR?mT weighing matrix
• Since all data streams are the same, h can
  identify the compromised node x if x didnt
  transmit the supposed symbol.
• For example, assuming that the primary node x in
  A is compromised and it is the jth node in A, if
  sj from step 2 is not the same as the information
  recovered from the majority of the other nodes, x
  will be classified as a compromised node

10
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes
Scenario 1 The transmitting and receiving
clusters have different number of nodes
(2)

• Scenario 2 More than one compromised node is
  present in the transmitting cluster
• In step 3, the recovered data streams from
  different transmitting nodes will be sorted into
  groups, where nodes are assigned to the same
  group if they contain identical symbols.
• The group with largest number of nodes is assumed
  to contain the IDs of trustworthy nodes. All the
  other nodes are classified as compromised nodes.
• If all the groups have the same number of IDs,
  all the nodes in transmitting cluster are
  classified as compromised nodes.

11
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• Scenario 3 Compromised nodes are present in the
  receiving cluster
• The CH h in B performs inverse channel detection
  on the set of received data from each node in B.
• If two or more of the estimates of the
  transmitted symbols are the same, then this node
  in B is not compromised. Its ID will be saved in
  the trustworthy node group.
• If none of the estimates of the transmitted
  symbols are the same, then this node in B is the
  compromised node.
• After all the compromised nodes in B are
  identified, all the trustworthy nodes in B will
  be used to identify the compromised nodes in A.
  Using the algorithm in Scenario 2, the
  compromised nodes in A can be detected

12
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes
Symbol recovery

• When the compromised nodes are detected in the
  transmitting cluster
• the CH h in B will send each node in its cluster
  the information of the compromised nodes in A.
• When the node v in B receives this information,
  it decodes the message by simply setting the
  columns in channel matrix that corresponds to the
  compromised nodes to zero
• When the compromised nodes are detected in the
  receiving cluster
• The CH h in B will send the IDs of the
  trustworthy nodes to the CH in its next relay.
• In the next relay, simply setting the column in
  channel matrix that corresponds to these
  compromised nodes to zero.
• If compromised nodes are detected in the last
  receiving cluster, the sink node will use the
  received data from the trustworthy nodes for
  symbol detection

13
Cooperative Transmission and Security Key
Management Schemes
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• The cooperative transmission and security key
  management schemes proposed in our previous work
  will be used to provide secure cooperative
  communications.
• This security scheme provides secured
  communication between uncompromised nodes.
• By combining the compromised nodes detection
  algorithm, the security of the system will be
  largely enhanced.

14
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes
Type of Keys (1) Each cluster each node has a
cluster key C-key(A) for local transmission (2)
Each link AB each node in A and B has a key
L-key(A,B) for long-haul transmission between A
and B
Secured cooperative relay Step 1 (Local
transmission at A) Each node in A encrypts its
information with C-key(A), and broadcasts it to
other local nodes using different timeslots. Each
node uses C-key(A) to decrypt the received m
information back.
Step 2 (long-haul transmission between A and B)
Each node i in A encrypts sequence I with key
L-key(A,B), and it acts as ith antenna encoding
the encrypted I. Then, all m nodes in A broadcast
the encrypted and encoded I to the nodes in B at
the same time. When a nodes in B receives m
copies of the information, it decrypts them with
L-key(A,B), and then decode them back to I.
15
Simulations
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• The channels are block Rayleigh fading channels.
• It is constant during the transmission of one
  symbol, but is randomly changing between symbols
• Different channels are identically distributed
  and statistically independent.
• Binary phase shift keying (BPSK) is chosen as the
  modulation scheme.
• 100 received symbols are used in the proposed
  algorithms for compromised nodes identification.
• The maximum likelihood detector is used for
  symbol demodulation

16
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• For scenario 1 and 2

• Detection Accuracy
  Performance Comparison

17
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• For scenario 1 and 3

• Detection Accuracy
  Performance Comparison

18
Conclusions
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes

• Algorithms for compromised nodes detection are
  proposed for three more general scenarios.
• The proposed compromised nodes detector are then
  used in a cooperative MIMO communication system
  to enhance the security.
• The compromised nodes detector has high detection
  accuracy.
• The reliability of the information is
  significantly improved.

19
College of Engineering, Technology, and Computer
Science Tennessee State University
Secure cooperative MIMO communications under
active compromised nodes
Thank you! Questions?