Adaptive Network Paper List and Categorization

1
Adaptive Network Paper List and Categorization
2
Categorization

• MAC
• Change schedule
• Lei
• S-MAC
• W. Ye, J. Heidemann, and D. Estrin, "An
  energy-efficient mac protocol for wireless sensor
  networks," IEEE INFOCOM, 2002
• T-MAC
• van Dam and K. Langendoen, "An adaptive
  energy-efficient mac protocol for wireless sensor
  networks," SenSys 2003
• D-MAC
• Gang Lu, Bhaskar Krishnamachari and Cauligi
  Raghavendra, "An Adaptive Energy-Efficient and
  Low-Latency MAC for Data Gathering in Sensor
  Networks," IEEE WMAN 2004
• Jun
• Mihail L. Sichitiu, Cross-Layer Scheduling for
  Power Efficiency in Wireless Sensor Networks,
  IEEE INFOCOM 2004
• Change data transmission rate
• Lei
• CARA
• Jongseok Kim, Seongkwan Kim, Sunghyun Choi, Daji
  Qiao, "CARA Collision-Aware Rate Adaptation for
  IEEE 802.11 WLANs," IEEE INFOCOM, 2006

3

• Network
• Ariffin
• Probabilistic techniques for adaptability in
  network layer
• Gianni Di Caro, Marco Dorigo "AntNet Distributed
  Stigmergetic Control for Communications
  Networks", Journal of Artificial Intelligence
  Research 1998 2.
• David Braginsky, Deborah Estrin "Rumor Routing
  Algorithm for Sensor Networks," ACM WSNA 2002
• Jun
• Yan Yu, Ramesh Govindan and Deborah Estrin,
  Geographical and Energy Aware Routing A
  Recursive Data Dissemination Protocol for
  Wireless Sensor Networks, UCLA Computer Science
  Department Technical Report UCLA/CSD-TR-01-0023,
  May 2001.
• One presentation out of the two
• Well-known Internet adaptive routing
• Yutaka
• ???
• Irene
• D. G. Andersen, H. Balakrishnan, M. F. Kaashoek,
  and R. Morris. "Resilient overlay networks",
  Proc. of the 18th Annual ACM Symposium on
  Operating Systems Principles, Banff, Canada,
  Octoboer 2001.

4

• Transport
• Ariffin
• Mario Gerla, Bryan K. F. Ng, M. Y. Sanadidi,
  Massimo Valla, Ren Wang "/ TCP Westwood with
  adaptive bandwidth estimation to improve
  efficiency/friendliness tradeoffs " ,/ Computer
  Communications, Volume 27, Issue 1, 1 January
  2004, Pages 41-58.
• Keita
• R2CP
• H.-Y. Hsieh and R. Sivakumar, A Receiver-centric
  Transport Protocol for Mobile Hosts with
  Heterogeneous Wireless Interfaces, Proc. ACM
  MOBICOM, San Diego, CA, Sept. 2003.
• Jun
• Transport of sensor net
• Chieh-Yih Wan, Andrew T. Campbell, Lakshman
  Krishnamurthy, PSFQ A Reliable Transport
  Protocol for Wireless Sensor Networks, ACM WSNA
  2002.

5

• Application
• Data dissemination
• Ryota P2P
• V. Kalogeraki, D. Gunopulos, and D.
  Zeinalipour-Yazti, "A local search mechanism for
  peer-to-peer networks," In Proc. CIKM, 2002
• Irene - WSN
• F. Ye, H. Luo, J. Cheng, S. Lu, and L. Zhang, "A
  Two-Tier Data Dissemination Model for Large-scale
  Wireless Sensor Networks", Proc.of ACM MOBICOM,
  Atlanta, GA, September 2002.
• Topology management
• Ryota
• Y. Chawathe, S. Ratnasamy, L. Breslau, N. Lanham,
  and S. Shenker, Making Gnutella-like P2P Systems
  Scalable, In Proceedings of ACM SIGCOMM 2003
• Target tracking
• Jun presented
• Wensheng Zhang and Guohong Cao, Optimizing Tree
  Reconfiguration for Mobile Target Tracking in
  Sensor Networks, IEEE INFOCOM 2004.

6

• Data processing
• In-network data processing
• Jidong identifies two famous papers and Yutaka
  presents one of them
• R. Ahlswede, N. Cai, S.-Y. R. Li, and R. W.
  Yeung, "Network Information Flow", IEEE
  Transactions on Information Theory, IT-46, pp.
  1204-1216, 2000
• Keita
• X. Fu, W. Shi, A. Akkerman, and V. Karamcheti,
  "CANS Composable, Adaptive Network Services
  Infrastructure," USENIX Symposium on Internet
  Technologies and Systems (USITS), March 2001.
• End system data processing (e.g., Adaptive
  video/audio coding)
• Keita
• I. Akyildiz, Y. Altunbasak, F. Fekri, and R.
  Sivakumar, AdaptNet An Adaptive protocol suite
  for the next generation wireless internet, IEEE
  Communications Magazine, vol. 42, no. 3, pp. 128-
  136, March 2004.
• Active networks
• Etc

7

• ContextWare Support for Network and Service
  Composition and Self-adaptation Network
  Composition A Step towards Pervasive Computing
  Keita
• An Adaptable Network Architecture for Multimedia
  Traffic Management and Control - Keita
• CANS Composable, Adaptive Network Services
  Infrastructure Keita
• An Adaptive Architecture for Multi-stream
  Authoring and Presentations in Distributed
  Networks - Keita
• AIDA Adaptive Application-Independent Data
  Aggregation in Wireless Sensor Networks - jidong
• Adaptive Protocols for Information Dissemination
  in Wireless Sensor Networks lei
• An Adaptive Web Page Recommendation Service -
  Yutaka
• Adapting to Network and Client Variability via
  On-Demand Dynamic Distillation Yutaka
• Adaptive situation aware mobile computing
  Yutaka
• Neuro Grid Ryota
• GIA Ryota
• PlanetP Ryota
• Bittorent Ryota

8

• Coverage
• Jun can identify classic papers but may not
  present
• Backcasting Adaptive Sampling for Sensor
  Networks jidong
• Power Conservation and Quality of Surveillance in
  Target Tracking Sensor Networks - Jun
• Dynamic Clustering for Acoustic Target Tracking
  in Wireless Sensor Networks Jun
• Jun can identify mobile sensor papers about
  coverage but may not present
• How about mobile sensor papers in other field???
• Bio
• Using Adaptive networks to model and control drug
  delivery - Yutaka

9

• What to adapt to? read into the paper and
  create slides following the new template
• Network
• available resources (CPU, memory, channel speed),
  heterogeneity, traffic
• Failures (node/link)???
• User
• user requirements (QoS), popularity, mobility
• Why adapt?
• efficiency, robustness, reliability,
  connectivity, scalability, load balancing
• Survivability/reproduce/fitness
• Collision avoidance in robotics
• How to adapt? Open loop/close
  loop/optimization
• Adapt to certain predetermined conditions
• Autonomously adapt
• Local communication among entities
• Long term (evolution)/short term (learning)

10

• Three opinions
• Layer by layer can do now
• What to adapt to? (need more time)
• Hybrid Layer by layer first (can do now), then
  sub-categorize using the most appropriate
  criteria (need more time).
• Dont know which one to go.

11
MAC layer
12
Flexible Power Scheduling for Sensor Networks -
Jun

• B. Hohlt, L. Doherty and E. Brewer
• IPSN 2004

13
What adapts?

• Node scheduling
• Duty cycles

14
What to adapt to?

• Traffic demands

15
How

• To reduce power consumption while supporting
  fluctuating demand in the network for data
  collection.
• It schedules transmitting and receiving time
  slots ineach nodes power schedule and sleeps
  during idle periods.
• Local power schedules dynamically adapt to
  network changes.
• The protocol supports data collection
  applications
• Routing is assumed to be done
• Communication is assumed to be primarily one-way,
  toward a base station, and there may be multiple
  basestations.

16
Cross-Layer Scheduling for Power Efficiency in
Wireless Sensor Networks - Jun

• Mihail L. Sichitiu
• North Carolina State University
• INFOCOM 2004

17
What adapts?

• Node scheduling
• States sample, receive, transmit, idle

18
What to adapt to?

• Data flow information

19
How

• The research investigates the problem of
  power-saving of data transmission
• By dynamically setup schedules of nodes on the
  path from the source to the destination
• The setup and reconfiguration phase takes place
  during the initialization of the network, and
  subsequent to any changes in the network queries
  and the availability of the routes.
• The steady state phase takes place between
  consecutive setup and reconfiguration phases.
• The steady state phase utilizes the schedule
  established in the setup and reconfiguration
  phase to forward the data to the base station.

20
CARA Collision-Aware Rate Adaptation for IEEE
802.11 WLANs Lei Zan

• Jongseok Kim, Seongkwan Kim, Sunghyun Choi, and
  Daji Qiao
• ECE, Iowa State University
• IEEE INFOCOM'2006

21
What adapts?

• A MAC protocol to adjust transmission rate

22
What to adapt to?

• Successful transmission, collision and channel
  error

23
How to do Adaptability

• This paper deals with the issue, when to decrease
  the transmission rate, given multiple
  transmission rates available in 802.11 WLAN
• Existing open-loop rate adaptation does not
  differentiate frame collision from frame
  transmission error caused by channel errors
• So rate adaptation malfunctions when collisions
  exist
• A collision-aware rate adaptation scheme is
  proposed
• Use RTS/CTS and Clear Channel Assessment (CCA) to
  differentiate collision from channel errors
• Rate is adapted based on three thresholds,
  consecutive success threshold, consecutive
  failure threshold, probe activation threshold (to
  active RTS/CTS probe)

24
An adaptive Energy-Efficient and Low-Latency MAC
for Data Gathering in sensor networks Lei Zan

• Gang Lu, Bhaskar Krishnamachari, Cauligi S.
  Raghavendra
• EE, USC
• Int. Workshop on Algorithms for Wireless, Mobile,
  Ad Hoc and Sensor Networks (WMAN)

25
What adapts?

• Sensor node duty cycles (active slots in each
  interval)

26
What to adapt to?

• Traffic load of sensor nodes

27
How to do adaptation

• DMAC mainly tackles issue the data forwarding
  interruption problem
• This paper proposes a scheme, called DMAC
• DMAC adjusts node duty cycles (sending and
  receiving) adaptively according to the traffic
  load by varying the number of active slots
  scheduled in an interval

28
PMAC An adaptive energy-efficient MAC
protocolfor Wireless Sensor Networks Lei Zan

• Tao Zheng, Sridhar Radhakrishnan,Venkatesh
  Sarangan
• CS, University of Oklahoma
• IEEE IPDPS05

29
What adapts?

• Scheduling of nodes wakeup

30
What to adapt to

• Traffic of a sensor node and its neighbors

31
How to do adaptation

• In this paper, the sleep-wakeup scheduling of
  sensor nodes are adaptively determined
• The scheduling is decided based on a nodes own
  traffic and that of its neighbors

32
ASCENT Adaptive Self-Configuring sEnsor Networks
Topologies - Jidong

• Infocomm 02

33
What adapt

• Participation in the multi-hop network topology

34
What to adapt to?

• environmental dynamics and terrain conditions
  which lead to non-uniform communication density

35
How

• Problem
• To propose Adaptive Self-Configuring topology
  management scheme for wireless sensor network
• Solution
• each node assesses its connectivity
• Each node make its state decision to participate
  in network topology
• Changing between test, active, passive and
  sleep states
• E.g.
• Signals when it detects high message loss,
  requesting additional nodes in the region to join
  the network in order to relay messages.
• Reduces its duty cycle if it detects high message
  losses due to collisions.
• Probes the local communication environment and
  does not join the multi-hop routing
  infrastructure until it is helpful to do so.

36
Network Layer
37
An adaptive network routing strategy with
temporal differences - Yutaka
Internet routing
38
What adapts?

• Routing table in Internet

What to adapt to?

• Varying network conditions, e.g., traffic pattern
  and topology (link/node failure)

39
How

• Goal
• Providing more efficient routing algorithm.
• High level idea
• By applying temporal differences method that is
  an incremental learning algorithm utilizing
  differences between successive states.

40
Rumor Routing Algorithm For Sensor Networks -
Ariffin

• David Braginsky Deborah Estrin

41

• Overview
• This research focuses on improving robustness of
  sensor event attribute retrieval.
• Assumes that sensing data can be identified by
  attributes
• What adapts
• Dynamic Node Resources
• Scheme does not rely on some active-nonactive
  pattern for sensors. Robust to sensor failure.
• Queries may find attributes even if some sensors
  are off which was on before
• How does it adapt
• Paths to a certain attribute (information) is
  updated periodically by agents

42
Dynamic Adaptive Routing for a Heterogeneous
Wireless Network - Jidong

• Mobile Networks and Applications 9, 219233, 2004

43
What adapts?

• Architecture and Routing protocol adapt to
• heterogeneous wireless networks (integrating
  cellular, WLAN networks)

44
What to adapt to?

• Heterogeneity of two different networks
• Compatibility of protocols and applications.
• Routing
• QoS requirements

45
How

• Problem
• integrates cellular network with an ad hoc
  network in WLAN
• To allow mobile users versatile communication
  with anyone or any device at any place and
  anytime
• reserves advantages of sizable coverage in a
  cellular network and high data rate in deployable
  ad hoc network
• High level idea (2-3 sentence)
• Proposed an integrated architecture of a
  Heterogeneous Wireless Network (HWN) and
• a dynamic adaptive routing protocol (DARP) for a
  HWN.
• Compatibility of protocols and applications
• Transparent networking
• Effective user location management
• QoS support

46
Energy Sensitive Routing in Ad hocNetworks -
Jidong
47
What adapt

• Routing

48
What to adapt to?

• energy consumption
• energy residue of intermediate nodes

49
How

• Problem
• To propose an energy aware routing protocol for
  ad hoc network
• Solution
• Battery residual information is exchanged during
  routing process
• Choose the routes which has maximum battery
  residual

50
Transport Layer
51
Application Layer
52
Neuro Grid - Ryota

• Overview
• P2P system to provide the functionality of
  keyword search for web pages
• What adapts
• Routing path of a query searching for resource
• What to adapt to
• user feedback/preference

53

• How it does the adaptation
• NeuroGrid nodes keep history information of each
  neighboring nodes about how many good results
  they returned.
• Measure
• of user clicking on the returned URL / of
  recommendations
• Users returns a feedback (whether user really
  click on the returned URL) into the system, which
  is used to update history
• NeuroGrid nodes change history information to
  decide where to forward a query to

54

• Reference
• Neurogrid Web site
• http//www.neurogrid.net
• Neurogrid whitepaper
• http//www.neurogrid.net/php/whitepaper.php
• Adaptive Routing in Distributed Decentralized
  Systems NeuroGrid, Gnutella Freenet, Sam
  Joseph
• http//www.neurogrid.net/php/si-simulation03.zip

55
GIA - Ryota

• Overview
• GIA is intended to design Gnutella-like P2P
  system that can handle much aggregate query
  rates.
• GIA adopts base design of Gnutella, and add
  several adaptive schemes on top of it to improve
  scalability
• Dynamic Topology adaptation
• Change the topology such that nodes with high
  degree actually have capacity to handle many
  queries
• An active flow control scheme
• Avoid overloaded hot spots by assigning
  flow-control tokens to nodes based on available
  capacity.

56

• What adapts to what
• Dynamic Topology adaptation
• Overlay topology adapts to heterogeneous node
  capability
• In order to balance node capacity and degree
• An active flow control scheme
• Topology adapts to nodes available resource to
  balance load

57

• How it does the adaptation
• Dynamic Topology adaptation
• Node want to maintain the following optimal state
• Nodes capacity total of (its neighboring
  nodes capacity / degree)

Capacity 20, of links 4
Capacity 10, of links 5
B
C
5
2
A
Capacity 10 2(B) 5(C) 3(D)
3
Capacity 9, of links 3
D
58

• Dynamic Topology adaptation
• If the optimal state is not satisfied, a node
  tries to add a link to a new neighboring node
• A node (say X) selects the node (say Y) with max
  capacity from a list of candidate nodes
• X starts handshake process with Y, and If Y
  agrees, X can establish a link to Y
• Y compares X with existing neighboring nodes
• If X has higher capacity and lower degree than
  any existing neighboring node, then Y agree to
  take X, and replace the existing neighboring node

59

• An active flow control scheme
• A node periodically assign flow-control tokens to
  its neighbor.
• One token permission to send me one query
• Node X can send a query to Node Y only when Y
  receives a toke from X
• of tokens are proportional to nodes capacity
• Proactively prevent overload

10 queries
20 queries
2 queries
5 queries
Y
Z
Y
Z
20 token
10 token
2 token
5 token
X
X
overloaded
60

• An active flow control scheme
• Topology adapts to node available resource

61

• Reference
• Y. Chawathe, S. Ratnasamy, L. Breslau, N. Lanham,
  and S. Shenker, Making Gnutella-like P2P Systems
  Scalable, In Proceedings of ACM SIGCOMM 2003

62
PlanetP - Ryota

• Overview
• PlanetP is a P2P research project in Rutgers
  University based on ad-hoc network.
• To explore the construction of a reliable peer to
  peer content search and retrieval service
• no centralized directories or managements, no
  complex distributed data structure such as hash
  table across the entire community to support the
  name-based object location, using randomly
  circulated global state between peers (gossip
  algorithm)

63

• What adapts to what
• Unpopular Content Replication
• Content distribution is adaptive to popularity (
  of requests)

64

• How it does the adaptation
• Peers collaborate to replicate unpopular
  contentpopular content is naturally replicated
  via caching
• Periodically, every Tr seconds, each peer
  randomly selects a file in either its data store
  or its replication store. The peer then checks
  for the availability of the file using its set of
  Bloom filters.
• If the availability is sufficient (the available
  peer number i and a threshold k), then it does
  nothing.
• If the availability is deemed insufficient, it
  actively replicates this file.
• Specifically, the replicating peer fragments the
  file using a variation of the Reed Solomon
  erasure coding, where nm fragments are created
  but only n are required to reconstruct the file.
• Finally, it pushes the nm fragments to peers
  chosen using hints about available space in the
  peers' replication store.

65

• Reference
• Francisco Matias Cuenca-Acuna, Richard P. Martin,
  and Thu D. Nguyen. PlanetP Using Gossiping and
  Random Replication to Support Reliable
  Peer-to-Peer Content Search and Retrieval.
  Technical Report DCS-TR-494, Department of
  Computer Science, Rutgers University, july 2002.

66
BitTorrent - Ryota

• Overview
• BitTorrent provide efficient download support of
  files for multiple users
• In BitTorrent, a file is slices into pieces.
• Users download a file piece by piece in a
  collaborative manner where a user who downloaded
  a piece uploads the piece for other users to
  download
• Users follow tit-for-tat strategy to prevent
  free-riders

67

• What adapts to what
• Downloading rate of users (gain) adapts to
  uploading rate of users (contribution)
• How it does the adaptation
• Suppose that user Y is downloading a piece from
  User X
• X may choke bandwidth to Y if Y is not uploading
  other pieces to other users

68

• Reference
• BitTorrent protocol specification
• http//www.bittorrent.org/protocol.html

69
ContextWare Support for Network and Service
Composition and Self-adaptationNetwork
Composition A Step towards Pervasive Computing -
Keita
70
What adapts

• Overlay
• Logical overlay network for each user (called
  Ambient Network)
• Similar to Personal Area Network (PAN)

71
Adapts to what

• Adapts to user's context
• e.g., location, network type etc.

72
How

• A superpeer monitors the user context and respond
  to context change by modifying its Ambient Network

73
An Adaptable Network Architecture for Multimedia
Traffic Management and Control - Keita
74
What adapts

• Application traffic
• I.e., QoS class of application traffic

75
Adapts to what

• Adapts to resource constraints

76
How

• Routers automatically identify a class of the
  application's traffic based on the resource
  availability
• The class is used for QoS Diffserv

77
An Adaptive Architecture for Multi-stream
Authoring and Presentations in Distributed
Networks - Keita
78
What adapts

• Application data
• The quality of multimedia stream content
  (video/audio)

79
Adapts to what

• Adapts to predefined quality requirements

80
How

• Quality requirements are predefined in Quality on
  Demand Markup Language
• When streaming multimedia content, its quality is
  modified based on the specified quality

81
CANS Composable, Adaptive Network Services
Infrastructure - Keita

• Surveyed by Keita

82
What adapts

• Overlay / topology
• Data path between a client and a service

83
Adapts to what

• Adapts to environmental changes
• E.g., buffer size change, link/node failures etc.

84
How

• Three ways proposed
• Intra-Component Adaptation using Distributed
  Events
• Upon detecting an event, each driver and service
  may change its policy to react to the event
• Data Path Reconfiguration and Error Recovery
  using Semantic Segments
• Upon detecting an event, the plan manager
  inserts, deletes or reorders drivers along an
  active data path
• Planning and Global Reconfiguration
• Upon receiving a connection request, the plan
  manager creates a plan to determine how to deploy
  a data path

85
An Adaptive Web Page Recommendation Service -
Yutaka
Web technology related
86
What adapts?

• Results of web page recommendation service.

What to adapt to?

• User request and similarity of users.

87
How

• Goal
• Providing web recommendation service that returns
  more personalized result.
• High level idea
• There are two typical approach to this problem
  i) analyzing document content, ii) taking the
  similarity between users into consideration.
• They propose a new system where two approaches
  mixed together.

88
Adapting to Network and Client Variability via
On-Demand Dynamic Distillation - Yutaka
89
What adapts?

• Data retrieved via HTTP

What to adapt to?
What to adapt to?
What to adapt to?

• Client variations, e.g., effective bandwidth,
  processing power, processing power, and ability
  to handle specifc data encodings

90
How

• Problem
• Difficulty in providing appropriate level of
  service that is sensitive to client variations.
• High level idea
• Implementing a HTTP proxy that
• Intercept HTTP requests from users with slow
  links
• Applies datatype-specific lossy compression
  (e.g., downsampling images in the frequency
  domain, and dividing long HTML pages into small
  parts)

91
Adaptive situation aware mobile computing -
Yutaka
Wireless-mobile computing
92
What adapts?

• Application software working on client

What to adapt to?

• Change in environmental state, e.g., change of
  client location and application specific data.

93
How

• Goal
• Providing application adaptability to the change
  of environmental state.
• High level idea
• Implementing environmental-aware API
• API provides models of several events common to
  clients and mechanism to deliver events to
  application.

94
Adaptive Protocols for Information Dissemination
in Wireless Sensor Networks Lei Zan
Joanna Kulik, Wendi Rabiner, Hari EECS, MIT ACM
Mobicom' 1999
95
What adapts?

• Information Dissemination

96
What to adapt to?

• Available resources (energy)

97
How to do adaptation

• This paper attempts to improve the following
  deficiencies introduced by the classic flooding
  approach to disseminate information between
  sensors
• Implosion waste energy by sending redundant data
• Overlap waste energy by sending overlap data to
  common neighbors
• Resource blindness no adaptation to available
  energy
• This paper proposed SPIN protocols to efficiently
  disseminate information among sensor nodes by two
  key innovations
• Negotiation eliminate data redundancy
• Use of meta-data eliminate overlap problem
• Resource are monitored and activities are adapted
  to available resources
• e.g. when energy is below a threshold, certain
  activities involved in data forwarding are cut
  back

98
AIDA Adaptive Application-Independent Data
Aggregation in Wireless Sensor Networks - AIDA

• ACM Transactions on Embedded Computing System,
  Special issue on Dynamically Adaptable Embedded
  Systems, 2004.

99
What adapt

• communication protocols for sensor networks
• Degree of aggregation at MAC layer

100
What to adapt to?

• Constraints
• bandwidth, energy, and throughput, etc.
• Dynamics
• unpredictable traffic patterns
• dynamic network topologies

101
How

• Problem
• Perform efficient and application independent
  data aggregation to improve end-to-end
  information throughput
• Solution
• Proposed an adaptive aggregation layer module
• resides between the data-link and networking
  layer
• aggregate packets through network unit
  concatenation.
• combines network units into a single outgoing
  AIDA payload to reduce the overhead incurred
  during channel contention and acknowledgment
• No semantics of the data in the network units are
  used.
• Aggregation decisions are made in accordance with
  an adaptive feedback-based packet-scheduling
  scheme
• dynamically controls the degree of aggregation in
  accordance with changing traffic conditions

102
Conflict-Free Motion of Multiple Mobile Robots
Based on Decentralized Motion Planning and
Negotiation - Ariffin

• Kianoush Azarm and Gunther Schmidt

103

• Overview
• Robots have specific destination goals.
• Knows static map
• What adapts
• Robots Adapts
• To avoid collision
• To share narrow spaces
• How does it adapt
• Robots sense when in proximity with other robots
  and negotiate on a one to one basis.

104
Coverage Layer
105
Dynamic Clustering for Acoustic Target Tracking
in Wireless Sensor Networks - Jun

• Wei-Peng Chen,
• Jennifer C. Hou, and Lui Sha
• IEEE TRANSACTIONS ON MOBILE COMPUTING, VOL. 3,
  NO. 3, JULY-SEPTEMBER 2004

106
What adapts?

• Monitoring structure
• Cluster formation

107
What to adapt to?

• Target movement

108
How

• A hierarchical sensor network that is composed of
  
• a static backbone of sparsely placed
  high-capability sensors which will assume the
  role of a cluster head (CH) upon triggered by
  certain signal events
• moderately to densely populated low-end sensors
  whose function is to provide sensor information
  to CHs upon request.
• A cluster is formed and a CH becomes active, when
  the acoustic signal strength detected by the CH
  exceeds a predetermined threshold.
• The active CH then broadcasts an information
  solicitation packet, asking sensors in its
  vicinity to join the cluster and provide their
  sensing information.
• Issues addressed
• how CHs cooperate with one another to ensure that
  only one CH (preferably the CH that is closest to
  the target) is active with high probability
• when the active CH solicits for sensor
  information, instead of having all the sensors in
  its vicinity reply, only a sufficient number of
  sensors respond with non-redundant, essential
  information to determine the target location
• both the packets that sensors send to their CHs
  and packets that CHs report to subscribers do not
  incur significant collision.

109
Power Conservation and Quality of Surveillance in
TargetTracking Sensor Networks - Jun

• Chao Gui and Prasant Mohapatra
• UC Davis
• Mobicom 2004

110
What adapts?

• Node scheduling
• Sleep and wakeup

111
What to adapt to?

• Target movement

112
How

• To guarantee surveillance quality of a moving
  target, it is necessary that each node be
  proactively informed when a target is moving
  toward it.
• Each sensor node has four working modes Waiting,
  Prepare, SubTrack, and Tracking mode.
• The Waiting mode represents the low power mode in
  surveillance stage.
• Prepare and SubTrack modes both belong to the
  preparing and anticipating mode, and a node
  should remain active in both modes.
• Node state transit between the four states and
  nodes are alerted about the potential targets

113
Backcasting Adaptive Sampling for Sensor
Networks - Jidong
114
What adapt

• Sampling of environment by sensor network

115
What to adapt to?

• adapt to application requirement and save energy
  consumption

116
How

• Problem
• To propose adaptive sampling scheme for Sensor
  Networks
• Considering meeting application requirements with
  optimized energy consumption
• Solution
• Proposed backcasting (an feedback mechanism)
• first having a small subset of the wireless
  sensors communicate their information to a fusion
  center.
• This provides an initial estimate of the
  environment being sensed, and guides the
  allocation of additional network resources.
• the fusion center backcasts information based on
  the initial estimate to the network at large,
  selectively activating additional sensor nodes in
  order to achieve a target error level

117
Cross Layer
118
AdaptNetAn Adaptive Protocol Suite for the
Next-Generation Wireless Internet - Keita
119
What adapts

• Protocol stack
• on mobile terminals connected to wireless Internet

120
Adapts to what

• Adapts to architectural heterogeneity and diverse
  QoS requirements

121
How

• Propose a set of adaptive protocols
• Application layer
• source and channel-adaptive coding to handle data
  and bit error rate fluctuations of the wireless
  channel.
• Transport layer
• an adaptive mobile-host-centric transport layer
  with an adaptive congestion control algorithm
• Link layer
• adaptive medium access control (A-MAC) framework
  to perform seamless medium access control over
  heterogeneous networks
• Data link
• an adaptive error correcting system that
  functions with only one encoder and
• decoder at the sender and receiver respectively,
  but still can change the coding rate based on the
  channel conditions to maintain acceptable quality
  of service (QoS).

122
Bio Layer
123
Using Adaptive networks to model and control drug
delivery - Yutaka
Biology related (neural network)
124
What adapts?

• Drug Delivery System

What to adapt to?

• Environment that has some physical
  characteristics (e.g., blood pressure and body
  temperature)

125
How

• Problem
• It is time-consuming process to adjust drug to
  human body, where several physical
  characteristics can be observed.
• High level idea
• Modeling Drug Delivery System as nonlinear system
  using Adaptive network
• Adjusting drugs to human body based on
  simulations on this model.