A Survey on Wireless Multimedia Sensor Networks
2008/04/18 14:40
I. F. Akyildiz, T. Melodia, and K. R. Chowdhury, "A survey on wireless multimedia sensor networks," Computer Networks, vol. 51, pp. 921-960, 2007.
-
Introduction
WMSN :
networks of wirelessly interconnected devices that allow retrieving video and audio streams, still images, and scalar sensor data
Possible new applications (existing apps - tracking, home automation, environment monitoring) :
- Multimedia surveillance sensor networks
- Storage of potentially relevant activities
- Traffic avoidance, enforcement and control systems
- Advanced health care delivery
- Automated assistance for the elderly and family monitors
- Environmental monitoring
- Person locator services
- Industrial process control
Advantages :
-
Enlarging the view.
- distributed system of multiple cameras and sensors -> multiple view points
-
Enhancing the view
- redundancy by multiple -> provide enhanced understanding
- same area, different views
-
Enabling multi-resolution views
- can be used recognize people based on facial characteristics
Needs :
-
provide QoS
- network layer metrics : latency, jitter
- minimize the energy consumption
Challenging for QoS :
- Resource constraints : battery, memory, processing, data rate
-
Variable channel capacity
- depending on interference level
- power control, routing, rate policies
-
Cross-layer coupling of functionalities
- due to the shared nature of the wireless communication channel
- should not be treated separately
-
Multimedia in-network processing
- the processing and delivery of multimedia content are not independent and their interaction has a major impact on the levels of QoS that can be delivered
- cross-layer optimization of communication process
-
in-network processing of raw data streams
-
Design Factors
WMSN :
convergence of communication and computation
(signal processing, control theory, embedded computing)
Factors :
-
Application-specific QoS requirements
- Snapshot or streaming multimedia
-
High bandwidth demand
- Crossbow's MICAz or TelosB : 250kbit/s
-
Multimedia source coding techniques
- QCIF(176x120) : 21Kbytes -> 5Mbit/s at 30 fps
- considering processing and energy constraints
-
Multimedia in-network processing
-
require new architectures to extract semantically relevant information
- increase the system scalability by reducing tx of redundant information
- merging data originated from multiple views, multiple resolutions
-
distributed filtering techniques can create a time-elapsed image
- H. Stockdon, R. Holman, "Estimation of wave phase speed and nearshore bathymetry from video imagery," J. Geophys. Res. 105 (C9) (2000) 22, 015-22, 033.
-
-
Power consumption
-
protocols, algorithms and architectures to maximize the network lifetime
- providing QoS required by application
-
- Flexible architecture to support heterogeneous applications
-
Multimedia coverage
- larger sensing range and directivity
-
Integration with Internet (IP) architecture
- for the commercial development (anywhere, anytime)
-
approaches
- use of application level gateways
- overlay IP networks
-
Integration with other wireless technologies
- create large-scale sensor network
-
need sacrificing on the efficiency of the operation within individual technology
-
-
Network Architecture
Focus :
- Scalability
Reference Architecture :
-
single-tier network of homogeneous video sensors
- processing hubs : higher processing capabilities
- wireless gateway : interconnected to a storage hub and to a sink
- storage hub : storing multimedia content locally for subsequent retrieval
- sink : front-end for network querying and tasking
-
single-tiered clustered architecture of heterogeneous sensors
-
cluster head (processing hub)
- perform intensive multimedia processing
- relay the gathered content to wireless gateway and storage hub
-
-
multi-tiered heterogeneous network
- resource-constrained, low-power scalar sensors : simpler tasks
- resource-rich, high-power device : complex tasks
Coverage :
-
sensing range -> camera's field of view (FoV)
S. Soro, W.B. Heinzelman, "On teh coverage problem in video-based wireless sensor networks,"
in Proc. of the IEEE intel. conf. on Broadband Communications, Networks nd Systems (BroadNets), Boston, MA, USA, October 2005
-
Collaborative In-network Processing
Objective :
-
to avoid transmitting large amounts of raw streams
Needs :
- application-specific querying and processing
-
architectures allowing data fusion or other complex processing operations in-network
Data Alignment and Image Registration :
-
Different Viewpoints (Multi-view Analysis)
- larger 2D view or a 3D representation : in remote sensing, computer vision and 3D shape recovery
-
Different times (multi-temporal analysis)
- find and evaluate changes in time in the scene
-
Different sensors (multi-modal analysis)
- acquired by different sensors
Registration methods :
-
feature detection
- detect distinctive objects (closed-boundary regions, edges, contours, line intersections, corners...)
-
feature matching
- establish correspondence between detected features and those detected in the reference images
-
transform model estimation
- estimate the types and parameters of mapping functions
-
image re-sampling and transformation
-
transform sensed image using the mapping functions
-
-
-
Application Layer
Service :
- traffic management and admission control functionalities
- source coding
- flexible and efficient system software
- primitives for applications to leverage collaborative, advanced in-network multimedia processing techniques
Traffic classes :
Multimedia encoding techniques :
-
Design Objectives
- High compression efficiency
- Low complexity
- Error resiliency
-
Distributed Source Coding
-
B. Girod, A. Aaron, S. Rane, D. Rebollo-Monedero, "Distributed video coding," Proc. IEEE 93 (1) (2005) 71–83.
Z. Xiong, A.D. Liveris, S. Cheng, "Distributed source coding for sensor networks," IEEE Signal Process. Mag. 21 (September) (2004) 80–94.
-
- JPEG with differential encoding
- distributed coding of images taken by cameras having overlapping FoV
-
multi-layer coding with wavelet compression
S. Misra, M. Reisslein, G. Xue, A survey of multimedia streaming in wireless sensor networks, submitted for publication.
-
Wyner-Ziv video coder
- Pixel-domain Wyner–Ziv encoder
- Transform-domain Wyner-Ziv encoder
System Software and middleware :
-
Main desired characteristics of a system software
- provide a high-level interface
- application-specific algorithm for in-network processing on multimedia content
- long-lived
- shared among multiple heterogeneous applications
- shared among heterogeneous sensors and platforms
- scalable
-
trade-off between degrees of flexibility and network performance
- flexibility, inter-operability, reprogrammability
Open research issues :
- trade-offs between fidelity and energy consumption
- high layer abstraction for fast development (controlling efficiency of low-level operations)
-
high-level primitives for applications to leverage collaborative, in-network multimedia processing
-
Transport Layer
Consideration :
-
Congestion control need to :
- be tuned for immediate response
- avoid oscillations of data rate along the affected path
-
Packet re-ordering due to multi-path
- large burst data, but limited buffers at the intermediate sensor nodes
-
channel conditions not permitting high data rate for the entire duration
Open research issues :
- Trade-offs between reliability and congestion control
- Real-time communication support
-
Relation between multimedia coding rate and reliability
-
-
Network Layer
Tasks :
- providing variable QoS guarantees
- supporting multimedia applications with lack of global knowledge, reduced energy end computational ability of node
Open research issues :
- Identification of the optimal routing metrics
-
End-to-end QoS guarantees
-
Cross-layer Design
Why not layered Design?
- Redundancy for service interface
-
Functions handled at different layers are highly coupled
- scheduling with rate allocation (MAC & Routing)
- congestion and power control (capacity based on transmission power)
Advantages of layered design : modularity
- Simplicity
- Easy debugging
- Easy to standardize
- Flexibility to deploy new protocols (easy upgradeable)
Cross-layer design advantages :
- Exploits the interactions between layers
- Promotes adaptability at all layers based on information exchange between layers
- In wireless networks: tight interdependence between layers
Cross-layer design disadvantages :
- Hard to characterize the interactions between protocols at different layers
- Joint optimization across layers may lead to complex algorithms
- Potential to destroy modularity
-
Other Research Issues
Convergence of Sensing and Actuation
- Wireless Sensor and Actor Networks : WSANs
Network Synchronization
- to aggregate according to packet generation times
Inter-media synchronization
- sound vs. image from RoI
Localization
- to allocate resources to events, decide sensing precision and ensure complete monitoring
'연구도 하고 > 레퍼런스' 카테고리의 다른 글
| A Survey on Wireless Multimedia Sensor Networks (0) | 2008/04/18 |
|---|---|
| WSN Research Lab Lists (1) | 2008/04/03 |
| Readings for WSN research (0) | 2008/04/03 |
| Tython 과 TinyViz의 Power Profiling 연동 (4) | 2006/09/12 |
| TinyOS 재전송 기능 사용 / QueuedSendM 버그 (1) | 2006/06/24 |
| 유전 알고리즘 (Genetic Algorithm) (0) | 2006/05/22 |
PowerProfilePlugin.java
