Outline Wireless Sensor Networks --- Concepts and Challenges Basic Concepts Applications Characteristics and Challenges 2 Traditional Sensing Method Basic Concepts Signal analysis Wired/Wireless Object Sensors Sensors are close to objects. Sensors only generate data streams. Sensors do not have computation ability. Sensors do not communicate. 4 1
Internet or Satellite Current Sensing Method Sensor network Sink Object Sensor node Sensing + Computation + Communication Small weight + Small size Low energy consumption User User Sensing Area A sensor network covers a sensing area. Each sensor is responsible for the object nearby. Sensors cooperate to complete the sensing task. Multi-hop routing is employed to report the sensed data to users. 5 6 Components of Sensor node Current Sensor Networks Sensing Unit Sensor ADC Processing Unit Communication Unit Sensing technology + Embedded computation technology + Distributed information management technology + Wireless communication technology Densely deployed sensor nodes Power Unit Storage Nested Software Static or mobile sensor nodes Ad Hoc network 7 Multi-hop routing 8 2
Sensor Network Definition A sensor network is an ad hoc wireless network which consists of a huge amount of static or mobile sensors. The sensors collaborate to sense, collect, and process the raw information of the phenomenon in the sensing area (in-network), and transmit the processed information to the observers. Applications 9 Applications of Sensor Networks Example 1 Military applications Environmental Monitoring friendly applications forces, equipment and ammunition Health Forest applications fire detection Battlefield surveillance Other Flood Telemonitoring commercial detection of applications human physiological data Reconnaissance of opposing forces and terrain Precision Tracking Environmental and agriculture monitoring control patients office buildings and doctors Battle damage assessment Drug Interactive administration museums Nuclear, Inventory biological control in and warehouses chemical attack detection and Vehicle reconnaissance tracking and detection Detecting and monitoring car thefts Military applications 11 12 3
Environmental Monitoring Example 2 Biological Systems Example 3 13 14 Example 4 Example 5 Traffic Control 15 16 4
Example 6 Contaminant Flow Monitoring Example 7 17 18 A Beautiful Hope Problem How to obtain information from locations in 3D space that are difficult to reach for human beings? Possible Solutions Wireless networks containing infinite nodes Connectivity Adaptive topology Each node Sensing + Computation + Communication Functional independence Can be re-programmed locally or remotely to have more functions anytime Survive in any environment Small size, low cost, low energy consumption 19 20 5
Characteristics and Challenges Limited battery power Each sensor node has very limited power Prone to failure due to lack of power The limitation of power is an obstacle for applications of sensor networks Current power unit Challenge cannot satisfy 1 requirements of sensor networks How to save energy, balance energy consumptions, Communication consume maximize more network energy than lifetime? computation Develop an energy complexity theory Develop an energy balance theory Design and analyze energy efficient algorithms 22 Limited Computation Ability Sensors have embedded processors, memories and computation abilities. However, the Challenge processor capability, 2 memory size and energy are all limited, How resulting to design the efficient limitation distributed to the computation algorithms for ability. a large number of sensors that have limited computation abilities in wireless distributed environment? Limited Communication Ability Scarce wireless channel Transmission range is only from several meters to several hundred meters, and greatly affected Challenge by energy. 3 How to complete transmission, query, analyze and mine sensed data with limited communication ability? Minimize energy consumption Minimize time and message complexities 23 Develop algorithms with minimized communication complexity 24 6
Multi-source, Multi-hop Communication Method Many sensors transmit information to a single destination. The transmission of multi-source information needs routes consisting Challenge multiple 4 sensor nodes, and each information source can have multiple routes. How to select transmission routes? Develop an optimized or approximate routing theory in static or dynamic environment. 25 Mobility and Frequent on-and-off In dynamic sensor networks, nodes move frequently. Nodes maybe on and off frequently, resulting communication filatures. Factors including Challenge mountains, 5 buildings, obstacles and environmental interference may How to make re-construct some sensors routes isolated and from how network to make in a long time. routes adaptive? Develop a random network connectivity theory Design adaptive algorithms capable of re-constructing routes 26 Huge Number of Sensors and Big Deployment Region A huge number of sensors are densely deployed. The deployed region is big and under tough conditions. It is difficult Challenge to maintain 6 sensor networks. How to make sensor networks robust and fault-tolerant in both software and hardware? Consider maintenance when design algorithms. 27 Distributed Actuators Many sensor networks need to control their monitored objects. Sensors need to be equipped with actuators. Challenge 7 How to manage a huge amount of actuators? 28 7
Infinite Sensed data Each sensor timely generates infinite data streams. Each sensor only has limited memory and computation ability. Challenge 8 How to query, analyze and mine the huge amount of data and where to store data? 29 Data-centric Network Users are interested in data, not network. Attribute-based naming is used and nodes do not have global ID. A sensor network is a data space. Management and manipulation of data are purposes of network Challenge applications. 9 How to construct data-centric networks? Management and manipulation of sensed data Combine with network technologies Construct an efficient sensed data space or database Users can manage sensed data as using traditional databases. 30 Multiple types of sensors Physical sensors Biological sensors Chemical sensors Challenge 10 How to construct new sensor concepts, theories, technologies and how to coordinate different types of sensors? 31 Sensor deployment Coverage Topology control Localization Synchronization Security Signal processing Other challenges 32 8