Download presentation
Presentation is loading. Please wait.
Published byAdele Stevens Modified over 9 years ago
1
Optimizing Sensor Networks in the Energy-Latency-Density Design Space Curt Schurgers, Vlasios Tsiatsis, Saurabh Ganeriwal, Mani Srivastava, IEEE TRANSACTIONS ON MOBILE COMPUTING,2002 Speaker : hsiwei
2
Outline INTRODUCTION STEM SIMULATION CONCLUSIONS
3
INTRODUCTION Sensor networks are made up of a large number of tiny devices,call sensor node Sensor nodes have only a small battery as a power source To achieve satisfactory network lifetime,energy efficiency is really a problem Network communications as the radio is the main energy comsumer in a sensor node
4
INTRODUCTION(con.) Monitoring state : The network is only sensing its environment. Transfer state : Once an event happens, data needs to be forwarded to the data sink. 目的 : It reduces the energy consumption in the monitoring state to a bare minimum while ensuring satisfactory latency for transitioning to the transfer state.
5
INTRODUCTION(con.) Sleep state:In the monitoring state, where there is no traffic to forward. Working state: The radio is only turned on if the processor decides that the information needs to be communicated to other nodes. To forward traffic, nodes on the multihop path need to be awakened, or,equivalently, transition from the monitoring to the transfer state.
6
INTRODUCTION(con.) Initiator node:The node that wants to communicate. Target node: initiator node polls the node it is trying to wake up. Once the link between nodes is activated data is transferred using a MAC protocol The reason is that MAC protocols are designed to organize access to the shared medium. STEM offers an alternative by trading energy for latency.
7
STEM Each node periodically turns on its radio for a short time to listen if someone wants to communicate with it. Since this aggressive nature is needed to limit the wakeup latency
8
STEM The solution : is to completely separate data transfer from wakeup. transfer state wakeup band
9
STEM STEM-B : As soon as the target receives a beacon, it turns on its data radio in band f1 and also sends back an acknowledgment in band f2. STEM-T : a target node never sends back an acknowledgment. contains the MAC address of both the target and initiator node.
10
STEM Target node has received the beacon correctly or seen a collided packet and turned on its data radio in either case. If they do not receive any traffic in band f1 after some time, they time out and return to the monitoring state.
11
SETUP LATENCY Fig. 11. Analysis of the setup latency of STEM-B without collisions.
12
SETUP LATENCY K=1….K
13
SETUP LATENCY NO collision
14
SETUP LATENCY Beacon collision
15
SETUP LATENCY Fig. 12. Analysis of the worst-case setup latency of STEM-T. Tone -based
16
SIMULATION Were written on the Parsec platform, an event- driven parallel simulation language distribute N nodes in a uniformly random fashion over a field of size L x L transmission range : R the average number of neighbors of a node: The node turns its data radio back off if it has not received any traffic for 20 seconds.
17
SIMULATION Setup latency of a link
18
ENERGY CONSUMPTION power of wakeup radio and data radio sleep power for the data and wakeup radio The average time the radio is on during one such data communication phase The number of such transitions per second the node sets up a link
19
ENERGY CONSUMPTION there is only one radio that is never in the sleep state,
20
ENERGY CONSUMPTION 其中 Setup frequency
21
SIMULATION
23
existing topology management schemes, such as GAF and SPAN.
24
CONCLUSIONS STEM: A topology management technique that trades power savings for path setup latency in sensor networks. the combination of STEM and GAF can reduce the energy to 1 percent or less of that of a network without topology management
25
Summary 利用兩個不的 channel, 達到 sensor node 節省 電力. 透過數學的分析及模擬, 得到不錯的省電效果
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.