Presentation is loading. Please wait.

Presentation is loading. Please wait.

1 An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks The First ACM Conference on Embedded Networked Sensor Systems (SenSys 2003) November.

Similar presentations


Presentation on theme: "1 An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks The First ACM Conference on Embedded Networked Sensor Systems (SenSys 2003) November."— Presentation transcript:

1 1 An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks The First ACM Conference on Embedded Networked Sensor Systems (SenSys 2003) November 2003 T. van Dam and K. Langendoen Delft University of Technology The Netherlands

2 2 / 23 Outline Introduction Introduction Related Work Related Work T-MAC Protocol Design T-MAC Protocol Design Simulation Simulation Real Implementation Real Implementation Conclusions and Future Work Conclusions and Future Work

3 3 / 23 Introduction Communication patterns Communication patterns Local uni-/broadcast Local uni-/broadcast Real-world event in the network occurs, nodes to perform some in-network processing Real-world event in the network occurs, nodes to perform some in-network processing This will generally involve local messages being exchanged between neighbors This will generally involve local messages being exchanged between neighbors Nodes to sink reporting Nodes to sink reporting Nodes may want to report something Nodes may want to report something After processing a local event After processing a local event Periodically Periodically

4 4 / 23 Introduction Idle listening problem Idle listening problem Most energy in traditional MAC protocols is wasted by idle listening Most energy in traditional MAC protocols is wasted by idle listening 5mS Tx 5mS Rx 990mS NOTHING HAPPENS Average rate of one per second Do nothing for 99% of the time

5 5 / 23 Introduction EYES nodes EYES nodes Texas Instruments MSP430F149 processor, 2KB RAM and 60KB Flash memory Texas Instruments MSP430F149 processor, 2KB RAM and 60KB Flash memory Variable clock rate, with a maximum of 5MHz Variable clock rate, with a maximum of 5MHz 115kbps radio (RFM TR1001, 868.35MHz) 115kbps radio (RFM TR1001, 868.35MHz) JTAG, RS232, 2 LEDs, JTAG, RS232, 2 LEDs, 16 general purpose I/O pins (8 with ADC capability) 16 general purpose I/O pins (8 with ADC capability) 3V supplied by two AA batteries 3V supplied by two AA batteries

6 6 / 23 Related Work TDMA TDMA-based protocols are naturally energy preserving TDMA-based protocols are naturally energy preserving Duty cycle built-in Duty cycle built-in Do not suffer from collisions Do not suffer from collisions Maintaining a TDMA schedule in an ad-hoc network is not an easy task Maintaining a TDMA schedule in an ad-hoc network is not an easy task TDMA divides time into very small slots TDMA divides time into very small slots The effect of clock drift can be disastrous The effect of clock drift can be disastrous Keeping a list of neighbor ’ s schedules takes valuable memory capacity Keeping a list of neighbor ’ s schedules takes valuable memory capacity Allocating TDMA slots is a complex problem that requires coordination Allocating TDMA slots is a complex problem that requires coordination

7 7 / 23 Related Work Another way Extra radio the so-called wake-up radio Extra radio the so-called wake-up radio Operates on a different frequency Operates on a different frequency It needs no data processing and therefore uses much less energy It needs no data processing and therefore uses much less energy CSMA protocol (802.11) CSMA protocol (802.11) Only uses a single frequency requires some kind of in-band signalling Only uses a single frequency requires some kind of in-band signalling 802.11 presumption that all nodes are located in a single network cell 802.11 presumption that all nodes are located in a single network cell

8 8 / 23 Related Work TinyOS project Includes a sensor-networks specific optimization of the basic CSMA protocol Includes a sensor-networks specific optimization of the basic CSMA protocol By sending out a very long preamble, By sending out a very long preamble, Receivers only need to weak up periodically to sense activity Receivers only need to weak up periodically to sense activity Shifts the cost from the receiver to the transmitter Shifts the cost from the receiver to the transmitter

9 9 / 23 Related Work S-MAC Single-frequency contention-based protocol Single-frequency contention-based protocol Time is divided into fairly large frames Time is divided into fairly large frames Every frame has two parts Every frame has two parts Active part Active part Communicate with its neighbors Communicate with its neighbors Send any messages queued during the sleeping part Send any messages queued during the sleeping part Sleeping part Sleeping part A node turns off its radio to preserve energy A node turns off its radio to preserve energy

10 10 / 23 Related Work S-MAC Needs some synchronization Needs some synchronization Is not as critical as in TDMA-based protocols Is not as critical as in TDMA-based protocols Time scale is much large Time scale is much large Trades used energy for throughput and latency Trades used energy for throughput and latency

11 11 / 23 T-MAC Protocol Design Every node periodically wakes up to communicate with its neighbors Every node periodically wakes up to communicate with its neighbors RTS, CTS, Data and ACK scheme RTS, CTS, Data and ACK scheme Collision avoidance Collision avoidance Reliable transmission Reliable transmission TA determines the minimal amount of idle listening per frame TA determines the minimal amount of idle listening per frame

12 12 / 23 T-MAC Protocol Design Basic protocol Basic protocol Messages between active times must be buffered Messages between active times must be buffered The buffer capacity determines an upper bound on the maximum frame time The buffer capacity determines an upper bound on the maximum frame time Clustering and synchronization Clustering and synchronization When a node hears nothing, it chooses a frame schedule and transmits a SYNC packet When a node hears nothing, it chooses a frame schedule and transmits a SYNC packet Nodes retransmit their SYNC once in a while Nodes retransmit their SYNC once in a while Nodes can detect the existence of different schedules Nodes can detect the existence of different schedules This allows new and mobile nodes to adapt to an existing group This allows new and mobile nodes to adapt to an existing group

13 13 / 23 T-MAC Protocol Design RTS operation and choosing TA Fixed contention interval Fixed contention interval Like 802.11, nodes wait for a random time within a contention interval Like 802.11, nodes wait for a random time within a contention interval This interval is tuned for maximum load This interval is tuned for maximum load RTS retries RTS retries When the sending node receives no answer within the interval TA, it might go to sleep When the sending node receives no answer within the interval TA, it might go to sleep A node should retry by re-sending the RTS two retries A node should retry by re-sending the RTS two retries

14 14 / 23 T-MAC Protocol Design RTS operation and choosing TA Determining TA Determining TA A node should not go to sleep while its neighbors are still communication A node should not go to sleep while its neighbors are still communication TA > Contention interval + RTS + Turn-around time TA > Contention interval + RTS + Turn-around time

15 15 / 23 T-MAC Protocol Design Asymmetric communication Asymmetric communication – Early sleeping problem Asymmetric communication – Early sleeping problem Future request-to-send (FRTS) Future request-to-send (FRTS) Taking priority on full buffers Taking priority on full buffers Early sleeping problem

16 16 / 23 T-MAC Protocol Design Asymmetric communication Future RTS

17 17 / 23 T-MAC Protocol Design Asymmetric communication Taking priority upon receiving RTS

18 18 / 23 Simulation OMNeT++ discrete event simulation package OMNeT++ discrete event simulation package 100 nodes in a 10 by 10 grid 100 nodes in a 10 by 10 grid

19 19 / 23 Simulation Load (byte/node/s) 0 20 40 0 20 40 60 80 100 CSMA S-MAC T-MAC Message Length = 20 Message Length = 100 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 energy used [avg. mA/node] Early Sleeping problem

20 20 / 23 Real Implementation 20 18 16 14 12 10 8 6 4 2 0 20 18 16 14 12 10 8 6 4 2 0 0 2 4 6 8 10 time [s] current [mA] Tx Rx CTS

21 21 / 23 Real Implementation Average Electrical Current

22 22 / 23 Conclusions and Future Work Conclusions Conclusions To solve the problem of idle listening To solve the problem of idle listening Dynamically adapts a listen/sleep duty cycle Dynamically adapts a listen/sleep duty cycle Saving as much as 96% of the energy compared to a traditional protocol Saving as much as 96% of the energy compared to a traditional protocol Future Work Future Work Applying virtual clustering to groups of mobile nodes Applying virtual clustering to groups of mobile nodes Multi-hop synchronization Multi-hop synchronization

23 23 / 23 Thanks

24 24 / 23

25 25 / 23 T-MAC Protocol Design Activation event Activation event The firing of a periodic frame timer The firing of a periodic frame timer The reception of any data on the radio The reception of any data on the radio The sensing RSSI (Received Signal Strength Indication) The sensing RSSI (Received Signal Strength Indication) The end-of-transmission of a node ’ s own data packet or ACK The end-of-transmission of a node ’ s own data packet or ACK Overhearing prior RTS and CTS Overhearing prior RTS and CTS


Download ppt "1 An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks The First ACM Conference on Embedded Networked Sensor Systems (SenSys 2003) November."

Similar presentations


Ads by Google