+ All Categories
Home > Documents > DOT3 Radio Stack Sukun KimJaein Jeong A DOT3 Mote Design & Implementation Motivation Evaluation MICA...

DOT3 Radio Stack Sukun KimJaein Jeong A DOT3 Mote Design & Implementation Motivation Evaluation MICA...

Date post: 20-Jan-2016
Category:
View: 212 times
Download: 0 times
Share this document with a friend
1
DOT3 Radio Stack Sukun Kim Jaein Jeong A DOT3 Mote Design & Implementation Motivation Evaluation MICA is not enough for large scale applications. • DOT3 is a new platform with CC1000 radio chip. • We aim to have a working network stack for DOT3 in nesC harnessing its improved performance of radio. Network Stack implementation on a new platform of wireless sensors Jaein Jeong and Sukun Kim University of California at Berkeley Electrical Engineering and Computer Science A MICA mote, the current generation of wireless motes in Berkeley A DOT3 mote with its radio chip (CC1000) in the middle •The sender sends a number of packets and the receiver counts how many packets it received from the sender as we vary the distance between the nodes 100 through 1200ft. Ratio of successfully received packets is an indicator of effectiveness of each transmission method • Transmission with error correction code, no packets were dropped within 800ft compared to 500ft for non- ECC version. • Retransmission reduced the packet losses with additional costs. • Using multiple channels reduced the packet losses due to collision. • Retransmission reduced most of the packet losses due to collision with a little high costs (over 6 times in case of 4 senders). Discussion and Future Works Comparison with MICA Pros: Better coverage and reliability Cons: Slower transmission (60 sec vs. 9 sec for 512 packets) caused by - Slower clock rate of radio (19Kbps vs. 40Kbps) - Less efficient interrupt handler Modifying interrupt handler (from SPI to timer interrupt) will address this. Problems with our reliable transmission method Effective for moderate collision, but not for high collision. Introducing exponential back-off is expected to be helpful. Using multiple channels Reduces collision. Currently statically determined, vulnerable to misconfiguration. Dynamic frequency allocation is needed. IDLE Send a packet READING Receive a byte FIND_SYNC Detected Preamble Detected Start Symbol Not detected Start Symbol Init A packet is received State diagram for packet decomposition and reassembly Schematic of reliable transmission Ready Wait Send Done Time Out Begin Success (Ack received) Fail (Repeated timeout) Sender Receiver Data Ack Src #1 Acknum #1 Src #2 Acknum #2 Src #n Acknum #n Ack table Send Data flow in DOT3 network stack The effectiveness ofEC C (256 packets) 0% 20% 40% 60% 80% 100% D istance (ft) R atio ofreceived packets BestEffort BestEffortw /o EC C The effectiveness ofretransm ission (256 packets) 0% 20% 40% 60% 80% 100% 100 200 300 400 500 600 700 800 900 1000 1100 120 0 D istance (ft) R atio ofreceived packets BestE ffort Retransm it 2 Retransm it 3 Retransm it 5 The effects ofm ultiple channels on collision (128 pakcets pernode) 0% 20% 40% 60% 80% 100% 1 2 4 (case 1) 4 (case 2) N um berofchannels used R atio ofreceived packets B estEffort Retransmission The effects ofm ultiple channels on transm ission tim e (128 packets pernode) 0 50 100 150 200 250 300 1 2 4 (case 1) 4 (case 2) Num berofchannels used Transm ission tim e (s) B est Effort Retransmission The effects ofretransm ission on collisions (128 packets pernode) 0% 20% 40% 60% 80% 100% 1 2 3 4 N um berofsenders Ratio ofsuccessfully received packets B estE ffort Retransmission Effects ofretransm ission on transm ission tim e (128 packets pernode) 0 20 40 60 80 100 120 140 1 2 3 4 N um berofsenders Tim e to com pletion (s) BestEffort Retransmission O utdoorrange ofM IC A (256 packets) 0.0% 20.0% 40.0% 60.0% 80.0% 100.0% 0 50 100 150 200 250 300 350 400 D istance (ft) R atio ofreceived packets O ne side ofC ory H all(240 ft) MICA has noticeable packet drops within a single building (Cory Hall)! Multiplexing messages for different applications and media (radio, UART) Calculates CRC. Encode/decode data for ECC Retransmit dropped packets using Acknowledgement Bit s Bytes Packets Sends and receives data in bytes and notifies data arrival Setting the parameters for CC1000 radio chip ReliableComm GenericComm AMStandard RadioCRCPacket RFComm Chipcon SpiByteFifoC SecDedEncoding ChannelMonC Application Radio Packet decomposition and reassembly *:newly made or modified from existing TinyOS network stack * * * * * How to decompose and reassemble a packet to and from raw bytes? Sending: sends a byte when the byte buffer is empty Receiving: detects the start of a packet using preamble and start symbol triggers an event when all the bytes are ready. How to transmit messages reliably? •Add source address and Ack number to packets.
Transcript
Page 1: DOT3 Radio Stack Sukun KimJaein Jeong A DOT3 Mote Design & Implementation Motivation Evaluation MICA is not enough for large scale applications. DOT3 is.

DOT3 Radio Stack

Sukun KimJaein Jeong A DOT3 Mote

Design & Implementation

Motivation Evaluation

• MICA is not enough for large scale applications.

• DOT3 is a new platform with CC1000 radio chip.

• We aim to have a working network stack for DOT3 in nesC harnessing its improved performance of radio.

Network Stack implementation on a new platform of wireless sensors

Jaein Jeong and Sukun KimUniversity of California at Berkeley

Electrical Engineering and Computer Science

A MICA mote, the current generation of

wireless motes in Berkeley

A DOT3 mote with its radio chip (CC1000)

in the middle

•The sender sends a number of packets and the receiver counts how many packets it received from the sender as we vary the distance between the nodes 100 through 1200ft.

• Ratio of successfully received packets is an indicator of effectiveness of each transmission method

• Transmission with error correction code, no packets were dropped within 800ft compared to 500ft for non-ECC version.

• Retransmission reduced the packet losses with additional costs.

• Using multiple channels reduced the packet losses due to collision.

• Retransmission reduced most of the packet losses due to collision with a little high costs (over 6 times in case of 4 senders).

Discussion and Future Works Comparison with MICA

• Pros: Better coverage and reliability• Cons: Slower transmission (60 sec vs. 9 sec for 512 packets) caused by

- Slower clock rate of radio (19Kbps vs. 40Kbps)- Less efficient interrupt handler

• Modifying interrupt handler (from SPI to timer interrupt) will address this.

Problems with our reliable transmission method

• Effective for moderate collision, but not for high collision.• Introducing exponential back-off is expected to be helpful.

Using multiple channels

• Reduces collision.• Currently statically determined, vulnerable to misconfiguration.• Dynamic frequency allocation is needed.

IDLESend a packet

READINGReceive a byte FIND_SYNC

Detected Preamble

DetectedStart Symbol

Not detectedStart Symbol

Init

A packetis received

State diagram for packet decomposition and reassembly

Schematic of reliable transmission

Ready

WaitSend Done

Time Out

BeginSuccess (Ack received)Fail (Repeated timeout)

Sender Receiver

Data

Ack

Src #1 Acknum #1

Src #2 Acknum #2

Src #n Acknum #n

Ack table

Send

Data flow in DOT3 network stack

The effectiveness of ECC (256 packets)

0%

20%

40%

60%

80%

100%

100

200

300

400

500

600

700

800

900

1000

1100

1200

Distance (ft)

Rat

io o

f re

ceiv

ed p

acke

ts

Best EffortBest Effort w/o ECC

The effectiveness of retransmission (256 packets)

0%

20%

40%

60%

80%

100%

100

200

300

400

500

600

700

800

900

1000

1100

1200

Distance (ft)

Rat

io o

f re

ceiv

ed p

acke

ts

Best EffortRetransmit 2Retransmit 3Retransmit 5

The effects of multiple channels on collision(128 pakcets per node)

0%

20%

40%

60%

80%

100%

1 2 4 (case 1) 4 (case 2)Number of channels used

Ra

tio o

f re

ceiv

ed

pa

cke

ts Best Effort

Retransmission

The effects of multiple channels ontransmission time (128 packets per node)

0

50

100

150

200

250

300

1 2 4 (case 1) 4 (case 2)

Number of channels used

Tra

nsm

issi

on

tim

e (

s)

Best Effort

Retransmission

The effects of retransmission on collisions(128 packets per node)

0%

20%

40%

60%

80%

100%

1 2 3 4Number of senders

Rat

io o

f suc

cess

fully

rece

ived

pac

kets

Best Effort

Retransmission

Effects of retransmission on transmission time (128 packets per node)

0

20

40

60

80

100

120

140

1 2 3 4Number of senders

Tim

e to

com

plet

ion

(s) Best Effort

Retransmission

Outdoor range of MICA (256 packets)

0.0%

20.0%

40.0%

60.0%

80.0%

100.0%

0 50 100 150 200 250 300 350 400

Distance (ft)

Rat

io o

f re

ceiv

ed p

acke

ts

One side of Cory Hall (240 ft)

MICA has noticeable packet drops withina single building(Cory Hall)!

Multiplexing messages for different applications and media (radio, UART)

Calculates CRC.

Encode/decode data for ECC

Retransmit dropped packetsusing Acknowledgement

Bits

Bytes

Packets

Sends and receives data in bytesand notifies data arrival

Setting the parameters forCC1000 radio chip

ReliableCommGenericCommAMStandard

RadioCRCPacketRFComm

ChipconSpiByteFifoC

SecDedEncoding

ChannelMonC

Application

Radio

Packet decomposition and reassembly

*:newly made or modified from existing TinyOS network stack

**

*

*

*

How to decompose and reassemble a packet to and from raw bytes?

• Sending: sends a byte when the byte buffer is empty• Receiving: detects the start of a packet using preamble and start symbol triggers an event when all the bytes are ready.

How to transmit messages reliably? •Add source address and Ack number to packets.

Recommended