Fan Yang, Qian Zhang, Member, IEEE, Wenwu Zhu, Senior Member, IEEE, and Ya-Qin Zhang, Fellow, IEEE

Fan Yang, Qian Zhang, Member, IEEE, Wenwu Zhu, Senior Member, IEEE, and Ya-Qin Zhang, Fellow, IEEE

Chungyong Cha

Fig. 1. General illustration of video streaming over wireless Internet.

The end-to-end packet loss can be caused by◦ The congestion loss occurred in the wired network◦ The erroneous loss occurred in the wireless network

• The variation in end-to-end delay is large◦ Packet loss ratio and round-trip time(RTT)is usually used

by streaming protocol to adjust sending rate

• The streaming protocol should be friendly to TCP

WMSTFP congestion control(sender)◦ Adjust sending rate based on the feedback information

WMSTFP network monitor(receiver)◦ Analyze the erroneous loss rate(wireless) and congestive

loss rate(wired)◦ Estimate the end-to-end available network bandwidth

Network-adaptive ULP channel encoder◦ Protect different layers of PFGS according to their

importance and network status using RS codes Loss differentiated R-D based bit allocation

◦ Make the total sending rate adapt to the estimated network conditions

Accurate loss differentiation◦ Detect packet losses caused by the erros in

wireless channels using the information acquire at the link-layer

Forward loss ratio estimation◦ Packets have different loss patterns(different loss

burtiness lengths) Smoothed RTT measurement

WMSTFP consists of a sender part and a receiver part◦ Estimating loss rate(congestive and erroneous)◦ Estimating RTT and retransmission time out(RTO)◦ Estimating the available network bandwidth and adjusting the sending rate.

RTT=RTTwired+RTTwireless

RTTn=Ts′n-Tsn-offset RTTi≈ RTTn-2[(Trn-Tri)-(Tsn-Tsi)],(0≤ i≤ n-1)

is the current estimated RTT is the smoothed estimation of the

variation of RTT

is the estimated RTTin the last round

Use the link-layer information to differentiate the wireless erroneous loss and congestive loss.

In the 3G wireless communication system, we can deduce a packet loss caused by wireless errors based on the information provided in the radio link control layer(RLC)

We can even get more detailde statistical information such as frame error rate at the radio resource control layer(RRC)

If(estrate>currate)

else

Rf:

Applying ULP scheme to different layer to provide prioritized transmission

When the net work is in good status, more bit budget should be assigned for source coding and fewer bits should be assigned for channel coding.

On the contrary,when network condition is bad,it is necessary to allocate more bits for channel coding, thus fewer bits should be allocated for source coding.

Protection ratios under different burst lengths

Different loss patterns have different impact on the perceived Qos quality in video streaming.

Fig. 6. Comparisons of the 55th reconstructed frame under different bit allocation schemes.

DT=Ds(Rs)+Dc(Rs,Rc)

DT:end-to-end distortion Ds:source distortion(caused by quantization& rate control) Dc:channel distortion(caused by packet loss) Rs:source coding rate Rc:channel coding rate

Allocate the available bit rate such that the optimal Rs and Rc are obtained by minimizing end-to-end distortion under the constraint Rs+Rc≤ Rr.

Dc(Rs,Rc)=Dc,wired(Rs,Rc)+Dc,wireless(Rs,Rc)

Average throughput of WMSTFP&:TCP connection

Friendliness measure defined as

Note that the closer to 1 the value of F is,the frien-dlier WMSTFP is to TCP

Comparisons of throughput for TCP and WMSTFP connections.

Variations of WMSTFP & TCP difined as:

The smoothness measure is then defined as:

S≤ 1means the ith WMSTFP is smoother than TCP

(a) FER = 0:1.(b) FER = 0:2. (c) FER = 0:3.

Table 1OVERALL PACKET LOSS RATIO

UNDER DIFFERENT FER

FULP-T◦ Fixed ULP without loss pattern differentiation over

TFRC FULP-W

◦ Fixed ULP without loss pattern differentiation over WMSTFP

AULP-W◦ Adaptive ULP over WMSTFP

(a)FER=0.3

(b)FER=0.2

FER=0.3

FER=0.2

(a)FER=0.3

(b)FER=0.2

(a)FER=0.3

(b)FER=0.2

FULP-W

FULP-T

AULP-W

The streaming protocol, WMSTFP, is proposed which is friendly to TCP in wired-line IP networks, and can achieve higher throughput than TCP-friendly in wirelessnetworks.

A loss differentiated R-D-based bit allocation sch-eme is further proposed by applying the network-adaptive ULP scheme.