Transmission rate control method and communication device

1. A transmission rate control method comprising the steps of:

in a streaming delivery on a network which includes a wireless environment, transmitting data to a reception device at a most recent transmission rate R SND t−1 and receiving an RTCP packet including a reception rate R RCV t of data from the reception device;

when the network is in a congestion state, or when the network is in a non-congestion state and a most recent reception achievement rate R AR t−1 is higher than the reception rate R RCV t , calculating a reception achievement rate R AR t as a long-term EWMA (exponential weighted moving average) of the reception rate, and

when the network is in the non-congestion state and the most recent reception achievement rate R AR t−1 is lower than the reception rate R RCV t , or when a time-out of the RTCP packet occurs, calculating the reception achievement rate R AR t as a short-term EWMA of the reception rate; and

setting a transmission rate R SND t based on the reception achievement rate R AR t .

2. The transmission rate control method according to claim 1 , wherein in the step of calculating the reception achievement rate R AR t , the reception achievement rate R AR t is calculated as the long-term EWMA or the short-term EWMA of the reception rate represented by the following formula:

when the reception achievement rate R AR t is calculated as the long-term EWMA of the reception rate,

R AR t ←α L ·R AR t−1 +(1−α L )· R RCV t , and

when the reception achievement rate R AR t is calculated as the short-term EWMA of the reception rate,

R AR t ←α S ·R AR t−1 +(1−α S )· R RCV t ,

wherein 0<α S <α L <1.

3. The transmission rate control method according to claim 1 , wherein in the step of setting the transmission rate R SND t , the transmission rate R SND t is set, when the network is in the non-congestion state, by the following formulae:

R SND t ←( R RCV t ·RTT t +1)/( RTT t +ΔRTT)

R SND t ←β·R SND t +(1−β)· R AR t , and

further, when R SND t >a limit rate R LR t , R SND t ←R LR t ; or

the transmission rate R SND t is set, when the network is in the congestion state, by the following formulae:

R SND t ←γ·R AR t ,

further, when R SND t >R RCV t , R SND t 77 R RCV t ,

R SND t′ ←( R RCV t ·RTT t +1)/( RTT t +ΔRTT), and

further, when R SND t′ <R SND t , R SND t ←R SND t′ +R SND t )/2,

wherein RTT t and RTT t−1 are the latest RTT and the most recent RTT, ΔRTT is a variation of the RTT (RTT t −RTT t−1 ), R LR is the limit rate defining an upper limit of R SND t , R SND t′ is a temporary transmission rate, 0<β<1, and 0<γ<1.

4. The transmission rate control method according to claim 3 , wherein the limit rate R LR t is calculated by the following formula:

when R SND t−1 <R MAX ,

R LR t ←R SND t−1 +α LR ·( R MAX −R SND t−1 )+ ΔR MIN , and

when R SND t−1 ≧R MAX ,

R LR t ←R SND t−1 +ΔR MIN ,

wherein R SND t−1 is the most recent transmission rate, R MAX is the maximum transmission rate in the current communication connection, ΔR MIN is the minimum increase limit rate (a constant, ΔR MIN ≧0), and 0<α LR <1.

5. The transmission rate control method according to claim 1 , wherein when RTT t is changed from a state of not more than RTT START to a state of more than RTT START , or when a packet loss occurs, a transmission rate control unit determines that the network is transferred to the congestion state, and

when RTT t is changed from a state of not less than RTT END to a state of less than RTT END , and when SRTT t /LRTT t <α, the transmission rate control unit determines that the network is transferred to the non-congestion state, and

further comprising a step of, even when the transmission rate control unit determines that the network is transferred to the non-congestion state, the congestion state of the network is continued for a congestion state duration Δ τ , represented by the following formula, from the point of time when the network transfers to the congestion state:

Δ τ ←μ·( RTT MAX −RTT MIN )/(2·(1−γ)),

wherein RTT t is the latest RTT, RTT START and RTT END are RTT upon the transfer of the network to the congestion state defined by a Spike method and RTT upon the transfer of the network to the non-congestion state defined by the Spike method, SRTT t and LRTT t are the short-term EWMA and the long-term EWMA of the RTT, α>1, 0<γ<1, and μ<1.

6. A communication device comprising:

a communication processing unit which, in a streaming delivery on a network which includes a wireless environment, transmits data to a reception device at a most recent transmission rate R SND t−1 and receives an RTCP packet including a reception rate R RCV t of data from the reception device;

a reception achievement rate calculation unit which, when the network is in a congestion state, or when the network is in a non-congestion state and a most recent reception achievement rate R AR t−1 is higher than the reception rate R RCV t , calculates a reception achievement rate R AR t as a long-term EWMA (exponential weighted moving average) of the reception rate, and when the network is in the non-congestion state and the most recent reception achievement rate R AR t−1 is lower than the reception rate R RCV t , or when a time-out of the RTCP packet occurs, calculates the reception achievement rate R AR t as a short-term EWMA of the reception rate; and

a transmission rate setting unit which sets a transmission rate R SND t based on the reception achievement rate R AR t .