JPH0413390A - Picture packet multiplexer - Google Patents

Abstract

PURPOSE:To prevent missing of picture information at receiver side and deterioration of a picture by comparing a moving average of input speeds of input signals to a buffer means with 1st and 2nd threshold levels respectively, giving priority to a signal of a terminal equipment and processing the signal into a packet when the moving average exceeds the 1st threshold level. CONSTITUTION:Moving average calculation sections 121-12n apply weight average calculation to a time series data of a picture coded signal by a weight function to obtain a moving average Y. Then threshold levels A1, A2 of the moving average Y are provided to set a normal area, a high load area and an overload area. A transmission assignment processing section 10 applies packet processing to a signal at a terminal equipment when the moving average Y enters the overload area with priority. When the moving average Y enters the high load area, the packet processing is stopped. When the moving average is a history for processing period, it represents prediction that the tendency at a current point of time will continue for a while. Thus, the capacity excess of an exchange network is prevented in advance.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image packet multiplexing device that is used in connection with a high-speed packet switching network and packetizes and transmits image signals from a plurality of terminals.

(Prior art) Hatahiko Yasuda, ``Video Packet Transmission Technology Jl, 1987 Image Coding Symposium IPC5J87) describes technical trends and technical issues related to image packet transmission technology. In order to efficiently transmit data information within the network, missing data due to transmission errors is retransmitted using an error control protocol.For this reason, the communication protocol is complex, and since it targets data information, transmission The upper limit of the speed is 64 Kbit/s, which is not suitable for transmitting real-time audio and image information.

On the other hand, high-speed packet switching simplifies the communication protocol within the network and increases the packet transmission speed to 1.5 Mbit/s or higher, which is the primary speed of the digital network, thereby making it possible to perform demanding packet communications in real time. do. High-speed packet-switched networks make it possible to handle various types of information such as voice, image, and data that have different transmission speeds in the same packet, and can flexibly respond to changes in various types of information over time. In addition, efficient operation of the transmission path can be expected due to the statistical multiplexing effect that utilizes the burst characteristics of different information sources.

In high-speed packet switching networks, all transmission information such as voice, images, and data is handled in fixed-length packets (for example, 32 bytes = 256 bits). An increase or decrease in the amount of different transmission information generated by each information source corresponds to an increase or decrease in the amount of high-speed packets generated.

Among various types of information, image information has a larger amount of information and is more bursty than audio information. In particular, since a television signal as moving image information is composed of still image signals (frame signals) continuous in the time axis direction and has large fluctuations, a high-efficiency code is used in the -shell to compress the image information.

The image packet transmitting device of the target high-speed packet switching network highly efficiently encodes the input image information, converts it into a coded signal, assembles it into image packets, and transmits it to the high-speed packet switching network. The image packet receiving device that receives this
Image information is reproduced by disassembling image packets and decoding encoded signals.

(Problem to be solved by the invention) In a high-speed packet switching network, transmission paths (
unlike conventional packet-switched networks, flow control and error control are not performed on links). Therefore, if the image packet transmitter in the network transmits an unlimited number of packets, the communication network will become overloaded, packet overflow will occur at the node, packets will be discarded, and the image information will be discarded on the receiving side. It is missing and the image deteriorates.

The present invention eliminates the shortcomings of the prior art, autonomously limits the transmission speed of image packet output, prevents abnormal congestion of communication traffic in a high-speed packet switching network, and moreover, allows information on multiple terminals to be An object of the present invention is to provide an image packet multiplexing device that multiplexes and transmits image packets.

(Means for Solving the Problems) An image packet multiplexing device according to the present invention is connected to a plurality of terminals in a packet switching network, is compatible with the terminals, inputs an image signal from the terminal, and outputs a coded signal. buffer means for inputting the encoded signal, temporarily storing it and outputting it; packet assembling means for converting the encoded signal from the buffer means into a packet signal and outputting it; and a plurality of terminals for inputting the packet signal. packet multiplexing means for generating a multiplexed packet containing image information and outputting it to a transmission path;
moving average value calculation means for calculating the input speed of the encoded signal input to the buffer means as a moving average value for a predetermined period; buffer usage rate calculation means for calculating the usage rate of the /to777 means; and bar777 means. buffer reading control means for stopping the output signal of the buffer means, and a moving average value of the input speed of the input signal to the buffer means calculated by the moving average value calculation means is set to a first threshold value and a second threshold value of the moving average value. Instructs the packet multiplexing unit to perform high load control in which the packet signals from the buffer means that exceed the first threshold are selected preferentially and packetized, and the packet signals that exceed the second threshold are
and transmission control means for instructing the buffer readout control means to control readout based on non-selection with respect to the packet signal from the G.777 means.

(Operation) The image packet multiplexing device according to the present invention connects to a plurality of terminals, converts image signals from the terminals into encoded signals, and stores the encoded signals in buffer means corresponding to the terminals. Then, it is extracted as a packet signal and the image information of each terminal is combined into one multiplexed packet.

The moving average value of the speed of the encoded signal input to the buffer and the buffer usage rate are
has a threshold of The moving average value and buffer usage are
When each first threshold is exceeded, the signal of that terminal is preferentially packetized. When the second threshold is exceeded, packetization of the terminal's signal is stopped.

(Embodiment) Next, an embodiment of the image packet multiplexing and transmitting apparatus according to the present invention will be described in detail with reference to the accompanying drawings.

This image packet transmitting device inputs image signals from a plurality of terminals in a high-speed packet switching network, converts them into multiplexed image packets, and outputs the multiplexed image packets. High-speed packet switching networks do not perform error control, simplify communication protocols within the network, and
By setting the packet transmission speed to 1.5 Mbit/g or higher, the transmission capacity of the switching network is limited, and in order to maintain normal service, the output of each terminal must be It is necessary to appropriately limit the transmission speed of packets transmitted in order to prevent packets from being discarded within the network.

Figure 3 shows the velocity waveform of an image encoded signal that encodes the image signal inputted to the image transmitter, which includes a peak at 8-stroke and is falling. A weighted average calculation is performed on the time series data using a moving average weight function to obtain a moving average value. The definition of the moving average is described, for example, in 1 Time Series Analysis 1 (Corona Publishing).

The observed time series X(t) is expressed as the sum of an average value function and a variation value function, as X (t) = m (t) + ε(1). Values at different times before and after a certain time t are respectively weighted, and the weighted average value is used as the value at time t. Time point t in the interval −L≦j≦L
The moving average y(t) obtained by multiplying the value of +j by weight qj is expressed as follows.

Therefore, the moving averaged average value function and fluctuation value function are respectively expressed as moving average average value function: Σqj m(t+D−2°1 moving average fluctuation value function: Σqj ε(t+j)I″1 .

This apparatus calculates the moving average value of the encoding speed i1 within a predetermined period (tl-t2) using equation (1) for the image encoded signal shown in FIG. 3, and obtains the moving average value shown in FIG. 4. Then, as shown in FIG. 5, the threshold value At.

For example, let A1 be the average value of the transmission speeds assigned to the terminals, and A2 be the maximum value of the transmission speeds. and,
When the moving average value enters the overload region, the signal of that terminal is prioritized and packetized. When entering a high load area, packetization is stopped. This regulation will be discussed later.

Since the moving average value is the history of a predetermined period of time, it indicates a prediction that the current trend will remain constant for some time.Therefore, regulations based on this moving average value are appropriate for terminals and are appropriate for switching networks. Capacity overflow can be prevented. Further, in the buffer, the apparatus sets the buffer usage rate threshold B in FIG. 6 corresponding to the thresholds Al and A2 described above.
1 and B2, and set a normal area, a high load area, and an overload area, and perform similar regulations.

FIG. 1 is a functional block diagram of an image packet multiplexing and transmitting device according to the present invention.0 Image packet multiplexing and transmitting device 1 converts image signals 11 to 1n from n terminals into multiplexed packets 20 and outputs the same. It is a device. Encoding units 21-2
n inputs image signals 11 to 1n, performs image band compression and redundancy reduction through information source encoding and entropy encoding, and outputs encoded signals 51 to 5n. Buffer memory sections 41 to 4n input the encoded signals 31 to 3n, store image information and convert speed by first-in first-out (FIFO) operation, and convert the buffered signals 51 to
~5n is output.

The packet signal units 61 to 6n input this signal and synchronously output packet signals 71 to 7n shown in FIG. 7. The multiplex packet unit 17 uses this signal as an information part, adds a header to it, generates a multiplex bag 20, and sends it to the transmission path.

The above signals are controlled by the following parts. The moving average calculation units 121 to 12n input the time series data of the encoded signals 31 to 3n input to the buffer memory units 41 to 4n, calculate a moving average value of the encoding speed, and convert the moving average signal 31
1 to 31n are output to the transmission allocation processing unit IO. The buffer usage rate calculation units 131 to 13n are based on the buffer memory unit 41.
~4n usage rate is calculated, and the buffer usage rate calculation signal 32
1 to 32n are output to the transmission allocation processing section 10. The bar 2 buffer read control units 151 to 15n control reading of the buffer signals 51 to 5n from the buffer memory units 41 to 4n according to instructions from the transmission allocation processing unit 10. The transmission allocation processing unit 10 generates moving average signals 311 to 31n, par 2
It inputs the buffer usage rate calculation signals 321 to 32n, and outputs instruction signals to the buffer readout control units 151 to 15n and the packet multiplexing unit 17 based on a built-in transmission allocation processing program. The processing conditions in this case are given by the transmission quality setting signal 30 at the time of initial setting.

The transmission allocation processing unit IO gives different instructions depending on the three states of FIG. 5 and FIG. 6: normal, high load, and overload.

Process this! This will be explained with reference to FIG. As shown in the figure, the process consists of a moving average value determination process 100 and a buffer usage rate profit process 110. In the process 100, first, it is determined whether the moving average value Y is greater than or equal to the threshold value A1 (step 102). If the moving average value Y is smaller than the threshold value A1, it is determined that the state is normal and instructs permission to read the buffer by sequentially selecting 1/-n channels (step 10EI); If so, it is determined whether Y is greater than or equal to the threshold value A2 (step 104). If the moving average value Y is smaller than the threshold value A2, it is determined that there is a high load state, and reading of buffers larger than the threshold value 4g A I is permitted. , that is, instructs permission to read the buffer based on priority selection (step 10B).
), if the moving average value Y is greater than or equal to the threshold value A2, Y is greater than the threshold value A2/he buffer read restriction, that is, buffer read restriction due to non-selection is instructed (step 114).
).

In the buffer usage rate determination process (110), processes 112 and 113 are similarly performed to determine whether or not the buffer usage rate F is greater than the thresholds B1 and B2. Step 11
In step 2, if the buffer usage rate F is smaller than the threshold value Bl, it is determined that the state is normal, and instructions are given to sequentially select buffer reading permission.

If the buffer usage rate F is greater than or equal to the threshold B1, it is determined whether the buffer usage rate F is greater than or equal to the threshold B2 (
Step 113). If the buffer usage rate F is smaller than the threshold B2, it is determined that the load is high, and an instruction is given to allow reading of the buffer by priority selection (step 108).

If the buffer usage rate F is equal to or higher than the threshold value B2, it is determined that there is an overload state, and an instruction is given to restrict reading of the buffer due to non-selection (step 114).

The packet multiplexer 17 distinguishes between normal, high load, and overload based on the selection signal 34 from the transmission allocation processor 10, and edits the multiplexed packet 20. As shown in Figure 7, during normal times, fixed-length packets a (1 = n channels), during high loads, variable-length buckets whose time length is shorter than the former).
Transmit b(1, m, channel). No packets are transmitted during overload.

(Effects of the Invention) According to the present invention, the signal input speed of a buffer corresponding to a plurality of terminals is monitored using a moving average value with a predictive function, and the regulation is performed using a threshold value of a moving average value, and the regulation is performed using a threshold value of a buffer usage rate. Execute parallel buffer output regulation by By multiplexing packets from all terminals during normal times, transmitting packets only from highly loaded terminals during high load times, and not outputting them during overloads, overload conditions in high-speed packet networks and packet discards at terminals are properly prevented. This has the effect of improving transmission efficiency.

[Brief explanation of the drawing]

Fig. 1 is a functional block diagram of an embodiment of the image packet multiplexing device according to the present invention, and Fig. 2 is! @ A flowchart of parallel control processing using the moving average and buffer usage rate by the device in Figure 1, Figure 3 is a waveform diagram of the image encoding signal, Figure 4 is a diagram of the moving average value, Figure 5 is the image encoding FIG. 6 is a diagram showing a signal threshold value, FIG. 6 is a diagram showing a buffer usage rate and its threshold value, and FIG. 7 is a diagram showing a multiplexed packet. Description of part codes i, . . . Image packet multiplexing device 10, . 21-2n, . 41-4n, . 61-6n, . 121 ~12n 131 ~13n 151 ~15n

Claims (1)

[Claims] 1. Encoding means connected to a plurality of terminals in a packet switching network, corresponding to the terminals, inputting an image signal from the terminals, and outputting an encoded signal; a buffer means for inputting, temporarily storing and outputting the encoded signal; a packet assembling means for converting the encoded signal from the buffer means into a packet signal and outputting the encoded signal; packet multiplexing means for generating coded packets and outputting them to a transmission path; moving average value calculation means for calculating the input speed of the encoded signal input to the buffer means as a moving average value for a predetermined period; and the buffer means. buffer usage rate calculation means for calculating the usage rate of the buffer means; and buffer readout control means for stopping the output signal of the buffer means, the input speed of the input signal to the buffer means calculated by the moving average value calculation means. The moving average value of is compared with a first threshold value and a second threshold value of the moving average value, and the high load control is performed such that packet signals from the buffer means exceeding the first threshold value are selected preferentially and packetized. and transmission control means for instructing the packet multiplexing unit and for instructing the buffer read control means to control readout by non-selection for packet signals from the buffer means that exceed the second threshold. image packet multiplexing device. 2. The apparatus according to claim 1, wherein the transmission control means compares the usage rate of the buffer means calculated by the buffer usage rate calculation means with a first threshold value and a second threshold value of the buffer usage rate, Instructs the packet multiplexing unit to perform high load control to preferentially select and packetize packet signals from the buffer means that exceed the first threshold, and An image packet multiplexing device characterized in that the buffer readout control means is instructed to control readout based on non-selection of packet signals.