KR20130116937A - Method for improved robust header compression with low signal energy - Google Patents

Abstract

A method is provided for a wireless system to reliably transmit context information of robust header compression at low signal energies. In particular, the methodology of the present invention provides an extended virtual transmission time interval (super) for the first (not the first) uncompressed packet (including the header of the upper layer) of the packet stream transmitted using ROHC compression. TTI). Improved reception reliability for uncompressed packets, in particular for decoding headers, is achieved by the application of TTI bundling techniques, RLC segmentation and super TTI of association with multiple HARQ processes, and the use of sufficient HARQ round trips. Caused by.

Description

Method for Improved Robust Header Compression with Low Signal Energy {METHOD FOR IMPROVED ROBUST HEADER COMPRESSION WITH LOW SIGNAL ENERGY}

The present invention generally relates to the coding of information in a wireless communication system.

Voice communication within wireless systems has conventionally been provided over circuit-switched wireless telecommunication networks of the prior art, which allows the mobile unit to establish a fixed connection or link to the network via a bearer. However, as wireless technology evolves from voice-centric to data-centric—a process that is expected to begin in third generation (3G) systems and complete in fourth generation (4G) systems, voice signals (data and Together) are transmitted progressively over a packet-switched network using the Voice Over Internet Protocol (sometimes referred to as VoIP).

For the implementation of VoIP in a wireless telecommunications network, voice data frames are inserted into Internet Protocol (IP) data packets. IP packet headers can increase the size of data packets transmitted by the wireless telecommunications network, thus reducing the capacity of the packet-switched wireless telecommunications network to some extent as compared to the circuit-switched wireless telecommunications network. However, IP packet headers can generally be compressed, reducing the overhead associated with transmitting voice data using VoIP. Robust header compression (ROHC), a commonly applied compression method, reduces the size of an IP packet header by removing predictable and / or static information from the header. Using ROHC, the size of the IP Hacket header can be reduced by approximately one order by reducing the information in the IP packet header for sequence number, context identifier and time stamp. Considering the uncompressed packet payload, the size of the total packet transmission (payload + header) is reduced by approximately half by this ROHC compression.

Not all IP packet headers can be compressed. For example, it is generally necessary to send uncompressed or partially compressed IP packet headers when the mobile unit is initialized or resynchronized, or when the network recovers from an error. Uncompressed or partially compressed IP packet headers may include static and / or dynamic information that may be used to resynchronize the mobile unit or recover from one or more transmission errors. This static and / or dynamic information, which is mainly characterized by a "context" and is transmitted primarily for only the first few packets of the packet system, provides a context for the decompression of subsequent compressed packets of the packet stream. . However, for mobile terminals located within weak signal regions, such as cell edges, the signal energy available even at the maximum transmit power of the terminal may contain a packet with this contextual information (of course, in addition to the voice payload of the packet). May be insufficient to reliably transmit.

A method is provided for a wireless system to reliably transmit context information of robust header compression at low signal energy. In particular, the methodology of the present invention provides an extended virtual transmission time interval (super TTI) for the first (not the first) uncompressed packet headers of the transmitted packet stream using ROHC compression. Improved reception reliability for uncompressed packet headers is essentially through the application of super TTI obtained by the combination of TTI bundling techniques, RLC segmentation and association with multiple HARQ processes and through the use of sufficient HARQ rounds. Caused. The use of the super TTI of the present invention necessarily results in a large delay for the next few (particularly two or three) packets in the packet stream, and these delayed packets are thus discarded with care. In another embodiment, the methodology of the present invention addresses alternative approaches for doing discarding of delayed packets.

The teachings of the invention can be easily understood by considering the following detailed description in conjunction with the accompanying drawings.

1 schematically illustrates the methodology of the present invention.
2 illustrates an alternative description of the methodology of the present invention.
3A is a first part of a flowchart illustrating one embodiment of the overall process of the methodology of the present invention.
3b is a second part of the flow chart illustrating one embodiment of the overall process of the methodology of the present invention.

In the following description, for purposes of explanation, and not limitation, certain details, such as particular structures, interfaces, techniques, and the like, are set forth in order to provide a thorough understanding of exemplary embodiments of the present invention. However, it will be apparent to those skilled in the art that the present invention may be practiced in other exemplary embodiments that depart from these specific details. In some cases, detailed descriptions of well-known devices, circuits, and methods are omitted so as not to obscure the description of the described embodiments with unnecessary detail. All principles, aspects, and embodiments are intended to include their structural and functional equivalents in conjunction with their specific examples. In addition, these equivalents are intended to include both presently known equivalents and equivalents to be developed in the future.

The present invention is described hereinafter with respect to the adaptation of known transmission-time-interval approaches to achieve reliable transmission of ROHC context information in conditions of low signal energy. Although the disclosed methodology of the present invention is generally described for an exemplary case of implementation within a wireless system that provides VOIP service in accordance with LTE system standards and applies the ROHC compression methodology, the concept of the present invention is a power-limited mobile. Although reliable reception of several packets from the transmitter can operate to establish a context for subsequent packets transmitted at a low information rate, the mobile transmitter needs the information rate needed for these several packets at its maximum transmit power level. It is obvious that it can be applied to other wireless applications, which cannot achieve.

As mentioned in the background section, the application of the ROHC algorithm for packet compression requires a small number of initial packets (possibly only one) to build up contextual information for decompression of subsequently transmitted compressed packet headers. For a single packet) the complete packet header needs to be sent. Since the information content (number of bits) for a complete packet header is much larger than that of headers later compressed via ROHC compression, the bandwidth and / or total signal energy required for the transmission of context information provided by the uncompressed header is It is also generally much larger than necessary for compressed headers. This is a significant issue for a mobile terminal transmitting ROHC compressed packets when the terminal is located in a weak signal region, such as at the cell edge. In such situations, the same maximum power, even though the signal energy per bit associated with transmission from the mobile terminal located at a considerable distance from the serving base station may be perfectly suitable for reliable transmission of relatively less compressed packet headers. The per-bit signal energy available from mobile terminals of the same distance to transmit to may be insufficient for reliable communication of even larger complete packet headers at the start of the transmitted packet stream using ROHC, because of complete header information This is because the data rate required for transmission cannot be achieved at the available signal energy level.

Before describing the methodology that enables the present invention to address such problems, it may be useful to briefly describe certain parameters or concepts known in the art to be mentioned during the course of such technology. First, a parameter called transmission time interval (TTI) is used in many wireless communication protocols to indicate the time interval required for transmission of a unit of data provided for transmission, which unit is generally a packet. The duration of TTI in LTE is 1 ms.

In order to improve the reliability of information transmission over the wireless link, error correction protocols are mainly applied. This powerful error protection protocol that is frequently applied is called Hybrid Automatic Repeat Request (HARQ). The HARQ process includes feedback from the receiver to the transmitter that the transmitted packet was received correctly (ACK), or alternatively that the packet was not received correctly (negative ACK or NACK). Through the HARQ process, error feedback is normally sent following each packet transmission, and a given packet will be retransmitted in response to receipt of a NACK (or absence of ACK and a timeout of a timer), and this retransmission of a given packet is received by the ACK. It lasts until The reception of the NACK from the start of packet transmission and the time of start of packet retransmission are referred to as HARQ round trip time or HARQ RTT. In addition, undecoded received symbols of previously failed transmissions of a given packet in HARQ are combined with the current one when attempting a new decoding-thus the energy spent on previously failed transmissions is not wasted and effectively accumulated.

One approach in the prior art to improve the likelihood that packets are received correctly for transmitted packets in the environment of available signal energy associated with a given TTI lower than required for reliable transmission of the information content of the packet is Known as RLC segmentation. In such an approach, the radio link control (RLC) protocol layer is capable of several transmission units (generally referred to as protocol data units (PDUs)) received from higher layers (such as the packet data convergence protocol (OPDCP) layer). It operates to subdivide into smaller units, each of these smaller units being transmitted in a separate TTI. While this approach is generally successful in overcoming signal energy limitations for remote mobile units, it suffers from the drawback of significantly increasing overhead and control signaling. Thus, an approach called TTI bundling has been developed to more effectively address the problem that mobile users cannot transmit the entire complete header packet during one TTI due to the lack of available signal energy. The TTI bundling approach is described later.

As is known, the transmission energy of a wireless signal is a function of transmitter power and transmission time. Thus, the higher transmission energy needed for a given required data rate from a distant mobile can be achieved through an increase in transmission power or through the use of longer transmission intervals. Since the transmit power is already considered at the maximum level at the cell edge, the only option is an increase in transmit time. In keeping with this constraint, TTI bundling operates to transmit a full-header packet—eg, an initial VoIP packet (corresponding to the PDCP PDU described above) —as a single PDU during a bundle of subsequent TTIs without waiting for HARQ feedback. HARQ feedback is only expected during the last transmission of the bundle. If the HARQ feedback is negative (ie, NACK), then the packet is retransmitted within the subsequent TTI bundle as further described below.

During the pronunciation interval, the user's voice packets are sent for all standard codecs to start transmission over the air under semi-persistent scheduling (at 0, 20, 40, ..., tick marks by established protocol). It is well known to arrive at ms intervals. For an exemplary LTE wireless system considered as the basis for this description, under normal TTI (1 ms) operation, the HARQ round trip time is 8 ms, so that up to five consecutive HARQ round trips of the first packet are for semi-persistent scheduling. It occurs at 0, 8, 16, 24 and 32 ms ticks, but it cannot be extended to 40 ms ticks because it can cause a collision with the third packet.

For TTI bundling, four TTI / subframes (4 ms bundle) per HARQ round trip are aggregated to transmit one RLC / MAC PDU. However, the balance for TTI bundling is that the HARQ round trip time (RTT) is also doubled to 16 ms.

Moreover, because the TTI bundling will be four times the normal TTI interval (4 ms), and since the corresponding HARQ RTT will be 16 ms, the result is that the voice packet is 0, 16, 32, 48, 64 ms tick markers. By using the five HARQ round trips that occur at, we technically use a delay budget of up to 80 ms without worrying about collisions with subsequent voice packets. Thus, the user now obtains four times as much transmission energy (i.e. improvement of 6 dB link budget through normal TTI operation) by the corresponding resource usage cost.

Increasing total transmission time, as in TTI bundling, generally suffers from two limitations. For one, the air-interface delay constraint (derived from the end-end delay constraint) requires the packet to be delivered within D seconds of its arrival at the transmitter of the mobile unit. This delay is usually related to the quality of service required for a particular application. For example, VoIP is relatively delay tolerant, while some data applications can tolerate fairly large end-to-end delays. Moreover, in a queue system with packets arriving every T seconds, the service time for processing and transmitting individual packets must not be greater than the length T of the packet arrival interval, so that the queue cannot grow indefinitely. Prior art TTI bundling approaches are packets that contain a complete context header and will generally not be able to accommodate these constraints for packets sent from mobile terminals at or near the cell edge.

In these situations-a packet with a complete context header sent from the cell edge, even with a packet service time equal to the packet arrival interval T, the available transmit power ensures successful packet reception for the initial ROHC packet. It is not likely to be enough. Although this scenario may seem to present a failed call to such an application—and this will be obvious for individual packets—but it is not necessarily the case for a stream or harmonization of packets containing such an application.

 Despite the improvements in usable transmission energy generated by TTI bundling, none of the prior art TTI bundling solutions are suitable for supporting large VoIP payloads without ROHC. This would be the case for both standalone legacy networks and overlay systems, where the next generation or upgrade technology network is overlaid with the legacy network, such as the example LTE network described herein that is overlaid with the legacy 3G network. To address this drawback, we disclose an improvement on the prior art TTI bundling approach, which is characterized herein as a super TTI.

The basic approach of the present invention temporarily violates the stable queue limitations (and semi-persistent scheduling limitations) of the disclosed techniques, allowing at least the first uncompressed ROHC packet to be transmitted with the desired high probability of success. By establishing a ROHC context, subsequent packets can be compressed for transmission by the ROHC. The super TTI bundling approach of the present invention is expected to provide transmit energy higher than 20 ms (60-80 ms) for the first VoIP packet.

The operation of the super TTI bundling methodology of the present invention is schematically illustrated in FIG. 1. According to the methodology of the present invention, after establishing the PDCP PDU in the drawing-such that the VoIP bearer queue of the mobile-terminal / user-equipment (UE) has data to transmit, the base station (eNB) is configured in (a complete header packet of VoIP). Schedule a grant for a transport block size (TBS) smaller than the PDCP PDU (VoIP) packet size. In a preferred embodiment, in order to force the UE RLC layer to subdivide the PDCP packet into four parts, RLC PDU1, RLC PDU2, RLC PDU3 and RLC PDU4 in FIG. 1, the TBS should be roughly determined according to the following equation:

TBS = (PDCP + Low Layer Overhead) / 4.

Each RLC segment is then transmitted on up to Q HARQ round trips with TTI bundling turned on (to minimize RLC subdivision and overhead). As shown, the RLC PDU1 includes a TTI including TTIs 16, 17, 18 and 19 during a second HARQ round trip, within a TTI bundle containing TTIs (0, 1, 2 and 3) during a first HARQ round trip. Within a bundle, within a TTI bundle containing TTIs 32, 33, 34 and 35 during a third HARQ round trip, and so on, for up to five HARQ round trips. RLC PDUs 2-4 are similarly transmitted in the next adjacent TTI bundles for up to five HARQ round trips, as shown in the figure. Thus one long "super TTI" is created with a length ('supertti' = Q * 4 * 4) (Q is a function of the number of HARQ round trips used, and insufficient transmission energy). An exemplary value of Q = 5 means a supertti of 80 ms. Early termination can reduce the realized value of supertti.

Two other points should be noted for the methodology of the present invention shown in FIG. First, as previously mentioned, voice packets arrive at the transmission buffer at 20 ms intervals, so new packets are sent TTI 20 and TTI 40 (20 and 40 ms from the start of the supertti transmission of the initial packet). Will reach the buffer). As described in more detail below, these two packets (and the third packet to arrive for transmission in the TTI 60, if full five HARQ round trips are used) are discarded for long-duration transmission queue delays. Will be. In addition, although the figure only illustrates the operation of the supertti methodology for the initial packet (including the header context information), after capturing the context information for such an initial packet through the use of the supertti methodology, subsequent transmitted packets are sent to the ROHC. It is to be understood that it is compressed and therefore transmitted using a normal HARQ process with a normal total transmission time (typically up to 20 ms) for long term transmission queue stability.

An alternative technique of the super TTI process is shown in FIG. 2, which has an extended time scale and describes packets queued in the transmission buffer at different times within the process. 2 shows a set of graphs describing time evolution for an example extended transmission (super TTI) of 64 ms. Five single shaded 230 indicator bars, 16 ms apart and 4 ms wide, are up to 5 HARQ transmissions with a delay budget of 80 ms for a total transmission time of 20 ms, and the first within a given packet stream. Represents a "best" normal behavior for sending uncompressed packets. The elongated transmission time over successive HARQ interlaces of the methodology of the present invention is shown by a line graph 250 line of square data points with falling lines. This causes the mobile's application queue to build up rapidly, as shown by double shaded 210 bars representing the number of packets queued. This construction is also the result of the constant arrival of new packets at the mobile terminal indicated by cross shaded 220 indicator bars spaced every 20 ms. After the extended transmission has ended successfully, most of the queue pressure is quickly alleviated by scheduling the number of packets in the next HARQ interlace, indicated as a peak in the line of triangular data point line graph 240 immediately after 64 ms. As can be seen, this operation causes the double shaded 210 bars to drop abruptly to 1 and ultimately to zero, ensuring that no transmission queue delay is performed for future scheduled packets. One embodiment of the present invention carefully confines scheduling of these packets into only one HARQ interlace (and deceptive acknowledgment) in order to prevent delay propagation downstream even if transmission failure is almost inevitable. As such, Spearhead Compressed Packets have a normal mode (normal total transmission time, when they arrive without any queue establishment and with the normal number of packets (typically not greater than 1) contained within each HARQ interlace. Ie typically up to 20 ms).

As mentioned above, the super TTI process will generally stop upon successful delivery of uncompressed ROHC packets or upon reaching a limit in the number of uncompressed ROHC packets to which the process applies. The number of packets transmitted via the super TTI process is then indicated by "N". Once successful reception of the uncompressed packet is obtained, two additional events (possibly overlapping) are triggered before the regular ROHC transmission prevails in SPS mode: one is real time initiation, as described below. Corruption "is the delivery of the actual ROHC feedback header on the downlink (DL) instructing the UE for ROHC transmission and (almost) simultaneous cancellation of packets.

Due to the extended transmission duration associated with the transmission of the first N packets (typically N = 1), the super TTI process generally results in inherent loss of several packets (typically 2-3) due to aging. And the number of "damaged" packets is indicated here by "K". In the preferred case, the scheduler will erase / delete the K corrupted packets from the UE buffer and will quickly resume packet transmission from the (N + K + 1) th packet. For this erase / erase function, two alternative additional embodiments are provided:

(1) a set of discard timers for these packets, i.e., a timer for discarding corrupted packets if an ACK is not received within an appropriately short interval after the transmission of the packet, or

(2) transmitting a scheduling grant proportional to K and scheduling for exactly one HARQ round trip with the lowest possible resource allocated; And then forcing the ACK upon receipt of the packet, regardless of the actual fate of the received packet (ie, wrong ACK).

In an alternative approach to the above, which continues the super TTI process until an uncompressed packet has been successfully received, the super TTI process has a first non-ROHC packet completing all HARQ round trips for all RLC segments. Then stop (ie N = 1). In this alternative, it will transmit a “false” ROHC feedback header on the prepared DL prior to the expected time of the end of the first non-ROHC transmission. If the packet was not successfully received at the end of the first super TTI, additional ROHC packets cannot be decoded by the application and the call is therefore aborted. If successful reception is achieved, ROHC synchronization is followed, the super TTI process is terminated, corrupted packets in ROHC mode are erased as above, and the regular ROHC transmission is later dominated (in SPS mode if necessary).

Flowcharts describing the overall process of the methodology of the present invention for one embodiment of the present invention are shown in FIGS. 3A and 3B. Referring to the figure, the process begins at step 301, which selects the TTI length and the number of HARQ interfaces to be performed (Q), and the VoIP packet via the base station transmitting this information to the mobile station via a scheduling message. It is characterized by the start of the event E, such as the start of the stream. The uncompressed packet counter (counting N) is also set to one. In a next step 303, the base station issues the number of scheduling grants (Q) to the mobile station for portions of the first packet in the provided packet stream, each portion being the sum of the overhead contribution divided by the initial packet size and Q. Is the same as Then, in step 305, for each scheduled partial packet segment, the base station grants a large number of HARQ round trips for the segment to increase the probability of successful decoding for that segment. As indicated in the figures (and hereinafter), the selection of the number of allowed HARQs is determined dynamically as a function of transmission times, segment sizes and transmission resources, but in general, as a limit required to maintain transmission queue stability It is expected to exceed the number of round trips that were previously considered.

In step 307 a determination is made whether transmission of any partial segments is successful after L HARQ round trips. If the answer is affirmative, the process proceeds to step 306 and if negative, the process proceeds to step 309. If at least one of the partial segments has not been successfully decoded (negative in step 307), the value of the uncompressed packet count N is checked in step 306. If this value is greater than a predetermined threshold value M (generally higher than that, determined as the number of iterations of the super TTI methodology in which successful transmission / decoding of the provided packet stream is impossible), the process proceeds to step 310 with call termination. Proceed to). On the other hand, if the counter coefficient N is less than the threshold value M, the counter value is increased by the counter step of 1 in the method step 308, and the process returns to the step 303.

If at step 307 it is determined that all the partial segments have been successfully decoded (and thus provided the necessary context for ROHC compression of subsequently transmitted packets), the process proceeds to step 309, which step ROHC compression is implemented in accordance with the protocol at the mobile station and the base station. Thereafter, in step 311, the base station issues a schedule grant for " corrupted " packets that are held in the transmission buffer as a result of the operation of the super TTI methodology, and then applies additional packets resulting from normal processing of these corrupted packets. In order to rule out additional transmission delays, the " false ACK "

Finally, at step 313, once ROHC compression is implemented and corrupted packets are cleared from the transmit buffer, the base station has a normal total transmission time (typically up to 20 ms) and a normal number (typically less than 1) within the HARQ interlace. Dynamic or semi-persistent scheduling with (not large) upper layer packets is used to continue scheduling the remaining packets in the provided packet stream.

In a pair of alternative embodiments of the invention, the entire PDCP packet is semi-persistently scheduled through a single interlace (HARQ process) with TTI bundling in the first alternative embodiment, and dynamic in the second alternative embodiment. It is passed in one RLC packet for the MAC to be scheduled as a single large over-the-air packet scheduled for. TTI bundling allows the maximum air interface delay to be extended from 33-34 ms to 68-80 ms, and allows a reasonable maximum number of HARQ round trips to be supported by a single interlace design (one RLC packet per PDCP packet). do. This allows enhanced transmission energy by 20 ms.

Both semi-persistently scheduled and dynamically scheduled embodiments use 100% of resources (time and HARQ processes), thus providing the highest transmission energy possible. 100% usage is also semi-persistently scheduled, since there is no separate freedom to introduce leverage to dynamically schedule at the time of the first HARQ round trip, and in this sense there will be no unnecessary separate DL overhead at all. It is meant that the disclosed embodiments will be preferred over the dynamically scheduled embodiments.

In the present specification, the inventors have disclosed a method for providing increased transmission energy for wireless terminals located in weak signal regions. Many modifications and alternative embodiments of the invention will be apparent to those skilled in the art in light of the above description.

Thus, the description is intended to be illustrative only, and is to be construed to teach those skilled in the art the best mode of carrying out the invention, and is not intended to show all possible forms of the invention. The words used are words of description rather than of limitation, details of the structure may be changed without departing substantially from the spirit of the invention and the exclusive use of all modifications falling within the scope of the appended claims is reserved. Will also understand.

Claims (10)

A method for packet transmission in a wireless communication system,
Dividing the packet provided for transmission into Q equal parts; And
For each portion, scheduling transmission over a plurality of transmission time intervals;
Method for packet transmission. The method of claim 1,
And the plurality of transmission time intervals are contiguous in time. The method of claim 1,
Upon successful decoding of all parts of the divided packet, the splitting step ends. The method of claim 3, wherein
Upon successful decoding of all parts of the divided packet, a context for ROHC compression of subsequently provided packets is provided. The method of claim 3, wherein
And packets provided for transmission subsequent to successful decoding of all portions of the divided packet are compressed using ROHC compression. The method of claim 3, wherein
Upon successful decoding of all parts of the divided packet, one or more packets waiting in the transmission queue are scheduled together and acknowledged as being received regardless of the actual output of the reception / decoding. The method of claim 1,
The total transmission time over multiple transmission intervals exceeds the limit needed for a stable transmission queue. The method of claim 1,
And the dividing step is implemented by dividing the provided packet size by Q and adding an overhead factor to each of the partial packet increments so determined. The method of claim 8,
A scheduling grant is issued for each of the Q packet portions by a serving base station. The method of claim 1,
Receiving, at the mobile station, scheduling grants for multiple (Q) HARQ interlaces over multiple transmission time intervals;
Subdividing the initial packet provided for transmission into Q equal parts at the base station, each part being scheduled on its HARQ interlace within the plurality of time intervals;
At the end of the L HARQ round trips for each of the Q portions, returning to the subdividing step for the next portion waiting in the transmission queue of the base station if any portion was not successfully decoded;
If all of the Q portions were successfully decoded using each of the less than L HARQ round trips, implementing ROHC compression for the subsequent portion provided for transmission;
Upon successful decoding of all parts for a given packet, discarding one or more damaged parts waiting in the transmission queue; And
Continuing transmission of the remaining portions provided for transmission based on ROHC compression;
Further comprising a packet transmission.

Images