Seminar. Ausgewählte Kapitel der Nachrichtentechnik, WS 2009/2010. LTE: Der Mobilfunk der Zukunft. Random Access. Almamy Touray. 02.

Transcription

1 Seminar Ausgewählte Kapitel der Nachrichtentechnik, WS 2009/2010 LTE: Der Mobilfunk der Zukunft Random Access Almamy Touray 02. December, 2009 Abstract An LTE User Equipment (UE) can only be scheduled for uplink transmission if its uplink transmission timing is synchronized. The LTE Random Access Channel (RACH), therefore, plays a key role as an interface between nonsynchronized UEs and the orthogonal transmission scheme of the LTE uplink radio access. 1 Introduction The growing demand for mobile Internet and wireless multimedia applications such as Internet browsing, interactive gaming, mobile TV, video and audio streaming has motivated development of broadband wireless access technologies in recent years. As a result, the 3rd Generation Partnership Project (3GPP) initiated the work on the long-term evolution (LTE) in late LTE will ensure 3GPPs competitive edge over other cellular technologies. The evolved UMTS terrestrial radio access network (E-UTRAN) substantially improves end-user throughputs, sector capacity and reduces user-plane and control-plane latencies, bringing significantly improved user experience with full mobility. With the emergence of Internet Protocol (IP) as the protocol of choice for carrying all types of traffic, LTE is expected to provide support for IP-based traffic with end-to-end quality of service (QoS). Voice traffic will be supported mainly as voice over IP (VoIP) enabling better integration with other multimedia services. Initial deployments of LTE are expected by 2010 and commercial availability on a larger scale will likely happen a few years later. Unlike its predecessors, which were developed within the framework of Release 99 UMTS architecture, 3GPP has specified the evolved packet core (EPC) architecture to support the E-UTRAN through reduction in the number of network elements and simplification of

2 2 Almamy Touray functionality but most importantly allowing for connections and handover to other fixed and wireless access technologies, providing the network operators the ability to deliver a seamless mobility experience. 3GPP has set aggressive performance requirements for LTE that rely on improved physical layer technologies such as orthogonal frequency division multiplexing (OFDM) and single-user and/or multi-user multiple-input multiple-output (MIMO) techniques. 2 Brief Overview of Random Access (RA) Process in LTE The Random Access Channel (RACH) is an uplink channel in mobile communication system that is used to transfer control information from a mobile terminal to the network, e.g. for initial access to set up a connection or for location area updates. The RACH channel may be contention based channel where several users might access the same resource. There is usually no knowledge about the required transmit power of the mobile terminal and thus an open loop power control method is applied. User Equipments (UEs) can only be scheduled for uplink transmission if its uplink transmission timing is synchronized. The RACH, therefore, plays a key role as an interface between nonsynchronized UEs and the orthogonal transmission scheme of the uplink radio. In UMTS, RACH is primarily used for initial network. Once uplink synchronization is achieved for a UE, the enodeb (Enhanced Node B or simply base station) can schedule orthogonal uplink transmission resources for it. 3 System Model of Random Access In the system description of Random Access model, we have N UEs. Each UE s signal arrives at the enodeb with different timing due to different propagation delays. When the enodeb receives a RA signal, it performs basic signal processing including CP removal, DFT, and sub-carrier de-mapping.the Random Access detector consists of a number of correlators and each correlator correlates the received signal to one of the available Random Access codes. Each correlator further investigates the maximum correlation output and compared it to the predefined threshold, and a final decision is determined (that is from which UE the detected preamble was transmitted) 4 What situations require Random Access? 1. UE in RRC_CONNECTED state but not uplink-synchronized but needs to send new uplink data or control information (e.g. an event-triggered measurement report).

3 Random Access 3 Figure 1: System Description of Random Access 2. UE in RRC_CONNECTED state but not uplink- synchronized but needs to receive new downlink data, and, therefore, to transmit corresponding ACK/NACK in the uplink, 3. UE in RRC_CONNECTED state handing over from its current serving cell to a target cell 4. transmission from RRC_IDLE state to RRC_CONNECTED. For example for initial access or tracking area updates, 5. Recovering from radio link failure. 4.1 RACH in UMTS versus RACH in LTE In both schemes, the primary role of RACH is for initial network access but there exists no possibility to transmit user data in LTE RACH. This is exclusively sent on the Physical uplink Shared CHannel (PUSCH).The main purpose of LTE RACH is to achieve uplink time synchronization for UEs. Also, in LTE RACH, there is no power ramping as in the case of UMTS, which can be a major course for latency and interference. 5 Random Access Process In this model, enodeb broadcasts information regarding random access and multiple UEs transmit randomly selected random access code and the enodeb allocates resources to detected UEs. Each UE transmits detailed information like the random access code etc using allocated resources. The enodeb confirms to each UE When the above steps are finished successfully, enodeb and each UE initiate data communication. The design of RACH comes in two ways:

4 4 Almamy Touray Figure 2: Message sequence chart for the Random Access procedure 5.1 Contention based This is a situation where several UEs may access the same resource and, therefore, the possibility of collision between them.the procedures here are as follows: 1. Preamble Transmission: The UE selects one of the 64 PRACH contention-free signatures. The enodeb can control the number of signatures in each subgroup according to the observed loads in each group 2. Random Access Response: The Random Access Response (RAR) is sent by the enodeb on the Physical Downlink Shared Channel (PDSCH), and addressed with an ID, the Random Access Radio Network Temporary Identifier (RA-RNTI), identifying the time-frequency slot in which the preamble was detected. If multiple UEs had collided by selecting the same signature in the same preamble time-frequency resource, they would both receive the same RAR. 3. Layer 2/Layer 3 (L2/L3) Message: This message is the first scheduled uplink transmission on the PUSCH and makes use of Hybrid Automatic Repeat Request (HARQ), also conveys UE identifier etc. It conveys the actual random access procedure message. 4. Contention resolution: The User Equipment s (UE s) behaviour upon reception of contention resolution message has three possibilities: The UE correctly decodes the message and detects its own identity: it sends back positive acknowledgment ACK. On the other hand, if the UE correctly decodes the message and discovers that it contains another UE identity (contention resolution), it sends nothing back (Discontinuous Transmission,

5 Random Access 5 Figure 3: Contention-based Random Access procedure DTX). Also, its possible that the UE cannot decode the message and here also, it sends nothing back (DTX). 5.2 Collision Detection In the collison detection figure, we have two power delay profile: the one with small cell size and the one with large cell size. In small cell size, collision detection is impossible and this is due to the small relative propagation delay between the cells. For example if one UE in close to the base station and the other closer to the cell border but because of the small cell size, the enodeb will end up decoding the two transmitted preambles by the two UEs as one preamble.on the other hand, in the other power delay profile with large cell size, collision detection is possible. 5.3 Contention free In Contention free mode, the enodeb assigns distinct preamble to each UE and hence the concern for collision and other collision related issues are non-existant. Contention-free Random Access can be used in areas where low latency is required, such as handover and resumption of downlink traffic for UE. Here, dedicated signature is allocated to the UE on a per-need basis. Contention-free may be used in Handover, which comprises of two types: 1. Intra-RAT, which is within one radio access technology (i.e. LTE -to-lte from one enodeb to another) 2. Inter-RAT, between radio access technologies e.g.: between LTE and GSM or 3G WCDMA, WIMAX or even wireless LAN.

6 6 Almamy Touray Figure 4: Collision Detection in LTE RACH Figure 5: Contention-free Random Access procedure

7 Random Access 7 Figure 6: PRACH multiplexing with PUSCH and PUCCH 5.4 Contention based versus Contention free Contention-based can be used by any accessing UE in need of an uplink connection while Contention free Random Access can be use in areas where low latency is required, such as handover and downlink data arrival events. Nevertheless,in both procedures, the Random Access preamble is transmitted by the accessing UEs. In LTE RACH,there exists 64 Random Access preambles allocated for each cell of an enodeb, and each enodeb dynamically configures two disjoint sets of preambles to be used by the two Random Access procedures seperately. 6 Physical Random Access Channel (PRACH) Design PRACH: Physical Random Access Channel, PUCCH: Physical Uplink Control Channel, PUSCH: Physical Uplink Shared Channel. The PRACH, PUCCH, and PUSCH are orthogonal to one another and as such, preambles interfering with user data will not occur as oppose to UMTS 6.1 PRACH Structure We have two preambles at the enodeb received with different timings due t propagation delay. We used the GT (Guard Time) to absorb the propagation delay. It s then possible to copy a section from the end of the symbol to the beginning, known as Cyclic Prefix, CP. The receiver can then sample the waveform at the optimum time and avoid ISI caused by reflection. The length of the CP is very important. If it is not long enough, then it will not

8 8 Almamy Touray Figure 7: The PRACH preamble received at the enodeb counteract the multi-path reflection delay spread. On the other hand, if it is too long, then it will reduce the data throughput capacity. The maximum cell radius is given by the CP (Cyclic Prefix) length. Similarly to WCDMA, the LTE PRACH preamble consists of a complex sequence. However, it differs from WCDMA preamble in that it is also an OFDM symbol, built with a CP, thus allowing for an efficient frequency-domain receiver at the enodeb. As shown in the figure 7 or Bild 7 below, the end of the sequence is appended at the start of the preamble, thus allowing a periodic correlation at the PRACH receiver. The UE aligns the start of the random access preamble with the start of the corresponding uplink sub-frame at the UE assuming a timing advance of zero and the preamble length is shorter than the PRACH slot in order to provide room for a Guard Time (GT) to absorb the propagation delay. In figure 7 or Bild 7, there exist shows two preambles at the enodeb received with different timings depending on the propagation delay. Regarding detection performance, one would intuitively expect that the higher the bandwidth, the better the detection, due to higher frequency diversity. Nevertheless, under certain conditions, a smaller bandwidth will perform better than a large bandwidth in a single-path static AWGN channel, given that no diversity improvement is to be expected from such a channel. 6.2 PRACH Preamble Formats Coverage performance: In general, a longer sequence gives better coverage, but better coverage requires a longer CP and GT in order to absorb the corresponding round-trip delay.

9 Random Access 9 Figure 8: Field durations and achievable cell radius of the PRACH preamble formats Figure 9: Functional structure of PRACH preamble transmitter

10 10 Almamy Touray Figure 10: Hybrid frequency/time domain PRACH generation 7 PRACH Implementation The PRACH preamble can be generated at the system sampling rate, by means of a large IDFT as shown in figure 9. The DFT block is optional as the sequence can be mapped directly in the frequency domain at the IDFT input. The cyclic shift can be implemented either in the time domain after the IDFT, or in the frequency domain before the IDFT through a phase shift. Disadvantage of the functional structure of the PRACH preamble transmitter: It does not require any time-domain filtering at baseband, but leads to large IDFT sizes (up to 24, 576 for a 20 MHz spectrum allocation), which are cumbersome to implement in practice. Solution to the problems of the functional structure of the PRACH preamble transmitter: We can solve this problem by generating the preamble using a smaller IDFT, actually an IFFT and Shifting the preamble to the received frequency location through time-domain up-sampling and filtering. This results in hybrid frequency/time-domain generation. 7.1 Hybrid frequency/time-domain PRACH generation The CP can be inserted before up-sampling and time-domain frequency shift, so as to minimize the intermediate storage requirements. 7.2 enodeb PRACH Receiver In both the frequency-domain and the hybrid time/frequency domain approaches, we have the removal of the CP, the Power Delay Profile (PDP) computation and signature detection. In the full-frequency-domain approach, the DFT computation cannot start until the complete sequence is stored in memory, which increases delay. On the contrary, the hybrid time-frequency domain method first extracts the relevant PRACH signal through a time-domain frequency shift and down-sampling and anti-aliasing filter. The use

11 Random Access 11 Figure 11: PRACH receiver options of down-sampling and the anti-aliasing filter are for generating PRACH time samples suitable for FFT computation at a sampling rate which is an integer fraction of the system sampling rate. Unlike the full-frequency-domain approach, the hybrid time/frequency-domain computation can start as soon as the first samples have been received, which helps to reduce latency. 8 Summary 1. RA provides uplink synchronization necessary for scheduling of UEs 2. RACH bandwidth in LTE is minimal compared to WCDMA used in UMTS 3. RACH in LTE fits into the orthogonal time-frequency structure of the uplink compared to UMTS which uses WCDMA 4. Random Access in LTE can be contention-free 5. Contention- free allows for handover and other scenarios which require low latency 6. We use the Zadoff-Chu (ZC) sequence as RA preamble for LTE networks 7. The PRACH preamble enables the enodeb to estimate UE transmission timing

12 12 Almamy Touray 9 References References [Pro05] J. G. Proakis: Digital Communications, New York McGraw-Hill, [Smi05] [Sha,Ogo,Hat] [Pan] [Fra,Zad,Hei] [Chu] W. W. Smith, J. M. Smith: Handbook of Real-time Fast Fourier Transforms, New York:Wiley Inter-Science, M. Shafi, S. Ogose, T. Hattori: Wireless Communications in the 21st Century, New York: Wiley Inter-Science, Panasonic: Random Acess Burst Design for E-UTRA, Panasonic and NTT DoCoMo, 3GPP TSG RAN WGI, meeting 46, Tallinn, Estonia, August 2006 R. L. Frank, S. A. Zadoff, R. Heimiller: Phase Shift Pulse Coses with good Periodic Correlation Properties, IRE IEEE Trans. on Information Theory, Vol. 7. D. C. Chu: Polyphase Codes with Good Periodic Correlation Properties, IEEE Trans. on Information Theory, Vol. 18. [Ets] ETSIEN : Radio Transmission and Reception (Release 1999), [Hua] Huawei: 3GPP TSG RAN WGI,