`Athens, Greece
`March 27 – March 30, 2006
`
`Agenda Item:
`Source:
`Title:
`Document for:
`
`1.
`
`Introduction
`
`10.2.3
`Motorola, [], []
` Random Access Channel TP
`Discussion and decision
`
` R1-06xxxx
`
`The random access channel is used for initial access to the network as well as to transmit small to medium
`amount of control information. This contribution proposes a TP of random access channel design for E-
`UTRA.
`
`Start of Text Proposal
`9.1.2.1
`Random access procedure
`
`The random access procedure is at least used when the UE uplink has not been time synchronized and shall
`allow the Node B to estimate, and, if needed, adjust the UE transmission timing to within a fraction of the
`cyclic prefix. The random access burst consists of at least a signature sequence. Inclusion of additional data
`symbols is FFS.
`Random access and data transmission are time and/or frequency multiplexed.
`
`The random access procedure is classified into two categories:
`•
`non-synchronized random access, and
`•
`synchronized random access.
`
`9.1.2.1.1. Non-synchronized random access:
`
`The non-synchronized access is used when the i) UE uplink has not been time synchronized or ii) UE
`uplink loses synchronization. The non-synchronized access allows the Node B to estimate, and, if needed,
`adjust the UE transmission timing to within a fraction of the cyclic prefix.
`
`9.1.2.1.1.1 Time Frequency Structure
`
`Non-synchronized random access and data transmission are time and/or frequency multiplexed as
`illustrated in Figure 1. The minimum bandwidth, BWRA, allocated for synchronized random access
`transmission is 1.25 MHz. For system bandwidths larger than 1.25 MHz, either the random access
`transmission uses a larger bandwidth, or multiple random access channels are defined. Multiple 1.25MHz
`random access channels might be especially useful for the second part of a two part random access channel,
`or for selecting a best block using frequency selective channel characteristics (TDD mode).
`The length of the non-synchronized random access burst, TRA, is less than (multiples of) 0.5 ms to allow the
`burst, and the required guard time to account for the uplink timing uncertainty, to fit within a subframe (or
`multiples thereof).
`Alternatively, the non-synchronized random access can also be CDM based (with full bandwidth
`allocation) as shown in Figure 2. CDM allows for random access transmission independent of scheduled
`data, i.e. full flexibility in allocating both time and frequency domains and does not consume any additional
`overhead. However, the effect of interference on scheduled data and vice-versa needs to be investigated
`further.
`
`9.1.2.1.1.2 Preamble and message part
`
`APPLE 1030
`
`1
`
`
`
`The non-synchronized random access preamble is used for time alignment, signature detection etc. The
`message part conveys upper layer signaling required for connection establishment.
`
`
`
`
`
`Can be used for other random- access
`channels or data transmission.
`
`Data transmission
`
`0.5 ms subframe
`(10 ms radio frame)
`
`TRA-REP
`
`TRA
`
`BWRA
`
`
`
`(Scheduled) Data transmission(Scheduled) Data transmission
`
`Random- access preambleRandom- access preamble
`
`
`Guard timeGuard time
`Figure 1. TDM/FDM Option Example using 1 sub-frame
`CDM of Random Access Preamble and Scheduled Channels
`Random Access Preamble spreading with a long sequence
`f
`
`f
`
`
`
`Scheduled Data Channel
`Scheduled Data Channel
`Scheduled Data Channel
`Scheduled Data Channel
`Scheduled Data Channel
`Scheduled Data Channel
`Scheduled Data Channel
`Scheduled Data Channel
`
`Scheduled Data Channel
`
`t
`
`+
`
`t
`
`Random Access Preamble
`Random Access Preamble
`Random Access Preamble
`Random Access Preamble
`Random Access Preamble
`Random Access Preamble
`Random Access Preamble
`Random Access Preamble
`
`Random Access Preamble
`
`
`
`Figure 2 Example of CDM of Scheduled Channels and random access preamble
`
`
`9.1.2.1.1.3 Non-synchronized random access procedure
`
`Prior to attempting a non-synchronized random access, the UE shall synchronize to the downlink
`transmission.
`
`
`2
`
`
`
`Two possibilities for the procedure are considered:
`• One-step approach, where the Node B responds to the non-synchronized random access attempt
`with timing information to adjust the uplink transmission timing and an assignment of uplink
`resources to be used for transmitting the message part. The message part is transmitted according
`to the assignment. It may be noted that the timing information can also be combined with the
`uplink data resource allocation. Furthermore, the uplink data resource allocation may be
`implicitly indicated by associating a reserved time frequency region with a preamble sequence.
`Finally, the access preamble may also contain message payload.
`• Two-step approach, where the Node B responds to the non-synchronized random access attempt
`with timing information and resource allocation for scheduling request. UE then sends the
`scheduling request at the assigned time-frequency bin using the shared data channel.
`The two possibilities are illustrated in Figure 3 and Figure 4 respectively.
`
`
`UE
`
`Node-B
`
`Access Preamble
`
`Timing information
`
`Uplink Data Resource Allocation
`
`UL Data Transmission
`
`Figure 3. Option 1: Non-Synchronized Access (one part)
`
`
`
`
`
`3
`
`
`
`
`
`Figure 4. Option 2: Non-synchronized access (two part)
`
`
`
`
`9.1.2.1.2. Synchronized random access:
`
`The synchronized random access procedure is used when the UE uplink is time synchronized by the
`Node B. The purpose is for the UE to request resources for uplink data transmission.
`
`
`9.1.2.1.2.1 Time Frequency Structure
`
`Synchronized random access and data transmission are also time and/or frequency multiplexed as
`illustrated in Figure 5. The minimum bandwidth, BWRA, allocated for synchronized random access
`transmission is 1.25 MHz. For system bandwidths larger than 1.25 MHz, either the random access
`transmission uses a larger bandwidth, or multiple random access channels are defined. Multiple 1.25MHz
`random access channels might be useful for selecting a best block using frequency selective channel
`characteristics (TDD mode). The length of the synchronized random access burst should be restricted to
`one or two DFT-SOFDM symbols with a period of x sub-frames (e.g. x=2).
`
`4
`
`
`
`Can be used for other random- access
`channels or data transmission.
`
`Data transmission
`
`BWRA
`
`0.5 ms subframe
`
`1 DFT-SOFDM Symbol
`
`(Scheduled) Data transmission
`
`Random Access Preamble
`Figure 5. Synchronized Random Access
`
`
`
`
`9.1.2.1.2.2 Synchronized Random Access Procedure
`
`For synchronized random access, Figure 3 and Figure 4 also apply, except the timing information may not
`be transmitted.
`
`9.1.2.1.3. Preamble Design Principle
`
`The random access channel sequence(s) (e.g. based on CAZAC/GCL) used to generate the transmitted
`random access preamble waveforms should have the following properties:
`
`
`1. Good detection probability while maintaining low false alarm rate e.g. by maximizing post-
`decoder Es/(Nt+Ne) for a occupied random access channel preamble where Ne is the residual
`interference due to other random access channel transmissions in a given random access channel
`and Nt is thermal noise.
`a. cross correlation of the sequences that occupy the same frequency and same cyclic shift
`value impacts achievable Es/(Nt+Ne) and false alarm rate
`2. Number of random access channel preamble waveforms should be defined to handle the maximum
`expected multiple access scenarios (traffic load) while guaranteeing low collision probability.
`a. Subsets of preambles could be defined such that performance is improved at lighter loads
`(e.g., first use cyclic shifts of a single CAZAC/GCL sequence before using additional
`sequences)
`3. Enable accurate timing estimation (e.g. good autocorrelation properties and sufficient occupied
`BW).
`4. Low power de-rating (low CM/PAPR).
`5. Resistant to the Near-far problem.
`
`
`
`
`5
`
`
`
`End of Text Proposal
`End of Text Proposal
`
`
`
`
`6
`
`