HSUPA/EUL

HSUPA/Enhanced Uplink

Basics:

• EUL is based on some of the same basic procedures as for HSDPA: fast scheduling, fast Hybrid ARQ and short (2ms) TTI
• The shared resource for the users in the uplink can be seen as the maximum tolerable interference at the nodeB which is generated by the transmission powers of all active UE’s in the cell.
• Contrary to HSDPA, the data buffers and the scheduler are located in different nodes (buffers are in the UE’s) which means that the buffer status must be signaled to the scheduler in the nodeB.
• The scheduler can adjust the data rate of the E-DCH by a power offset which is relative to the associated power-controlled uplink control-channel. TPC’s (Transmit Power Control) is transmitted on the F-DPCH (Fractional Dedicated Physical Channel).
• Unlike HSDPA, EUL supports soft handover.
• SF2 is introduced in EUL since it leads to a lower PAR (Peak to Average Ratio).
• The reason why it's difficult to use higher modulation in HSUPA is the fact that it requires more energy per bit to be transmitted. Since the transmitter in the UL is the UE, the available UL Tx power is limited and less than in the DL transmitter, which is the nodeB.

Scheduling

• The scheduling in EUL is based on scheduling grants from the nodeB and scheduling requests from the UE’s.
• Scheduling grants can be sent with every TTI or slower.
• Due to the limited transmission power of a single UE compared to the nodeB, in most cases multiple users can transmit simultaneously.
• Inter cell interference is controlled by a down-scheduling grant (overload indicator) which is transmitted by the non-serving cells while the UE is in soft handover.
• Absolute grants are sent from the serving cell only on the E-AGCH ( E-Absolute Grant Channel) and may increase or decrease the grant in arbitrary steps.
• Absolute grant value ranges from 0-31 and are directly mapped to E-DPDCH/DPCCH power ratio.
• Absolute grants are valid for one UE, a group of UE’s or all UE’s using a primary identity (E-RNTI) which is UE specific or a secondary identity (E-RNTI) which is assigned to multiple UE’s. In this way the scheduler has more flexibility in controlling the overall cell load.
• Relative grants are sent from all involved nodeB’s and typically only adjusts the grant by one step up/down at a time. It is sent on the E-RGCH (E – Relative Grant Channel)
• Relative grants from the non-serving cells can only have to values: DTX or “Down” and is transmitted to all UE’s (aka the overload indicator)
• The UE maintains a Serving grant which is used in the E-TFC selection algorithm unless the HARQ process is in retransmission. In that case, the E-TFC should be unchanged.
• In order to adapt to the fast changing variations of the channel quality, the scheduling grant is only seen as an upper limit of the used resources (total interference) and the UE is free to do its own instantaneous adjustments to the transmission parameters.
• Out-band signaling consisting of scheduling information is used in the form of a “happy bit” which should be set if the UE has more data in its buffer to transmit and has more available transmission power.
• The basic information that the happy bit transmits is whether or not the UE can empty its buffers in N TTI's where the value of N is signaled by the NW.
• The terminal may only increase the bit rate if there is an UP from the serving HSUPA cell and no DOWN from any other cell in the E-DCH active set.

Scheduling information

• The scheduling information in the MAC-e pdu consists of:
  -total E-DCH buffer status
  -highest priority logical channel ID
  -highest priority logical channel buffer status
  -UPH
• UE transmission power headroom (UPH) is defined as being the ratio of the maximum UE transmission power and the corresponding DPCCH code power.
• UPH indicates the available power resources, or if and how much the UE can increase the bit rate.
• The nodeB scheduler may use this scheduling information in addition to the happy bit in the E-DPCCH

HARQ

• In soft handover, the HARQ protocol is terminated at multiple (UE’s) locations.
• As long as one ACK is received from any of the involved nodeB’s. The UE will consider the transmission as successful.
• For continuous transmission to a single UE, multiple HARQ processes can operate in parallel.
• The number of parallel processes depends on the TTI: 8 for 2ms TTI and 4 for 10ms TTI
• Individual HARQ processes can be de/activated in order to adjust the bit rate.
• To make the most efficient use of the HARQ features and to achieve maximum bit rates, it has been shown that a typical rate of retransmissions up to 10% should be targeted.
• Same reordering mechanism as for HSDPA is also necessary in EUL however, due to the multiple nodeB’s involved in soft handover, reordering takes place in the RNC.
• The E – HARQ Indicator Channel (E-HICH) is a dedicated physical channel that is transmitted from each cell in the active set and is used for sending the HARQ ACK/NACK’s.
• E-HICH from multiple users are transmitted on the same channelization code where each user has it’s own orthogonal signature sequence for identification. This ensures that channelization code space is not wasted at the same time as each ACK/NACK is transmitted with sufficient energy.
• E-HICH from non serving cells only contains ACK's
• UL-HARQ is fully synchronous which means that even transmitted redundancy versions can be predetermined.
• One HARQ profile consists of a power offset attribute and a maximum number of retransmissions.
• Incremental redundancy is supported with Chase combining as a subcase
• The node B transmits the responding ACK/NACK a well-defined time after receiving each transport block. In this way the UE knows to which HARQ-process the ACK/NACK belongs.
• In HSUPA/EUL, HARQ retransmissions are operated in synchronous and non-adaptive mode which means that retransmissions can only be made at pre-defined times after the initial transmission and that the same transport format is used.
• A 2 bit Retransmission Sequence Number  (RSN) is used to distinguish retransmissions from new data.

MAC

UE

• A new MAC entity, MAC-es/MAC-e is introduced in the UE which handles: HARQ protocol, scheduling and E_DCH Transport Format Combination (TFC )selection
• The UE can autonomously select the data rate by choosing the E-TFC( E-DCH Transport Format Combination)  as long as the resulting data rate does’nt exceed the scheduling grant.

NodeB

• A new Mac entity, MAC-e is added in the node B to handle HARQ, Scheduling and MAC-e demultiplexing.
• There is one MAC-e entity in each node B for a UE in soft handover.
• There is one E-DCH scheduler function in the node B

RNC

• For each UE there is one MAC-es entity in the SRNC which handles: Reordering queue distribution, Reordering and combining of data from different node B’s in case of soft handover.
• The MAC-es header includes the Transmission Sequence Number (TSN) necessary for reordering (in-sequence delivery) in the RNC.
• Since the RNC does not know when a PDU has been transmitted the maximum number of times, a stall avoidance timer is used. On its expiry, the packet delivery from the reordering entity to RLC is resumed and a RLC retransmission is triggered for the missing PDU.

Mobility

• E-DCH keeps its own active set of nodeB’s. However, most of the time it’s the same as the DCH active set.
• Serving cell is the same for both E-DCH and HS-DSCH and is changed in the same way as for HSDPA

Physical channels

• E-DCH maps to one or multiple E-DPDCH
• A separate and parallel code channel E-DPCCH is carrying all the necessary information to receive the E-DPDCH.
• The 10 information bits on E-DPCCH is carrying the following information:
  E-TFCI, RSN and happy bit
• From a performance perspective, coding is always better than spreading implying that the number of channelization codes (E-DPDCH’s) should be high and the spreading factor low.

RRM for HSUPA

• RNC: admission control, handover control, resource management, packet reordering, congestion control
• NodeB: resource management, packet scheduling, congestion control, HARQ
• UE: packet scheduling, power control, HARQ

• the HSUPA scheduler has instant information of uplink interference since it is located in the nodeB.
• the RNC can send a UE-specific congestion indicator to the nodeB