6 Functional description of the PSAP data modem

26.2673GPPeCall data transferGeneral descriptionIn-band modem solutionRelease 17TS

This clause describes the different functions of the PSAP data modem.

6.1 PSAP transmitter

The PSAP transmitter generates the signal sent on the downlink. This signal is required to control the transmission of the MSD message on the uplink. The different blocks of the PSAP transmitter are shown in Figure 15.

Figure 15: PSAP transmitter block diagram

6.1.1 Message encoding

The PSAP transmitter is designed to send up to 16 different link-layer feedback messages to the IVS. Three of them are used currently as follows:

1) START signal, i.e. the signal that triggers start of the IVS MSD transmission.

2) NACK, i.e. negative acknowledgement upon CRC check failure.

3) ACK, i.e. positive acknowledgement upon CRC check success.

A fourth link-layer message is defined for exclusive use in the higher-layer ACK message (see Table 3).

6.1.2 BCH encoding

The link-layer feedback message code is error protected by a shortened (60, 4) BCH block code which is derived from a (63, 7) BCH code. The messages and their encoded representations are shown in Table 3.

Table 3: Downlink message encoding

Link-layer feedback message	Binary representation	BCH encoder output, bi (hexadecimal)
START trigger	0000	A72 F298 41FA B376
CRC flag = 0 (NACK)	0001	4C4 1FD6 6ED2 7179
CRC flag = 1 (ACK)	0010	97A 8C41 FAB3 7693
reserved	0011	DBE 9397 9461 07EA
Not used	0100 to 1111

The binary representations of the link-layer feedback messages defined in Table 3 are re-used for the compressed higher-layer ACK (HL-ACK) message, in which four data bits (i.e., two of the binary link-layer message representations) are transmitted in a different feedback message format (see clause 6.1.4)

6.1.3 Modulation

The encoded binary data stream bits are grouped into symbols. Each symbol carries 4 bits of information and modulates one basic downlink waveform.

The duration of the downlink waveform is 4 ms or 32 samples at 8 kHz sampling rate. Therefore 5 modulation frames correspond in length to one speech frame. Each modulated waveform carries one symbol of 4 bits of binary information and the modulation data rate is 1 000 bits/s for the downlink transmitter (modulation data rate for FEC-encoded bits, not considering muting gaps and synchronization frame).

The basic downlink waveform is defined for n = 0,…,31 as follows:

p_DL(n) = (40, -200, 560, -991, -1400, 7636, 15000, 7636, -1400, -991, 560, -200, 40, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)

Table 4 describes the symbol modulation mapping between symbol and the downlink waveform. The downlink waveform is derived from the basic downlink waveform p_DL(n) by a cyclic right-shift by k samples, denoted by , and multiplication with a sign q.

Table 4: Symbol modulation mapping (downlink)

Symbol		Downlink waveform wDL(n) = (n= 0,…,31)
		sign q	cyclic shift k
0	0000	1	0
1	0001	1	4
2	0010	1	8
3	0011	1	12
4	0100	1	16
5	0101	1	20
6	0110	1	24
7	0111	1	28
8	1000	-1	28
9	1001	-1	24
10	1010	-1	20
11	1011	-1	16
12	1100	-1	12
13	1101	-1	8
14	1110	-1	4
15	1111	-1	0

Since there are only few messages for the downlink and the modulated waveform for each message is relatively short (480 samples), the modulated downlink waveforms are stored in ROM tables to save computation complexity at runtime. The downlink waveforms can be found in 3GPP TS 26.268 [2].

6.1.4 Downlink signal

6.1.4.1 Link-layer feedback messages

Every downlink message starts with a synchronization frame (as defined in clause 5.1.6) and continues with a feedback frame. For the link-layer control messages, the feedback frame consists of a single DL-Data field surrounded by muting periods as follows:

1) 3 frames of muting, M1 (60 ms).

2) 3 frames of modulated data, DL-Data (60 ms).

3) 1 frame of muting, M2 (20 ms).

Each DL-Data field includes one of the three types of link-layer messages in a block-encoded representation as described in clause 6.1.2.

6.1.4.2 Higher-layer acknowledgement messages

For the higher-layer acknowledgement messages, the synchronization frame defined in clause 5.1.6 is inverted (i.e., each sample is multiplied with -1). The feedback frame consists of two DL-Data fields preceded by a muting period as follows:

1) 1 frame of muting, M1 (20 ms).

2) 3 frames of modulated data, DL-Data 1 (60 ms).

3) 3 frames of modulated data, DL-Data 2 (60 ms).

Each DL-Data field includes one of the four types of block-encoded two-bit binary message representations as described in clause 6.1.2. The feedback frame for higher-layer acknowledgement messages therefore transports four information bits for use by the higher-layer application protocol (HLAP). These information bits can be used to satisfy the eCall requirements [1], e.g. to clear down the call.

6.1.4.3 Handling of downlink messages

The PSAP transmitter retransmits the START message multiple times until it has detected the uplink synchronization frame. Upon detection of the synchronization frame, the PSAP transmitter sends a series of NACK messages, until a successful CRC check of the MSD message is obtained. After successful MSD detection, the PSAP transmitter sends link-layer ACK messages and/or the higher-layer ACK (HL-ACK). This operation is illustrated in Figure 16.

Figure 16: Downlink signal format

If the PSAP transmitter fails to obtain uplink synchronization, it will not start transmitting NACK on the downlink. Instead, START messages are repeated. The IVS awaits the reception of a NACK after transmission of the first uplink synchronization frame. If instead repeated START messages are received, it interrupts the current MSD transmission attempt and starts with a new synchronization frame and MSD rv0. If the receiver is not able to decode the message successfully within 8 redundancy versions or if the Sync Tracker indicates that synchronization has been lost on the uplink, it asks the IVS to restart the transmission with a new synchronization frame and MSD rv0. This is achieved by switching from NACK to START messages.

The repetition of START messages by the PSAP (in case it does not obtain any uplink synchronization) continues until the PSAP modem is manually switched off by the operator. An automatic PSAP timeout mechanism should be added to the PSAP implementation if this behaviour is not desirable. The PSAP timeout mechanism is not part of this specification.

6.1.5 Synchronization

Synchronization signals for the PSAP transmitter are as described in clause 5.1.6, except that a value of 5000 is added to the PN sequence pulses (i.e., the resulting pulses have amplitudes of 25000 and -15000), and the original zeroes are replaced by samples with a value of 12000.

For higher-layer acknowledgement messages, the synchronization frame defined above is inverted (i.e., each sample is multiplied with -1).

6.1.6 Multiplexing

The multiplexer combines synchronization, muting and feedback frames to form the downlink signal for the PSAP transmitter.

6.2 PSAP receiver

The PSAP receiver demodulates the MSD message from the IVS and checks the integrity of the received MSD by evaluating the CRC field. The different blocks of the PSAP receiver are shown in Figure 17.

Figure 17: PSAP receiver block diagram

The PSAP receiver continuously monitors traffic from the speech decoder in idle or standby state. When the PSAP receiver is in idle state, speech from the speech decoder passes through as in normal voice call.

Once an eCall sync burst is detected, the PSAP transitions out of idle state and the speech path to audio out is muted.

6.2.1 Synchronization detector/tracker

Basically, the uplink synchronization detector/tracker works as described in clause 5.2.1. Some differences are given in the following.

The detection of a single synchronization preamble is sufficient to trigger the PSAP. A Sync Observer checks the received signal for another 10 speech frames after a preamble has been detected. This is to ensure that synchronization does not find a more reliable preamble, which would mean that the previous detection would have been a false detection (wrong delay value). In case synchronization does find a better preamble it restarts the reception of the MSD.

The Sync Check function continuously checks the validity of the identified delay value, based on the subsequently transmitted uplink synchronization fragments. If the delay value is found to be invalid, the Sync Tracker searches for a new valid delay value within a pre-defined search window. The maximum setting for the search window at the PSAP receiver is +/– 240 samples. If, after an invalid delay value, the Sync Tracker cannot identify a new valid delay value for a certain number of subsequent synchronization fragments (default value is four unsuccessful tracking attempts), it resets the PSAP transmitter in order to re-initiate the MSD transmission by sending START messages to the IVS.

A tone detector based on a DFT evaluation at the two reference frequencies is used to evaluate the frequency of the sync tone. If the frequency can be detected reliably, then this decision is taken to decide on which modulator is used for demodulation. If a reliable detection is not possible, the fast modulation scheme is chosen if it is the first time that a preamble has been detected successfully while receiving the current MSD. Otherwise, the robust demodulator is chosen.

6.2.2 Timing Unit

As described in clause 5.2.2.

6.2.3 De-multiplexer

As described in clause 5.2.3.

6.2.4 Data demodulator

The data demodulator is represented by a correlator matched to the modulated waveform applied by the data modulator of the IVS transmitter. Specifically, correlations for all possible symbols are calculated:

where n =15 for the fast modulation mode and n =31 for the robust modulation mode.

The correlation values are then normalized by their variance and subtracted by a mean value to be converted to soft symbols for the HARQ FEC decoder [2].

6.2.5 HARQ FEC decoder

The HARQ decoder combines the demodulated and deinterleaved data signal with previously transmitted redundancy versions. For this operation it applies a two stage rate matching scheme and performs turbo decoding of the combined soft-information.

To speed up the MSD reception in adverse transmission conditions, the HARQ decoding is performed for partially received messages, starting from the second redundancy version, rv1. Decoding is then attempted after reception of each of the three data parts of the MSD frame (see clause 5.1.5). The decoding attempts based on partial messages is beneficial since, in many cases, the correct MSD can be decoded already after the incremental redundancy contained in the first data part D1 of rv1. Figure 18 summarizes the decoder algorithm.

After MSD data bits are decoded, a descrambling operation as described in clause 5.1.2 applies.

Figure 18: Turbo decoder

6.2.6 CRC handling

This function performs a CRC check and outputs a CRC flag. The CRC flag triggers transmission of an ACK or NACK message by the PSAP transmitter.

6.2.7 Push mode – push message detector

In push mode the PSAP receiver starts monitoring the incoming signal immediately after the call has been established. For the push message (IVS initiation sequence) detection, it applies the same receiver as described in clause 5.2. A push command is considered detected if two correct sync preambles have been detected and a subsequent push message (see Table 2b) has been identified.