Future Communications Infrastructure - Technology Investigations Description of AMACS

Transcription

1 EUROPEAN ORGANISATION FOR THE SAFETY OF AIR NAVIGATION EUROCONTROL Future Communications Infrastructure - Technology Investigations Edition Number : 1.0 Edition Date : 02/07/07 Status : Issue Intended for : ACP-CG EUROPEAN AIR TRAFFIC MANAGEMENT PROGRAMME

2 DOCUMENT CHARACTERISTICS TITLE Future Communications Infrastructure - Technology Investigations : Document Identifier Edition Number: 1.0Error! Reference source not found. Edition Date: 02/07/07 Abstract using the Technology Assessment template. Keywords future communication technologies FCI evaluation Datalink COCR Terrestrial systems Satellite systems aeronautical communications shortlist Requirements spectrum AMSS/AM(R)S STATUS, AUDIENCE AND ACCESSIBILITY Status Intended for Accessible via Working Draft General Public Intranet Draft EATMP Stakeholders Extranet Proposed Issue Restricted Audience Internet Released Issue ELECTRONIC SOURCE Path: P:\EATM\DAS\BD_CSM\CMU\FUTURE_COM\Technology investigations\step 2\AMACS\deliverables Host System Software Size Windows_NT Microsoft Word Kb Page ii Error! Reference source not found. Edition Number: 1.0

3 DOCUMENT APPROVAL The following table identifies all management authorities who have successively approved the present issue of this document. AUTHORITY NAME AND SIGNATURE DATE Edition Number: 1.0 Error! Reference source not found. Page iii

4 DOCUMENT CHANGE RECORD The following table records the complete history of the successive editions of the present document. EDITION NUMBER EDITION DATE INFOCENTRE REFERENCE REASON FOR CHANGE PAGES AFFECTED Page iv Error! Reference source not found. Edition Number: 1.0

5 CONTENTS 1. L-BAND DATALINK TECHNICAL DESCRIPTION Overview Key facts AMACS Functional Architecture General AMACS Network Architecture Services Provided & Key Features Data exchange services Supplemental services Timing concept Air Interface Description: PHY, MAC, Data-link & Network Sub-layers Physical & MAC Sub-layers Data-link Sub-layer Network Sub-layer Services offered by the network layer: Standards Technology Readiness Level (TRL) APPLICATION OF TECHNOLOGY TO ATM Concept of operation: cellular deployment Introduction Representative C/I derivation and Cellular deployment Applicable Frequency Band and electromagnetic compatibility Airspace Application ATM services supported Proposed Architecture for Technology System Avionics Range Link Budget Performance Assurance STATUS OF THE TECHNOLOGY Summary Status ANNEX A : MODULATION OPTIONS Edition Number: 1.0 Error! Reference source not found. Page v

6 A.1 Introduction A.2 Modulation design choice A.3 Practical modem implementation considerations ANNEX B : ERROR CORRECTION CODES B.1 Introduction B.2 Inner code properties B.3 Outer code properties B.4 Interleaving properties ANNEX C : ESTIMATION OF A LINK BUDGET ANNEX D : MESSAGE STRUCTURES D.1 Cell insertion message: CELL_INS D.2 Cell insertion reply message: GS_ALLOC D.3 Cell exit message: CELL_EXIT D.4 Cell exit reply message: EXIT_ACK D.5 Block reservation message: GS_BLOCK D.6 Framing message: GS_FRAME D.7 GS data uplink message: GS_DATA D.8 CoS1 keep-alive message: KEEP_ALIVE D.9 CoS1 data downlink message: DATA_COS D.10 ACK/CTS message to all aircraft: CTS_ACK_ALL D.11 GS ACK uplink message: GS_ACK D.12 CoS2 downlink message: DATA_COS D.13 A/C ACK downlink message: AC_ACK D.14 CoS2 random access short data message: DATA_RA_COS2_SHORT D.15 CoS2 random access RTS message: RTS_COS D.16 CoS2 random access CTS message: CTS_COS D.17 CoS2 random access long data message: DATA_RA_COS2_LONG ANNEX E : SYSTEM OPERATIONS Page vi Error! Reference source not found. Edition Number: 1.0

7 E.1 A/C cell insertion E.2 A/C has data to send E.3 A/C has no data to send E.4 CoS2 random access E.5 A/C-initiated cell exit E.6 GS request for A/C cell exit E.7 GS has data to send E.8 GS framing message E.9 GS changes the section sizes in the frame REFERENCES: Edition Number: 1.0 Error! Reference source not found. Page vii

8

9 1. L-BAND DATALINK TECHNICAL DESCRIPTION This document contains a description of the L-Band Datalink system called AMACS (Allpurpose Multi-channel Aviation Communication System). This document relies on a physical layer of the same family as the UAT standards and builds on the two E-TDMA studies performed by Sofréavia for the DSNA. Some elements have already been described in the ITT Phase 1 report [27] (Technology Assessment for the Future Aeronautical Communication System). See also the report on future communications prepared for LFV by Helios Technology [31]. 1.1 Overview AMACS is a multipurpose communication system, with cellular narrowband ( khz), operating in the MHz frequency allocation designed for flexible deployment. Its key design drivers are flexibility, scalability and robustness. E-TDMA and XDL4 concepts have been merged to provide an adapted technical solution for the data-link communications needs of The AMACS concept is intended to provide a data-only service with significant requirements for QoS for air/ground point to point, air/air point to point and broadcast modes. Its flexible slot structure is adaptable to meet local requirements. It can support different channel bandwidths and bit rates to cope with the various operational needs and traffic densities foreseen for Europe in the future. Its robust physical layer is based on the GSM/UAT modulation types associated with strong data coding, for achieving the highest QoS in terms of latency. A multi-level QoS system is proposed to permit use of channel resources according to the QoS level required. Specific channel slots are reserved for high QoS transmissions. The efficient handling of QoS is based on the TDMA structured MAC layer and gives a guaranteed transmission delay. These communications can support ATN or IPv6 networks. Common Signalling Channels (CSC), similar to those employed by VDL Mode 4, are proposed to maintain QoS levels during intervals of network degradation. Examples are the warm- and cold-start features if a ground station (or stations) should go off-line for any reason. CSCs would serve to broadcast new ground station frequencies to alert aircraft mobiles of the new channels to which they should tune. The AMACS frame length is designed for fast delivery of time-critical messages and has been set at 2 seconds, but could be adapted to a lower duration if necessary Key facts The AMACS concept is based on several fundamental performance requirements. These include: 1) A very robust physical layer using the already-validated modulation family (CPFSK) used by GSM or UAT. 2) Contributions to data integrity and certification goals through careful, fast, error detection and correction mechanisms. Edition Number: 1.0 Error! Reference source not found. Page 1

10 3) Use of modular error correction where a unique FEC code is used for headers and for short and long slot data. 4) A high-integrity deterministic MAC sublayer employing deterministic slot scheduling and potentially Statistical Self-Synchronization (S3) and deterministic slot scheduling for remote area applications (without ground stations). 5) For ranging functions, fairly imprecise positioning performance may be adequate. 6) High throughput using low overhead for headers, FEC, and transmitter ramping. AMACS is designed to simultaneously handle up to 175 aircraft per cell in high-density airspace. It has an efficient air-initiated cell handover mechanism, which uses aircraft knowledge of cell locations and characteristics through on-board databases, Electronic Flight Bags (EFB) or a Common Signalling Channel (CSC). Its initial deployment will be in the lower L-band for new ATM point-to-point services requiring a high QoS, thus giving support to SESAR or NEXTGEN future concept. Broadcast services will be provided in a segregated channel if the spectrum availability in the lower L-band is sufficient. Air-to-air data communication is also provided in other segregated channels. It is expected that AOC data communications can be achieved if the necessary extra spectrum is available. 1.2 AMACS Functional Architecture General A depiction of the AMACS functional architecture is shown in Figure 1.1. Figure 1.1 Functional Architecture of AMACS The aircraft's communications system is linked to the ground station for the cell in which it is located. This ground station will provide all of the communication services required, such as Page 2 Error! Reference source not found. Edition Number: 1.0

11 time synchronization. The aircraft will only need to initiate contact with other ground stations before it leaves the current cell. The information about cell parameters, ground station locations and frequencies necessary for the aircraft to initiate and maintain contact may be available through an on-board database (updated before the start of the flight), the EFB or CSC. Existing GSM/UAT radio technology will be used. Therefore new hardware radio development will be limited AMACS Network Architecture The expected network architecture in support of AMACS is depicted in Figure 1.2. Cluster 1 GNI ATN A/G Router ATN G/G Router ATN Applications GNI Cluster 2 IPv6 Router ATN A/G Router WAN GNI Cluster 3 IPv6 Router IP Router TCP/IP Applications Figure 1.2 AMACS network architecture The basic concept is that the ground AMACS infrastructure comprises a number of AMACS Ground Radio Stations, which are organized into clusters. Typically, the Ground Radio Stations in a cluster will be geographically adjacent, or may have overlapping areas of coverage (using different frequencies). Each Ground Radio Station in a cluster will be connected to some redundant concentrator, the Ground Network Interface (GNI), which interfaces it to the transport network via an ATN Air/Ground Router or to other types of router (e.g. an IPv6 Router). The Air/Ground Routers supporting each cluster will themselves be interconnected by a ground transport network, which will also support Ground/Ground Routers for interconnection with end-users. Edition Number: 1.0 Error! Reference source not found. Page 3

12 From this description, the ATN A/G Routers and the IPv6 Routers are ground-based users of the AMACS sub-network service and the airborne ATN and IP routers are mobile users of the AMACS sub-network service. 1.3 Services Provided & Key Features Data exchange services The AMACS system provides services for reliable data transfer, ensuring delivery on a perframe basis: the provision of an acknowledged connectionless service is expected to be sufficient in the context of ATN or TCP/IP communications where end-to-end connection will be ensured at transport level. AMACS has a strong robustness at physical layer level to ensure both the highest QoS in terms of latency and predictive behaviour in a typical distorted propagation channel. The communication service which is provided by AMACS addresses unicast destinations as well as multicast diffusion. AMACS will support the following specific communication types: air-to-ground point-to-point; ground-to-air point-to-point; air-to-air point-to-point; ground-to-air broadcast; air-to-air broadcast. The services that could be supported by AMACS are shown in Figure 1.3. Autonomous mobile broadcast Air-to-air point-to-point Mobile broadcast A/G pointto-point communication services Ground station broadcast Ground network Figure 1.3 Possible uses of AMACS Page 4 Error! Reference source not found. Edition Number: 1.0

13 AMACS is designed to be flexible and configurable, for use for h point-to-point and broadcast communications. The aircraft can use AMACS to communicate with each other as well as with the ground station (using the appropriate channels), and the ground station can selectively communicate with individual or all aircraft. Separate channels are used for point-to-point and broadcast communications Supplemental services The AMACS high-level supplemental services include: User/radio authentication; Data Compression: The AMACS data suite will include capabilities for both IP or ATN header and user data compression; Mobility management: The AMACS data suite will handle the mobility of the aircraft in order to maintain communications while they move, using handover algorithms; QoS management: The AMACS data suite will handle different classes of traffic (CoS1 and CoS2), integrity, priority and a guaranteed minimum QoS Timing concept UTC is the common time reference in AMACS, in support of the TDMA structure. A hierarchy of timing schemes will be defined in order to guarantee data exchange capability in situations where the the quality of time from available time sources is degraded. These schemes include range measurement functions that also can be used to support security functions in the lower communication layers. The timing accuracy needs varies depending on the actual type of communication. Air-to-air communications are in general more demanding in this respect than air-to-ground communications. The timing scheme for AMACS will be based upon the timing scheme developed for VDL Mode 4 1. The results of previous studies conducted to investigate the timing requirements for VDL Mode 4 will be taken into account [26, 32]. 1.4 Air Interface Description: PHY, MAC, Data-link & Network Sub-layers Physical & MAC Sub-layers Introduction The aim is to re-use where appropriate the physical layer specifications of the UAT/GSM systems, thus affording considerable advantage in the costs associated with the development, standardization and fielding of the technology. The derivation of the necessary physical layer properties is explained below. The system performance is designed to provide a required Residual Message Error Rate (MER) of 10-7 on the basis of a Physical Bit Error Rate of ICAO Manual [15] Part II clause and ICAO Manual Part I clause Edition Number: 1.0 Error! Reference source not found. Page 5

14 The target net data rate is about 500 kbps in order to accommodate the most demanding communications load requirement in high density airspace (ref. COCR v2, [25]). These are summarized in Table 1.1 and Table 1.2 for addressed and broadcast services in high density airspace. Airport Surface TMA Enroute Oceanic ATS AOC Combined Table 1.1 COCR addressed communications load (kbps) for combined uplink and downlink in High Density volumes Airport Surface TMA Enroute Oceanic C&P SURV ITP SURV M&S SURV SURV TIS-B Table 1.2 COCR broadcast communications load (kbps) in post 2020 The figures in Table 1.2 are for individual services the aggregate loading figure is close to 500 kbps. Previous work on E-TDMA has led to the definition of framing and error correcting codes. Much of this remains applicable to the design requirements of the AMACS system Modulation Introduction Although AMACS makes use of UAT and VDL Mode 4 characteristics, it is not essential for AMACS to use the same modulation schemes. A balance has to be struck between the bit rate, the Bit-Error Rate (BER), the Signal-to-Noise Ration (SNR), the bandwidth and power. A description of possible modulation schemes and an analysis of the proposals which could be used for AMACS is presented in Annex A Modulation design choice Given the fundamental principles of GMSK modulation shown in Annex A, three proposals were drawn up taking into consideration the desirable characteristics, design goals, and the spectral environment that define the theatre of operations for AMACS. Page 6 Error! Reference source not found. Edition Number: 1.0

15 In light of the observations, the known spectral constraints in the L-band, and cost advantages from reusing mass-market standards, the GMSK modulation proposal represents the best design compromise and is the design choice for the AMACS physical specifications. Chosen proposal : GMSK : h = 0.5 & BT = 0.3 Gross bit rate : ~ 540 kbps Channel bandwidth : 400 khz Expected C/I : ~ 9dB The use of concatenated error coding is considered in Annex B, to make the objective of a C/I of 9 db in co-channel interference attainable Introduction: Point-to-Point Air Interface The AMACS system makes use of a specific channel for point-to-point communications. This channel is designed to allow stations (air and ground) to have a minimum number of exclusive bits per slot for regular or high-qos transmissions, with more bits available on request. The broadcast air interface description is given in Section Impact of the physical layer features on airborne co-site issues One of the key problems for a future communication component operating in the L-band ( MHz) is co-siting with other radio transmitters that operate in the same frequency band. Even if a frequency separation is implemented, providing some decoupling with other spurious emissions, the robustness of the new communication link will be drastically affected by the proximity of other pulse transmitters on the same aircraft. Therefore the best solution will be to take advantage of Pulse Blanking Techniques that have been used in many cases to reduce the effect of strong interference (that is, the case on board aircraft due to very small system isolation). Such a pulse blanking mechanism has been defined in the UAT standards and has a common bus interconnecting the avionics elements that could benefit from the information provided (pulse blanking signal whenever a transmitter is on). Taking into account that two of the major interferers for the new communication system will be short pulsing transmitters (DME and SSR/Mode S), the duration of the jamming pulses will be equivalent to or lower than the AMACS bit duration: the impact of the interference will be therefore limited to a few bits in the frame for which data coding (presented in Section ) will be the appropriate answer to mitigate the impact of the interference on the frame error rate. On the other hand the impact of AMACS onboard implementation on DME or SSR/Mode S will be limited by providing a frequency separation between the AMACS channel and the first DME receiving channel (i.e. 978 MHz) and by taking into account the small duty cycle of AMACS (0.15% in average on the basis of a 3ms usable slot duration). Edition Number: 1.0 Error! Reference source not found. Page 7

16 Power control option In order to reduce the level of interference for point-to-point, a power control option is possible. This can be done rather easily, using a small capacity in the signalling channels (afforded by the high capacity offered by the GMSK modulation option). It requires the ground base station to perform a continuous measurement of the received signals from each aircraft and then return this information to each aircraft. On reception of this information, the aircraft terminal uses it to feed a power control algorithm, which is an adaptive algorithm that converges to the optimum power, i.e. the power which is required for normal operations and acceptable BER. Advantage can be taken from algorithms developed for GSM Access and frame structure For point-to-point channels, AMACS will use the MAC layer principles developed for E- TDMA. It will have deterministic organization and a deterministic access to the medium. AMACS will have a frame repeating every 2 seconds, with specific 'uplink' and 'downlink' sections. The frames are presented in Figure 1.4. frame (N-1) AMACS cycle frame (N) AMACS cycle frame (N+1) AMACS cycle Figure 1.4 AMACS Frames A frame consists of multiple slots and a slot consists of one burst. A frame is composed of successive time slots that each consists of: transmitter ramp-up, synchronization interval, flags and addresses, the data burst, FEC/CRC code bits, transmitter ramp-down, propagation guard time. This slot structure is depicted in Figure 1.5. Page 8 Error! Reference source not found. Edition Number: 1.0

17 individual slot structure Ramp-up Synch signalling and data active slot duration FEC/ CRC decay Guard time depending on cell size next slot total slot duration 4ms Figure 1.5 AMACS slot structure If a ground station or an aircraft is using several slots for one transmission, then the transmitted ramp-up time and synchronization interval will only be present at the start of the initial slot, and the transmitter ramp-down time and propagation guard time will only be present at the end of the last slot (the number and position of the FEC/CRC code bits will be dependent on the size of the transmission). This means that the size of the combined signalling and data bursts will be larger than the sum of signalling and data burst sections from separate, individual slots (the extra bit transmission capacity is of the order of 540 bits). This is shown in Figure 1.6. merged slot structure ramp-up synch signalling and data FEC/ CRC signalling and data FEC, CRC and decay guard time depending on cell size next slot total slot duration total slot duration Figure 1.6 AMACS merged slot structure Figure 1.7 shows a different view of the slot structure, showing the actual number of bits allocated to each section. It is based on a frame size of 2 seconds, a data rate of 540 kbps and a slot size of 4ms. Edition Number: 1.0 Error! Reference source not found. Page 9

18 4 ms Ramp up Flag Addresses plus flags User data n 1 octet 4 5 octets (typical) 148 octets Reservation header FEC / CRC Flag Guard time Ramp down 3 octets (if required) 47 octets 1 octet 0 9 ms m Figure 1.7 AMACS slot structure sizes The combined ramp-up and ramp-down time (m+n) is less than 0 1 ms. Slot characteristics: Active slot length: 4 ms (ramp + guard times) = 3 ms Bits per slot: Active slot length Bit rate = 1,620 bits Bits for FEC/CRC: ~30% of bits per slot = 376 bits Remainder: Bits per slot CRC = 1,244 bits = octets ISO flags + reservation header = 3 octets Addresses plus administrative flags (typical) = 4 5 octets User data space = 148 octets The use of the uplink sections in the frame is configurable (dynamically) by the ground station. These sections are ground-reserved areas for uplinks and ground-directed signalling. The two downlink sections are separated for different Classes of Service (CoS). The first one (CoS1) is intended for a high QoS and each aircraft is allocated one exclusive downlink slot in CoS1 for high QoS messages. More downlink slots are available on request in the lower QoS section (CoS2). The slots are allocated based on QoS requirements, and may be based on the application or may be functionally grouped. Figure 1.8 illustrates this concept. Page 10 Error! Reference source not found. Edition Number: 1.0

19 Start of UTC second Frame UP1 CoS1 UP2 CoS2 Framing message Cell insertion Reserved slots for uplink messages Exclusive primary slots for short, high QoS messages or RTS messages Second uplink for ACKs, CTS, reservations Shared slots, reserved or random access: used for any messages Uplink section Downlink section Uplink section Shared section Figure 1.8 AMACS Frame Structure Frame section usage Introduction In Figure 1.8, the CoS levels indicate service delivery levels. For the highest level, a dedicated time slot is reserved in CoS1 for each aircraft in the cell, and transit times and minimum throughput rates are guaranteed. The use of deterministic slot assignments is important for QoS performance. For the lower-level CoS2 time slots, the time guarantees are smaller since these slots are potentially shared among many aircraft and time guarantees are measured statistically. Specific uplink slots are reserved in each frame for the ground station framing message and for cell insertion messages. It is expected that the messages required for hand-off procedures will normally be exchanged in the UP1 and CoS1 sections, thereby taking place within one frame. The section lengths are not fixed and can be optimized by the ground station. The ground station will broadcast a framing message to all aircraft within the cell to indicate section length changes. Only the slot size and the overall frame size are fixed. If the ground station intends to change the frame section sizes, the aircraft will be notified a long time in advance (typically up to a few minutes). Edition Number: 1.0 Error! Reference source not found. Page 11

20 Proposed messages required for AMACS are provided in Annex D. These structures indicate the different message fields and the number of bits required for each message field. All except the longest messages will fit into the single slot shown in Figure 1.7. System operation exchanges are provided in Annex E, indicating how the aircraft and the ground station interact using the proposed messages Message identifiers and acknowledgements The 'message identifier' fields are used in addition to the message type fields so that stations can be certain of which of their transmissions have been acknowledged. (For example, a long data message could be transmitted in several parts, and if one part is not correctly received it would be inefficient to have to retransmit the whole message) The identifier acts as a rolling sequence number, which can have values from 1 to 64. This range is acceptable because it is not necessary for every message to have a unique ID, merely for the messages from a station to be distinguishable within a period of time. The receiving station will include the message identifier in its acknowledgements so that they show which of the individual messages have been received; the message type field on its own would not be sufficient Insertion mechanism for an aircraft entering the cell An aircraft entering the cell will know (from on-board information) the frequency of the ground station in the new cell. It will listen on this frequency for the framing message transmitted by the ground station, which contains information about the slot structure, and will then announce its presence to the ground station by transmitting a message in one of the dedicated cell insertion slots. The ground station will reply in UP1 in the following frame, telling the aircraft the position of its allocated high-qos slot in the CoS1 section. It will also give the aircraft a local address, used in the cell for identification instead of the longer 27-bit ICAO address. It is expected that the aircraft will be able to transmit in its allocated CoS1 slot very soon after reaching the new cell (as a framing message is transmitted by the ground station every 2 seconds) Reservation mechanism for downlink In order to reserve a slot or a series of slots in the CoS2 section of the frame, the aircraft includes a reservation request for CoS2 slots in its CoS1 slot transmission. These CoS2 slots are likely to be required when the aircraft has a large amount of data to transmit which cannot all be fitted into the CoS1 slot. A reservation flag (RTS) is set in the CoS1 slot transmission by the requesting aircraft and notice is implicitly provided to all members of the channel that future timeslots are requested by that aircraft. A reservation echo (CTS) is transmitted by the ground station, acknowledging and granting the request for time slots within the pool of secondary slots available in CoS2. Page 12 Error! Reference source not found. Edition Number: 1.0

21 An aircraft may also transmit a slot reservation request (RTS) in CoS2 if it has further information to send to the ground station in the same frame. The ground station will reply in the first available slot, indicating whether or not any slots are available for the aircraft to use. If slots are available, the ground station will transmit a CTS identifying the available slots. If none are free then the aircraft will have to transmit in the next frame. In order to prevent conflicting transmissions, all aircraft will listen to all CTS transmissions to record in their reservation tables which slots have been reserved Uplink transmissions The ground station has two blocks in each frame, UP1 and UP2, which are reserved for uplink transmissions. These are used by the ground station for normal data transmissions to aircraft, acknowledgements to downlink messages, CTS messages and cell insertion and exit exchanges. The slots in the uplink section are concatenated and do not require separate ramp-up and ramp-down nor guard-time in between messages. Consequently the number of bits available for data transmission within each slot is greater. If the ground station requires more slots for uplink transmissions, it will examine the reservation table for CoS2 and then broadcast a block reservation message to all aircraft. This message will indicate the start and number of slots of the reservation, which will only apply to the current frame. No aircraft will transmit in this block Hand-off mechanism for an aircraft leaving the cell There are two possible means of hand-off: controlled and uncontrolled Controlled hand-off Controlled hand-offs can be air-initiated or ground-requested air-initiated. An aircraft will know, from on-board information, when it is nearing the edge of the current cell. At an appropriate time, it will transmit a "cell exit" message in its dedicated CoS1 slot. The ground station will reply, in a UP2 slot, to confirm that the aircraft is leaving the cell and to indicate the number of a reserved CoS2 slot. As the "cell exit" message is being transmitted, the aircraft will also be searching for the ground station of the next cell, to ensure continuity of communications. When the aircraft receives the "exit confirmation" message, it will send a normal ACK message in this CoS2 slot, indicating that it is the "exit confirmation" message which is being acknowledged (this will be the last message that the aircraft sends to the ground station). When the ground station receives the ACK message from the aircraft, it will de-allocate the aircraft's CoS1 slot and will consider the link to be terminated. Edition Number: 1.0 Error! Reference source not found. Page 13

22 The ground-requested procedure is very similar the ground station will know, from the location information in ADS-B transmissions, when an aircraft is nearing the edge of a cell. If the ground station determines that a hand-off is appropriate, it will transmit a "cell exit" message to the aircraft. The aircraft, on receiving it, will attempt communication with the ground station in the next cell. If this is successful and the aircraft is allocated a CoS1 slot in the next cell, it will reply, in its current CoS1 slot, with an "exit confirmation" message. Then the current ground station will de-allocate the aircraft's CoS1 slot and will consider the link to be terminated. Otherwise the aircraft will maintain its communication with the current ground station. The hand-off from the current ground station will not be completed before the aircraft has made contact with the next cell Uncontrolled hand-off If communication between the aircraft and the ground station is lost, for more than a predetermined time period, then both the aircraft and the ground station will consider their link to be terminated. 1. The ground station will de-allocate the aircraft's CoS1 slot as it will assume that it is no longer required by the aircraft. 2. The aircraft will determine the appropriate (new) ground station to contact and will begin the "cell insertion" procedure on the new frequency. For this process to occur correctly, the appropriate value for the time-out period must be chosen (as it may be affected by local factors) Broadcast Air Interface The AMACS system makes use of a specific channel for broadcast communications. The AMACS broadcast channel has the same MAC structure as VDL Mode 4, modified for a single channel and for the AMACS frame structure. One superframe is 60 seconds long and contains 15,000 slots, with a corresponding slot length of 4 ms. There is an increase in the allowable basic message size, making AMACS more convenient for ADS-B. Each superframe starts with a short ground-quarantine section. The remaining slots are available to all stations (air and ground) by random access, using modified VDL Mode 4 reservation protocols. Most VDL Mode 4 broadcast protocols will be used, but no point-to-point transmissions will be permitted on the broadcast channel. Therefore some modifications will be required Data-link Sub-layer Segmentation and de-segmentation Page 14 Error! Reference source not found. Edition Number: 1.0

23 The data-link sub-layer will handle the segmentation of user data queued for transmission by higher layers into appropriate blocks for the MAC layer and the de-segmentation (reassembly) of received blocks from the MAC layer into a single user data packet for the upper layer Transmission The overall size of the message will already be known. The user data will need segmentation into blocks if the message size exceeds the maximum for one slot, which is 145 octets. Each additional block can contain up to 208 octets of data. These blocks shall carry sequencing markers to indicate both the total number of blocks and each block's place in the sequence Reception On reception of user data blocks from the MAC layer, the data-link sub-layer will know (from the sequencing numbers) how many blocks to expect. The re-assembly of the blocks will be done by using the sequencing numbers; the sub-layer shall know that the first block contains 145 octets of the user data but that the other blocks may contain up to 208 octets Error correcting scheme and interleaving A key factor of AMACS is provision of deterministic access to the radio channel to cover the stringent latency requirements of future data-link services. In order to achieve this goal it is not only necessary to provide the proper mechanism at the MAC layer to ensure deterministic access to each frame to any aircraft logged in a cell but also to ensure that the probability of message rejection due to data corruption is kept very low. This is only achievable through the use of specific data coding ensuring a high level of error corrections: this data coding will be therefore dependent on the QoS associated to the data (e.g. the useful data throughput will be lower for high QoS due to the level of associated data coding. Different error correcting schemes are used in data communications : cyclic block codes, convolutional codes, turbo code and low density parity check code. The focus has been on cyclic bloc codes. Annex B provides the reasoning behind the choice of the Reed-Solomon (RS) and Bose, Chaudhuri and Hocquenghem (BCH) codes. Edition Number: 1.0 Error! Reference source not found. Page 15

24 1.4.3 Network Sub-layer Figure 1.9 and Figure 1.10 present the end-to-end connectivity architecture involving the AMACS subnetwork. ATN Stack 1 ATN Stack ATN ground SNDCF WAN ATN Ground SNDCF WAN ATN AMACS SNDCF WAN 3 WAN 2 GNI AMACS datalink AMACS Physical 4 ATN AMACS SNDCF AMACS datalink AMACS Physical Ground router Air-Ground router AMACS Ground Station Airborne router Figure 1.9 Protocols stack model for ATN TCP/IP Stack 1 TCP/IP Stack IP ground SNDCF WAN IP Ground SNDCF WAN IP Mobile AMACS- SNDCF WAN 3 WAN 2 GNI AMACS datalink AMACS Physical 4 IP Mobile AMACS SNDCF AMACS datalink AMACS Physical Ground router Air-Ground router AMACS Ground Station Airborne router Figure 1.10 Protocols stack model for TCP/IP These figures show the main associations: 1) Association between end users: In the context of both ATN and TCP/IP communications, end-to-end connections are provided at the Transport level. Page 16 Error! Reference source not found. Edition Number: 1.0

25 2) Association between AMACS network service users: In the context of both ATN and TCP/IP communications, channels are established between an air-ground router and an airborne router. These channels are established in reference to the QoS. Compression mechanisms will be set up (e.g. Deflate). 3) Association between the GNI and the Air-Ground router: The GNI will report aircraft connectivity to the Air-Ground router. This will be performed through use of Join and Leave messages. This interface will permit to handle the AMACS QoS management as defined in ) Transfer of data between an airborne AMACS system and a GNI though an AMACS Ground Radio Station, using the AMACS medium access protocol Services offered by the network layer: Point-to-point air-ground Data transfer In the Airborne and Air/Ground routers, the data communication service is based on a dedicated AMACS SNDCF (which could be derived from the ATN Frame Mode SNDCF). The following data transfer services are provided: SN-Unitdata.request SN-Unitdata.Indication Two mobility management information events will be raised by the datalink layer and provided to the airborne router and the Air/Ground router: Join event, Leave event. These two events will be used as defined in Section Following a join event, packets may be up-linked to the aircraft from the Air/Ground router or down-linked from the aircraft to the Air/Ground router respectively, using the SN- UnitData.Request. The Unidata.Request will support the following parameters: Source address, Destination address, Data to be sent, QoS category, Class of Service, Priority, Integrity. Edition Number: 1.0 Error! Reference source not found. Page 17

26 The Source and Destination addresses will uniquely identify the Air/Ground router or the aircraft on the AMACS network. The ICAO 27-bit address will be used for the aircraft. The Air/Ground router will have a unique address coded over 3 octets. The maximum data size of each SN-Unitdata.request will be 145 octets. The values allowed for QoS category, Class of Service, Priority and Integrity parameters are defined in Section For uplink data transmission, the information associated with the Unidata.Request will be sent though a reliable transport network to the GNI from which a join event has been received for the destination address aircraft. The GNI will then handle the received parameters to send the requested information to the destination aircraft, using the required QoS. Handling of the requested QoS parameters is described in Section The data part of the PDU sent uplink by the GNI will consist of: The local identification of the Air/Ground router; The user data part. The local identification of the Air/Ground router corresponds to a local identifier (from the GNI point of view) that permits unique identification of the Air/Ground router among all the Air/Ground routers connected to the GNI. This identifier has been allocated locally by the GNI and was provided in the Join event triggered by the aircraft. It will permit identification of the data flows exchanged between the Air/Ground router and the airborne router. The GNI will use the received Source address information to find the local identification value to be used. The Destination address will be directly mapped to the corresponding AMACS datalink address. The GNI will implement message queuing mechanisms in order to temporarily store uplink transmission requests until they have been fully completed, i.e. sent and acknowledged at the AMACS MAC layer. In the case of failure to transmit the requested information, the GNI will react according to the requested QoS: If a high QoS has been requested (e.g. Guaranteed QoS requested with Deterministic CoS), it will report the transmission error to the A/G router, and the corresponding data packet will be discarded. If a medium QoS has been requested (e.g. Deterministic CoS), re-transmission will be attempted as soon as possible, taking into account potential concurrent requests with higher priority requests arriving in the meantime. The maximum permitted number of retries will be configurable. In case of failure after this number of retries, a transmission error will be reported to the A/G router, and the corresponding data packet will be discarded. If a low QoS has been requested (e.g. Concurrent CoS), re-transmission will be attempted as soon as possible, taking into account potential concurrent requests with higher priority requests arriving in the meantime. The maximum permitted number of retries will be configurable. In case of failure after this number of retries, the data packet will be discarded. Page 18 Error! Reference source not found. Edition Number: 1.0

27 When an aircraft receives an uplink PDU, the following information will be provided to the user through the SN-Unitdata.Indication service: Source address, Destination address, User data part. The Destination address will correspond to the sub-network address of the airborne router. It will be locally inserted. The Source address will correspond to the sub-network address of the Air/Ground router. It will be translated from the local identification received in the PDU. For downlink data transmission, the aircraft will send the data part to the GNI through the AMACS Ground Radio Station of the current cell within which it has been inserted. The transmitted PDU sent downlink will consist of: The local identification of the Air/Ground router The user data part. The local identification of the Air/Ground router will be selected, according to the requested Destination address, from the list of available connectivity with Air/Ground routers (notified locally through Join events). Handling of requested QoS parameters will be similar to the GNI behaviour for uplink transmissions, although queuing mechanisms have to deal with a unique sender. When a GNI receives a downlink PDU, it will identify the Air/Ground router to which it will forward the information through the reliable ground transport network. The GNI will translate the received local identification value to the real sub-network address of the Air/Ground router. Based on the datalink interface, it will also identify the sub-network address of the Source aircraft. The Air/Ground router will then provide the following information to the user through the SN-Unitdata.Indication service: Source address, Destination address, User data part. The Destination address will correspond to the sub-network address of the Air/Ground router. The Source address will correspond to the sub-network address of the airborne router. As the GNI may be inter-connected with several Air/Ground routers, through several AMACS Ground Radio Stations, it will be able to handle concurrent transmission requests coming from and going to different network entities Mobility management This section presents the mechanisms that are used for mobility management, taking into account the above AMACS architecture and protocols stack model. Edition Number: 1.0 Error! Reference source not found. Page 19

28 The aircraft AMACS Radio will use geometry information to identify the AMACS cell to which it should request insertion. The monitoring of the associated frequency and the description of the AMACS frame for this cell will allow the AMACS Radio to retrieve information regarding the presence of a ground station and its address, which will be encoded in one slot. When an aircraft first joins a 'user group' (i.e. an AMACS cell), this will result in a Join event being sent to the network user for both the ground A/G router and the airborne router. On the ground side, this will notify the service user that the identified aircraft has joined a 'user group' via the identified Ground Network Interface (GNI). On the air side, this will notify the service user of the GNI address as well as the supported A/G router(s). As the aircraft continues on its flight, a hand-off may take place to another Ground Station. When this occurs, it is signalled to both the airborne and ground users by another Join event. This Join event will identify the new GNI that both air and ground users must now use to communicate. On the airborne side, the Join event will include the following information: the address of the AMACS Ground Radio Station, the sub-network address of the connected Air/Ground router, Local identification of the connected Air/Ground router. The sub-network address of the connected Air/Ground router will permit unique identification of the Air/Ground router on the AMACS network. This value will be coded in three octets. The local identifier of the connected Air/Ground router will permit unique identification of the Air/Ground router from the GNI point of view (local ground reference). This value will be coded in one octet. It is noted that the Sub-network address of the GNI will consist of the initial part of the address of the AMACS Ground Radio Station through which the aircraft has been inserted. This will be ensured when deploying AMACS Ground Radio Stations: all such systems connected to the same GNI will have a common address prefix value that will permit unique identification of the connected GNI on the AMACS network. When a GNI is connected to several Air/Ground routers, the airborne system will receive one Join event per connected Air/Ground router. On the ground side, the Join event will include the following information: Sub-network address of the GNI, ICAO 27-bit address of the aircraft. If an aircraft moves from one GNI to another GNI which are both connected to the same Air/Ground router, the second Join event will be identified as a hand-off event. Seen from the points of view of the Airborne router and the Air/Ground routers, there will be two possible simultaneous paths for communication between them. The two paths will be distinguished by the different GNI addresses. Page 20 Error! Reference source not found. Edition Number: 1.0

29 Insertion into the new cell will occur before the aircraft leaves the previous cell, permitting the air-ground connectivity to be maintained throughout the flight. When the aircraft leaves a cell, a Leave event will be triggered on both the airborne and the Air/Ground routers. A hand-off between Ground Radio Stations in the same cluster (i.e. connected to the same GNI) will only have a minor impact on both air and ground. On the ground side, the aircraft will be inserted into the next AMACS Ground Radio Station cell and will leave the cell of the previous AMACS Ground Station. The GNI should thus not trigger a Join or Leave event, as routing information from the ATN or IP Air/Ground router's point of view is not impacted (the sub-network address of the GNI and the ICAO 27-bit address of the aircraft have not changed). On the airborne side, the AMACS Radio should recognize that it is still connected to the same GNI (same initial prefix of the address of the AMACS Ground Station Radio) in order to avoid triggering the Leave and Join events. The GNI will handle routing tables in order to identify the list of aircrafts connected and the identification of the AMACS Ground Radio Stations through which each is connected. When two routes exists to reach the same aircraft (handoff situation), the latest one established shall always be preferred Quality of Service (QoS) management AMACS QoS management The AMACS system will permit handling of QoS based on four parameters: QoS category, Priority, Class of Service (CoS), Integrity. QoS category: this flag will indicate whether the QoS must be provided on a best-effort or a guaranteed basis. From the point of view of the user (pilot or controller), the provision of data link services on a best-effort basis may not be satisfactory. For example, in the case of a trajectory negotiation, it would be better for the user to know that the data link network cannot deliver the requested service in time rather than trying to negotiate another route in a situation which may no longer be optimal. Although QoS shall be handled from an end-to-end viewpoint, the airground link is a potential bottleneck. Priority information will be used to distinguish the relative importance of the exchanged data with respect to gaining access to communications resources and to maintaining the requested QoS. The priority of different message categories has been specified by ICAO in terms of the ATN priority. Edition Number: 1.0 Error! Reference source not found. Page 21

30 When AMACS has multiple messages with different ATN priority, but the same AMACS transmit priority, queued to send then it shall take account of the ATN priority in deciding which messages to send first. Table 1.3 presents the priority mapping between the types of message category and the ATN network priority. Message category ATN Priority Network/systems management 14 Distress communications 13 Urgent communications 12 High priority flight safety messages 11 Normal priority flight safety messages 10 Meteorological communications 9 Flight regularity communications 8 Aeronautical information service messages 7 Network/systems administration 6 Aeronautical administrative messages 5 Unassigned 4 Urgent priority administrative and UN charter communications 3 High priority administrative and state/government communications Normal priority administrative 1 Low priority administrative 0 Table 1.3 Mapping between message category, ATN priority, and AMACS priority classification 2 AMACS will provide two types of Class of Service (CoS): Deterministic transmission (CoS1) The Deterministic transmission CoS will be used by applications that require a very high level of reliability for the transmission of short messages. Each aircraft will have a dedicated communication channel reserved for sending data at this CoS. This channel shall always be available and maintained by the datalink services. In case of failure to maintain such service, the user will be notified immediately. Concurrent transmission (CoS2) The Concurrent transmission CoS will permit transmission of longer messages, but without a guaranteed delivery time for transmission. Transmission of data at this CoS will be done in a concurrent way between all aircraft in the same cell. All data transmitted using this CoS will be characterized by a priority value. This Class of Service parameter will need to be understood in the context of the requested QoS category. The following table presents the impact of each of them where AMACS will handle a data transmission. Page 22 Error! Reference source not found. Edition Number: 1.0

31 Deterministic (CoS1) Concurrent (CoS2) Best Effort QoS requested In the case where there are no guaranteed QoS requests for the current CoS1 slot, then if the transmission fits in a CoS1 slot, transmit the data in the current CoS1 slot. If the transmission is too long for CoS1, then request a slot in CoS2. In case of reception failure (no ACK), then re-transmit the data in CoS2 slots (later in the same frame). In the case where there are guaranteed QoS requests for the current CoS1 slot, then transmit the data in CoS2 slots. Guaranteed QoS requested For the highest priority guaranteed QoS request for the current CoS1 slot, then if the transmission fits in a CoS1 slot, transmit the data in the current CoS1 slot. If the transmission is too long for CoS1, then request a slot in CoS2. In case of reception failure (no ACK), then re-transmit the data in CoS2 slots (later in the same frame). In the case where there are guaranteed higher priority QoS requests for the current CoS1 slot, then transmit the data in CoS2 slots in order of priority. Table 1.4 Impact of Class of Service on transmission The integrity parameter will offer the highest level of integrity over the AMACS communication channel. Based on the high level of integrity of the underlying channel, this service will consist of adding a simple CRC as part of the data message exchanged over the AMACS link. This CRC will permit the lowest residual bit error rate to be achieved. It should be noted that independently of the level of service to be provided, whether it is on a best-effort or guaranteed basis, mechanisms should be implemented to monitor the traffic load and usage in each AMACS cell. These mechanisms will allow a better anticipation of the capacity situation, permitting an early reconfiguration of the AMACS cell before capacity problems become a real issue End-to-end QoS management End-to-end communication will involve heterogeneous networks, including mainly an airground AMACS link and a ground transport network. Management of QoS on the AMACS link has been addressed in Section In order to be able to provide end-to-end QoS management between the airborne system and the ground controller system, two alternatives are envisaged: Implementation of QoS management mechanisms on the ground network infrastructure. Solutions based on IP based infrastructure, using IntServ or DiffServ model, are envisaged, Implementation of QoS management mechanisms at the transport level. This transport protocol shall be designed to be used over a network layer that provides besteffort service differentiation (called EDS Equivalent Differentiated Services). This solution has the advantage of providing this information directly to end users in order to decide whether the communication infrastructure is capable of providing the expected QoS. Edition Number: 1.0 Error! Reference source not found. Page 23

32 1.5 Standards The AMACS Physical Layer uses features and characteristics of GSM and UAT, for which international standards are available ([15], [16], [18], [19], [20]). The AMACS MAC and Data Link layers use features and protocols that have been standardized in VDL Mode 4. Thus while international standards are not yet in place for AMACS, the system is already well specified and the development of the appropriate AMACS standards will be facilitated by the availability of the existing material. 1.6 Technology Readiness Level (TRL) The system is effectively a collection of COTS components which are themselves extensively understood and deployed in the commercial marketplace. Standards for these COTS components are available and have been validated in an aeronautical context. A mention of these components is appropriate here: Physical layer (modulation) UAT and GSM MAC layer (access protocols) VDL4 These components have been tested, validated and demonstrated in relevant environments. The choice of components for AMACS is taking advantage of real-life experience gained from actual use of the UAT, GSM and VDL4 systems. Therefore the AMACS system is expected to meet TRL Level 5. Page 24 Error! Reference source not found. Edition Number: 1.0

33 2. APPLICATION OF TECHNOLOGY TO ATM This chapter describes the application of the AMACS system to aeronautical communications, which provides the basis for subsequent evaluation. This concept-of-use description involves: Concept of operation: description of how the technology is able to operate in the ATM environment. Applicable Frequency Band: the band or bands that are appropriate for the implementation of AMACS for aeronautical communications. Applicable Airspace: the airspace in which AMACS can practically provide aeronautical communications. Services Used and Performance required: the AMACS services that are best applicable to aeronautical communications and meet the required levels of performance. Architecture Integration: description of how the AMACS architecture integrates into the architecture for aeronautical communications. 2.1 Concept of operation: cellular deployment Introduction The obvious characteristic of this system which is different from VDL subnetworks is that AMACS requires the aircraft receiver to have an a priori knowledge of ground station positions. In today s VHF ATC and AOC systems, the ground station is implicitly identified by the aircraft through use of a pre-loaded channel map by sector. In other words, the system knows a priori which channel to tune to by virtue of knowledge of the sector being traversed. The AMACS system s use of cellular concepts is somewhat different. With regards to the identified operational scenarios, the AMACS system shall provide a spatial segmentation to take into account the typical message characteristics and requirements for different operational environment. Each cell shall be geographically distinct and will have its own dedicated ground station, which shall use a non-conflicting frequency (as described below). The cell partition will be built on the horizontal geographical plan, but will include vertical segregation as well. Each aircraft shall have knowledge of the parameters of all the cells including the relevant ground station frequencies. A cellular scheme will provide the adequate configuration to the airspace controlled by ATC. The size of the cell should (and could) be modulated according to the traffic. As a first assessment, three operational environments should be distinguished: En-Route Low Density cells, with a range of about the optical range 250NM (for lower airspace of the same type, smaller cells could be used taking into account the line of sight coverage limitations) Edition Number: 1.0 Error! Reference source not found. Page 25

34 En-Route High Density cells, with a maximal range of about 100 NM TMA cells, with a range of about 50NM modulated by the size of the airport The cells are tailored to operations, their sizes depending on: air traffic density deployed applications en-route, TMA, airport Figure 2.1 presents an example of such cellular deployment over the ECAC area. En-route ECAC Periphery 110-NM Radius Cell En-route ECAC Core Area 55-NM Radius Cell Major TMA Figure 2.1 AMACS cell deployment across ECAC Representative C/I derivation and Cellular deployment Application on a 12 frequencies pattern For further refinement of the AMACS system definition, the derivation of the C/I is a very important task. This is not a trivial exercise. As a first approach the C/I can be estimated from Page 26 Error! Reference source not found. Edition Number: 1.0

35 the C/N on a channel with AWGN. The degradation caused by a man-made signal, with the same characteristics (modulation, frequency, bandwidth, etc.) on the signal of interest, is smaller than the degradation caused by additive white Gaussian noise. This is because most of the time, the man-made signal is bound (the amplitude is limited, from a receiver point of view). This is not the case for a Gaussian random signal which corrupts the wanted signal. Some measurements on different systems have corroborated this hypothesis. The margin was about 3 db, for PSK systems. In order to derive an estimation of C/I from first principles, a cell planning based upon the reuse scheme presented below is considered. Figure frequencies pattern for cell planning With such a frequency plan, the worst-case interference scenario is an air-air interference, which is represented below in Figure 2.3. The following paragraphs present an evaluation of the C/I in the co-channel interference case. A cell radius R in the presented pattern is assumed. The distance from the interfered plane to its ground station is the wanted path d w = R. The distance from the interfered plane to its interferer on the same channel is the interfered path d i = 4R. The following propagation model is assumed : A(d) = (constant) + a.10 log(d) where the (constant) term stands for the contribution of frequency and other constant parameters to the attenuation, and a is the exponent applied to the distance. Here, a is assumed to take the value of 2 (free space model) and could range from 3.5 to 4 in ground cellular network. Given these elements, and assuming an omni-directional receiving antenna, if the transmitted powers are kept the same, the C/I at the receiver is simply a function of distance: C/I = A (d i ) A (d w ) = a.10 log(4r/r) C/I = a.6 db, and a 2 Edition Number: 1.0 Error! Reference source not found. Page 27

36 thus C/I 12 db The co-channel C/I ratio at the interfered terminal is always better than 12 db with this pattern and these assumptions. However considering the GMSK modulation choice, the GSM FEC rate 260/456 = 0 57 and a very light interleaving, 9 db C/I is considered sufficient. Furthermore, it is assumed here that the transmit power of the wanted signal (base station) is the same as the interfering signal (from an aircraft). For these reasons, at least the same performance is expected for the AMACS system. d w =R d i =4R Figure 2.3 Co-channel interference in a 12 channels re-use pattern 2.2 Applicable Frequency Band and electromagnetic compatibility AMACS systems shall be deployed in the lower L-band ( MHz) which already has an Aeronautical Radio-Navigation allocation. The use of this band is subject to WRC approval of co-prime allocation to AM(R)S. Additionally, a new channelization scheme will have to be provided in the band, to accommodate the AMACS system s use of channels ranging from 50 khz to 400 khz. One of the key problems for a future communication component that is intended operate in the L-band ( MHz) is co-siting with other radio transmitters that also operate in the L-band. Even if a frequency separation is implemented, to provide some decoupling with other spurious emissions, the robustness of the new communication link will be drastically affected by the proximity of other transmitters on the same aircraft. Page 28 Error! Reference source not found. Edition Number: 1.0

37 Therefore the solution is to take advantage of Pulse Blanking Techniques that have been used in many other cases to reduce the effect of strong interference (which is the case on board aircraft due to the very small system isolation). Taking into account that two of the major interferers for the new communication system will be short pulsing transmitters (DME and SSR/Mode S), the duty cycle of the jamming pulses will be lower than that of the AMACS bit duration: the impact of the interference will therefore be limited to a few bits in the frame for which data coding will be the appropriate answer to mitigate the impact of the interference on the frame error rate. Such a pulse blanking mechanism has been defined in the UAT standards and will have a common bus interconnecting the avionics elements that could benefit from the information provided (pulse blanking signal whenever a transmitter is on). 2.3 Airspace Application The use of AMACS systems will provide A/G communications in continental airspace (Core Area as well as periphery), which includes En-route (ENR) and Terminal (TMA) areas. We believe that the surface (APT) area should be covered by another terrestrial-based system (such as WiMAX ); oceanic and polar (ORP) communications should be supported by a satellite-based system. 2.4 ATM services supported AMACS is designed to support distinct modes of operation: The ground-supported mode where the aircraft fly within the range of ground datalink stations (these stations may be interconnected via ground links or not), The autonomous mode where the aircraft fly without any ground datalink infrastructure to support them. Hence AMACS is designed to support all existing and foreseen types of datalink application: Air-ground and ground-air point-to-point communications (as required today by AOC and also by emerging ATS applications such as COTRAC, ADS and CPDLC), Air-air, air-ground and ground-air multicast (i.e. locally broadcast) communications (as proposed for ADS-B, FIS-B and TIS-B), Air-air point-to-point communications (as envisaged for supporting autonomous separation assurance applications). Edition Number: 1.0 Error! Reference source not found. Page 29

38 2.5 Proposed Architecture for Technology System Avionics Figure 2.4 provides a notional view of the avionics required for a AMACS implementation of ADS-B and AOC and ATS functions. Figure 2.4 Possible Avionics for AMACS The air-to-air point-to-point connectivity is covered by the ADS-B function. For the air-to-air communication mode (generally supporting surveillance functions such as ASAS and ADS-B), two additional receivers are required in the avionics. While the primary receiver is tuned to the channel associated with the cell within which the aircraft is currently located, the additional receivers are tuned to downstream, adjacent cells to get any necessary signalling information associated with them Range The expected range of a typical ground station is 150 NM. The size of each cell is the area of a hexagon with a radius of 150 NM Link Budget See Annex C for an estimation of the Link Budget for AMACS. Page 30 Error! Reference source not found. Edition Number: 1.0

39 2.6 Performance Assurance Since even within the ATN framework the AMACS system will be in competition with other services, especially from commercially operated telecommunication systems, our design efforts have been focused at supporting very high performance data link services that will remain unattractive to general purpose telecommunication operators. Both the perspective of a doubling or more of air traffic densities in the most developed countries in the next twenty years and the progressive emergence of the "autonomous aircraft" operational concept will require that a future data link service simultaneously provides a very high integrity, a very high availability, an extremely short fault detection and recovery delay and a short and highly predictable transfer time. It cannot be assumed that the 95% maximum value of the transit time, which is the most usual metric today, will be a sufficiently rigorous specification in the future. Sooner or later, a datalink system supposed to address long-term needs yet unable to guarantee a 99% maximum transit time for certain categories of message is bound to become a problem rather than a solution. A high degree of confidence in the provided Quality of Service (QoS) will be required: critical in-flight data communication services for applications such as the ASAS or CPDLC in dense areas will be difficult to certify unless the whole system is designed ab initio with QoS verification in mind. Since relaxing the performance constraints is not really an option, the only cost-effective approach is to incorporate the demonstrability of performance into the very design of the AMACS system, on the one hand, and to avoid to translating the watchword of CNS/ATM integration into a multiplication of common failure modes, on the other hand. AMACS shall meet the COCR requirements for security, continuity, availability and integrity. For details of integrity see Section Edition Number: 1.0 Error! Reference source not found. Page 31

40 3. STATUS OF THE TECHNOLOGY 3.1 Summary The AMACS technology solution has been developed from a baseline of the existing UAT/GSM and VDL Mode 4 systems. AMACS has a robust physical layer, with appropriate UAT and GSM specifications which are re-used where appropriate, affording considerable advantage in the costs associated with the development, standardization and fielding of the technology. The GMSK modulation scheme is proven and meets the AMACS requirements. The highperformance MAC layer is based on the existing E-TDMA MAC layer concept. The frame structure is devised to meet the high-qos transmission requirements of AMACS. Existing VDL Mode 4 broadcast and reservation protocols are used. The broadcast experience from VDL Mode 4 is used to take advantage of known operating practices. 3.2 Status The design of AMACS is finalized at the Physical and MAC layer levels, with complete definitions of the frame, slot and message structures. The error correction coding definition is completed. The Channel structure, cellular deployment and network architecture are specified All of the AMACS message types have been defined and the definition of services has been provided. The protocols and system operations are defined for both point-to-point and broadcast communication Page 32 Error! Reference source not found. Edition Number: 1.0

41 ANNEX A : MODULATION OPTIONS A.1 Introduction UAT is a recent aeronautical digital link technology operating in the lower AM(R)S band which is considered as the starting point for designing an optimal physical layer for the AMACS system. UAT uses the CPFSK modulation scheme, with a modulation index of h = and a data rate of 1Mbps. UAT is designed to provide a reliable broadcast service in the L-band environment. The service usage profile is characterized by the fast transfer of short and periodic bursts of data, primarily in support of the ADS-B function. While attaining a high data rate for a safety-critical communications service is desirable, the physical layer needs to be designed around a stable and robust scheme to provide the most reliable transport medium possible. In order to serve this goal, an adaptation of the UAT physical specification is considered to address the main attributes and known weaknesses for a communications link. The main design goals considered when addressing the suitability of modulation options are: A low Bit-Error Rate (BER) at a low Signal-to-Noise Ratio (SNR). Good performance in multipath and fading environments. Occupation of least possible bandwidth. Introduce least amount of power in the RF environment. Low sensitivity to timing jitter (good decision thresholds). Easy and cost effective to implement. Of these, a major design criterion is spectrum efficiency. This is especially true in the targeted L-band, thus the design of the spectrum mask of the waveform is chosen to reduce the extent of secondary lobes. In order to achieve this, a pre-filtered variation of CPFSK is considered. The MSK (Minimum Shift Keying) family is a type of modulation offering a robust means of transmitting data in wireless systems where the data rate is relatively low compared to the channel bandwidth. MSK is a special form of FSK (Frequency Shift Keying). In MSK, the differential frequencies used to represent the data symbols are orthogonal, making reception particularly straightforward. The term 'minimum' results from the fact that from the FSK viewpoint, the two frequencies used are at the minimum allowable separation whilst maintaining orthogonality. The GMSK (Gaussian Minimum Shift Keying) variant is specifically being considered both for its technical qualities and because it is popular in the commercial mass market and is used with different characteristics in GSM, DECT and TETRAPOL. Hence it offers significant advantage in cost reductions for the development and manufacture of avionics and groundbased equipment. Edition Number: 1.0 Error! Reference source not found. Page 33

42 Gaussian filter IQ signals generation IQ modulation IF to RF Figure A.1 Gaussian filtered modulation Figure A.1 shows the block diagram of a GMSK modulator. The Gaussian filter used in GMSK is generally specified by its BT product, where B is the 3 db bandwidth of the Gaussian filter and T is the symbol duration. Note that this is in contrast with the CPFSK specified for UAT where the filtering is applied after the generation of the I and Q component signals. For the Gaussian filter used in GSM, BT = 0.3. The filter is specified by its impulse response given by : and BT = 0.3 for GSM., with Pre-filtering in FSK can be considered as an extension of the concept of continuous phase modulation. A basic FSK shows abrupt phase transition from one symbol to another, which cause wider spectrum occupation. Continuous Phase FSK addresses this problem by providing smoother transitions from one symbol to another. The concept of pre-filtered FSK is based upon the same considerations. With a CPFSK, the phase transitions are continuous, but can still be abrupt: the frequency transition is not continuous. Pre-filtered Gaussian CPFSK specifically targets this problem. All the transitions are much smoother, as shown in Figure A.2. It represents the phase tree, i.e. the possible values of the phase throughout time, during transmission. The straight lines show the phase transitions for a CPFSK, while the curved lines show the transition for a GMSK. Figure A.2 Gaussian filtered modulation These transitions result in a transmitted spectrum for which the secondary lobes are significantly reduced (Figure A.3). This is a major advantage in a cellular system where frequency reuse is mandatory. GMSK has high spectral efficiency, but it needs a higher power level than the more traditional schemes to reliably communicate the same amount of data. A firstorder estimation of the transmitter power likely required by an AMACS transponder is provided in Annex C, although this will largely depend on the deployment scenario (service volume size etc.). Page 34 Error! Reference source not found. Edition Number: 1.0

43 Figure A.3 Spectral density for MSK vs GMSK A.2 Modulation design choice Given the fundamental principles of GMSK modulation shown above, three proposals have been drawn up taking into consideration the desirable characteristics, design goals, and the spectral environment that define the theatre of operations for AMACS. First proposal (GSM baseline): GMSK : h = 0.5 & BT = 0.3 Gross bit rate : 270 kbps Channel bandwidth : 200 khz C/I = 9 db Note - The above characteristics are those specified for the GSM system, including convolutional coding (see Annex B). It is expected that better performance may be attained. Other alternatives are considered in order to explore any trade-offs that may exist between system capacity and spectral efficiency, using different channelizations. Second proposal : GMSK : h = 0.5 & BT = 0.3 Gross bit rate : ~ 540 kbps Channel bandwidth : 400 khz Expected C/I expected to be in the region of 9dB 2 2 Simulations will be required in order to confirm and substantiate this estimation. Edition Number: 1.0 Error! Reference source not found. Page 35

44 Concatenated coding is considered in Annex B. It is expected to be more efficient than the coding used in GSM, hence the objective of a C/I of 9 db in co-channel interference is attainable. In former study efforts, a CPFSK solution with a modulation index h of had been found to have suitable characteristics. Hence a third proposal is considered using this value, bearing in mind that, with all other parameters kept constant, the gross bit rate over a constant channel bandwidth is reduced as the modulation index increases. Third proposal : GFSK : h = & BT = 0.3 Gross bit rate : ~ 378 kbps Channel bandwidth : 400 khz In order to illustrate the influence of the modulation index, h, on the performances of CPFSK, the following discussion presents the comparison of three modulations: BPSK (used as reference), and binary CPFSK with h = and MSK with h = 0 5. The results presented in Figure A.4 are exact for non-filtered modulations. It is noted that for filtered modulations, a significant performance enhancement is expected. For all these modulations, in the presence of AWGN, the probability of error (BER) can be expressed as follows : 1 Eb 1 b Pe = erfc 2 N 2 with b=1 for BPSK, and b=sinc(2h) for any binary CPFSK. They are represented against the Signal-to-Noise Ratio (SNR) in Figure A.4, where the blue curve represents BPSK, the green curve represents MSK, and red curve represents CPFSK. As can be seen in the figure, the degradation in terms of power, with respect to the BPSK modulation, is 3 db for h = 0 5 and 2.2 db for h = Hence, using CPFSK (h = 0 715) instead of MSK (h = 0 5) provides a saving of 0.8 db in the power link budget for a given BER, but costs about 30% in terms of bit rate, if the channel bandwidth remains constant. 0 Figure A.4 Comparison of BPSK, MSK (h=0 5) and CPFSK (h=0 715) Page 36 Error! Reference source not found. Edition Number: 1.0