Wireless Communications Lecture 4: Multiple Access Technologies

Transcription

1 Wireless Communications Lecture 4: Multiple Access Technologies Module Representive: Prof. Dr.-Ing. Hans D. Schotten Lecturer: Dr. Vincenzo Sciancalepore Institute of Wireless Communication (WiCon) Department of Electrical and Computer Engineering TU Kaiserslautern SS 2018

2 Outline Transmission on shared means Multiplexing and Multiple Access Scheduled Access Protocols Random Access Protocols Channelization Protocols V. Sciancalepore Wireless Communications 2

3 How to share the means? Multiple information sent through the same physical channel This may create collisions and data loss Several tecniques may help to avoid collisions or recover from data loss

4 Multiplexing vs. multiple access Multiplexing allows to send data from one transmitter to multiple receivers (DOWNLINK) Multiple access is required to send data from multiple transmitters to one single receiver (UPLINK) V. Sciancalepore Wireless Communications 4

5 The multiple access in wireless networks Access to the same shared means can be managed by using: Random access Scheduled access Random access: Simultaneous transmissions may overlap and result in packet collision (data loss) Scheduled access: Sequential transmissions result in no collisions Channelized access: Transmission are centrally controlled to avoid collisions V. Sciancalepore Wireless Communications 5

6 Scheduled access λλ MM λλ MM λλ MM μμ 1 μμ 2 μμ MM When multiple transmitters have to send their packets on a shared means, we can assume different local queues that are emptied when scheduled Scheduling algorithms are defined to select which transmitter can occupy the channel and send data Polling scheme is a way to sequentially grant access (token) to any single transmitter V. Sciancalepore Wireless Communications 6

7 Exhaustive polling Transmitters can transmit all packets residing in the queue when the token has been received Assuming a global arrival rate λλ, packet transmission time TT and token passing time h, let us calculate the probability that the channel is occupied while one transmitter is sending a packet at a random time tt ρρ = λλλλ The average waiting time for a packet in the queue to be transmitted is EE WW = WW 1 + WW 2 + WW 3 WW 1 = EE NN cc TT = λλλλλλ WW = ρρρρ[ww] WW 2 = ρρ TT 2 + (1 ρρ) h 2 WW 3 = MM 1 2 EE WW = ρρρρ WW + ρρ TT ρρ h 2 + (MM 1) 2 h h V. Sciancalepore Wireless Communications 7

8 Exhaustive polling It yields the following EE WW = ρρ 2(1 ρρ) MM ρρ TT + h, where 0 ρρ 1 2(1 ρρ) Waiting time of a single queue Additional waiting time due to token passing time We can calculate the average token cycle time EE CC = λλλλ CC TT + MMM EE CC = MMM 1 ρρ V. Sciancalepore Wireless Communications 8

9 Gated polling Transmitters can transmit all packets that are in the queue at the time when the token arrives. Packets arriving in the queue while the transmitter is occupying the channel, need to wait for the next token. EE WW = ρρρρ WW + ρρ TT ρρ h (MM 1) + h + WW where WW 4 = ρρ MMM = hρρ is the additional cycle the packet may wait MM It yields EE WW = ρρ MM + ρρ TT + 2(1 ρρ) 2(1 ρρ) h V. Sciancalepore Wireless Communications 9

10 Limited polling Transmitters can send up to K packets when the token arrives. When K=1, the scheme is called Round Robin EE WW = ρρρρ WW + ρρ TT ρρ h 2 where WW 4 = ρρ MMM = hρρ and WW MM 5 = EE[NN cc] MM + (MM 1) 2 MMM = λλee[ww]h h + WW 4 + WW 5 It yields EE WW = ρρ 2(1 ρρ h+tt TT ) TT + MM+ρρ 2(1 ρρ h+tt TT ) h, where 0 ρρ TT TT V. Sciancalepore Wireless Communications 10

11 Random access When transmissions are performed simultaneously, collisions may occur. The transmitter needs to get feedbacks about the status of its previous transmission. In case of collision, recovering mechanisms are needed to retransmit (and reduce the probability of collision) The simplest random access protocol is called Aloha Prof. Norman Abramson was the inventor of this first wireless network (AlohaNet) V. Sciancalepore Wireless Communications 11

12 Aloha random protocol The transmitter may send the packet as soon as it arrives in the queue If there is a collision (data loss), the transmitter waits for a random time and retransmit the packet Let us consider the transmission starting times distributed as a Poisson process with rate λλ with a normalized rate GG = λλλλ The success probability to send and deliver a packet depends on the probability that no other transmitter is occupying the shared means in a 2TT interval: PP ss = ee 2GG Normalized throughput is S = GGee 2GG V. Sciancalepore Wireless Communications 12

13 Aloha random protocol When different transmitters are synchronized, we call the protocol Slotted Aloha. This improves the performance of the pure Aloha by reducing the vunerability time window to T Normalized throughput is S = GGee GG V. Sciancalepore Wireless Communications 13

14 Channelization protocols FDM/FDMA: Frequency Division Multiplexing / Frequency Division Multiple Access TDM/TDMA: Time Division Multiplexing / Time Division Multiple Access CDM/CDMA: Code Division Multiplexing / Code Division Multiple Access V. Sciancalepore Wireless Communications 14

15 Frequency division multiple access The total available bandwidth is divided into several (small bandwidth) channels. Simple technique used for several telecommunication systems, such as cable television system, fiber optic communications or aereospace telemetry. In fiber optic systems, it may be called Wavelength Division Multiple Access (WDMA), as the light may be reffered by the wave-length instead of its frequency. V. Sciancalepore Wireless Communications 15

16 Time division multiple access The time is divided into several time slots that are grouped into time frames. A single frame consists of N time slots. Each transmission occurs on a isolated time slot to prevent collisions. A guard time TT gg can be defined to ensure that different transmissions do not interfere each others (because of overlapping transmissions): TT gg = max(2ττ ii ) Propagation time for transmitter i V. Sciancalepore Wireless Communications 16

17 Time division multiple access The guard time can play a fundamental role in the overall system efficiency When the propagation time ττ is known in advance, the transmission time can be anticipated to compensate Due to the channel condition, the propagation time ττ can vary and must be estimated and sent back to the transmitter V. Sciancalepore Wireless Communications 17

18 Code division multiple access Useful bits are multiplied by a code Codes are orthogonal in the code division multiplexing while they have limited correlation in the code division multiple access TT NN 1 ss ii CC ii ii=0 CC kk = ss kk V. Sciancalepore Wireless Communications 18

19 Code division multiple access Using the code, we can expand the radio bandwidth and allow different signals to transmit on the same radio band The signal is multiplied by the code at the receiver and reduced V. Sciancalepore Wireless Communications 19

20 Orthogonal division multiple access The channel is divided into multiple narrow orthogonal bands spaced so as to avoid interference. Data to be transmitted is divided into multiple bit streams that are modulated onto the subcarriers. Time slots (within each subchannel) are used to package the data V. Sciancalepore Wireless Communications 20