1. Speaker: Po-Hung Chen Advisor: Dr. Ho-Ting Wu 2016/10/12
A Model-based Beacon Scheduling algorithm for IEEE e TSCH networks
Speaker: Po-Hung Chen
Advisor: Dr. Ho-Ting Wu
2016/10/12

2. Outline Introduction IEEE 802.15.4e TSCH Network Formation Process
Model-Based Beacon Scheduling Algorithm
Algorithm for Comparison
Simulation Setting
Simulation Result
Conclusion
Reference

3. Introduction
Time Slotted Channel Hopping (TSCH) is an emerging MAC protocol defined in the IEEE e standard, combining time slotted access with multi-channel and channel hopping capabilities.
In this paper we focus on the network formation process of TSCH networks. This relies on periodic advertisement of Enhanced Beacons (EBs), however, the standard does not specify any advertising strategy.
Finally, we propose a new Model-based Beacon Scheduling (MBS) algorithm that approximates the optimal strategy in real networks.

4. IEEE 802.15.4e Application Presentation
IEEE is a standard which specifies the physical layer and mac layer for low-rate wireless personal area networks (LR-WPANs) and it has defined it in 2003.
IEEE MAC is used CSMA/CA.
IEEE Frequency is (k-11)MHz, k=11~26
IEEE e is the first MAC amendment which has defined in 2012.
Application
Presentation
Session
Transport
Network
Data Link
Physical
(0~15)
IEEE e MAC
IEEE PHY

5. IEEE e(Cont.)

6. IEEE e(Cont.)
Tree Topology

7. TSCH
The Time Slotted Channel Hopping (TSCH) protocol is one of the new MAC Behavior modes introduced by the IEEE e standard.
TSCH combines
Time slotted access
Predictable and bounded latency, Guaranteed bandwidth
Multi-channel communication
increased network capacity (More node can communication at the same time using different channels)
Channel hopping
mitigates the effects of interference and multipath fading, improves reliability

8. TSCH(Cont.)
Time slotted access

9. TSCH(Cont.) TSCH provides two types link: Dedicated Links: Tx, Rx
Shared Links
Channel hopping
𝑓: physical channel
ASN: absolute slot number
N c : number of channels
𝑓= 𝐴𝑆𝑁+𝑐ℎ𝑎𝑛𝑛𝑒𝑙𝑂𝑓𝑓𝑠𝑒𝑡 % 𝑁 𝑐

10. TSCH(Cont.) The special Enhanced Beacons(EBs) for TSCH
Synchronization information
Channel hopping information
Timeslot information
Initial link and slotframe information
The special EBs for TSCH
1. Synchronization information:
allows new devices to synchronize to the network
2. Channel hopping information:
allows new devices to learn the channel hopping sequence
3. Timeslot information:
describes when to expect a frame transmission and when to send an acknowledgment
4. Initial link and slotframe information:
allows new devices to know:
a. when to listen for transmissions from the advertising device
b. when to transmit to the advertising device

11. Network Formation Process
A node wishing to join a TSCH network turns its radio on and starts scanning for possible EBs. As soon as it receives a valid EB, its switches the device into TSCH mode.
When radio-on for a long time, the joining node will waste a lot of energy and long time to join the network.
Therefore, have a good algorithm for how to advertise and when to advertise an EB is a very important thing.
But IEEE e standard does not specify any EB advertising strategy.

12. Network Formation Process(Cont.)
In TSCH networks nodes use a link to exchange messages, each link is represented by a couple [timeslot, channelOffset].
𝑁 𝑠 : the number of timeslots in a slotframe
𝑁 𝑐 : the number of available frequencies
𝑁 𝑠 and 𝑁 𝑐 are coprime.
𝑓=(𝐴𝑆𝑁+ 𝑐ℎ𝑎𝑛𝑛𝑒𝑙𝑂𝑓𝑓𝑠𝑒𝑡 ) 𝑚𝑜𝑑 𝑁 𝑐

13. Network Formation Process(Cont.)
An example, 𝑁 𝑠 =3, 𝑁 𝑐 =4
The sequence of channels used for communication in a certain link repeats every 𝑇 = 𝑁 𝑐 ∗ 𝑁 𝑠 .
At successive slotframes, all links change frequency.
All links follow the same channel hopping sequence.

14. Network Formation Process(Cont.)
We assume that the joining node activates in a random timeslot and starts listening for EBs on a frequency 𝑓𝑗.
Hence, it has a probability 1⁄𝑇 to be initially in any state 𝑆 𝑖 (𝑖=0,1,…,𝑇−1).
State 𝑆𝑖: EBs are sent on frequency 𝑓𝑗 during this slot.
𝑝 𝑙𝑜𝑠𝑠 : the probability a non-valid EB is received, due to collisions and/or channel errors.
(1− 𝑝 𝑙𝑜𝑠𝑠 ): the probability a valid EB is received.

15. Network Formation Process(Cont.)
Discrete Time Markov Chain

16. Network Formation Process(Cont.)
And this paper propose a average joining time 𝜏 𝑗𝑜𝑖𝑛 equation by calculating the average absorption time 𝜂 in the DTMC.

17. Network Formation Process(Cont.)
We implemented an ad-hoc simulator in C++.
It’s performed replications, and 95% confidence intervals.

18. Model-Based Beacon Scheduling Algorithm
An example of optimal schedule.
𝑁 𝑠 =23, 𝑁 𝑐 =16 𝑎𝑛𝑑 𝑁 𝑏 =5
𝑇= 𝑁 𝑠 ∗ 𝑁 𝑐 =368
𝑑 𝑗 =𝑑( 𝑏 𝑗 , 𝑏 𝑗+1 )
We used the AMPL modeling language, in conjunction with the CPLEX solver.
𝑑 0 =73, 𝑑 1 =74, 𝑑 2 =74,
𝑑 3 =74, 𝑑 4 =73

19. Model-Based Beacon Scheduling Algorithm (Cont.)

20. Model-Based Beacon Scheduling Algorithm (Cont.)
In fact, it may not be possible for a node to transmit all the Nb EBs specified in the schedule, for two main reasons.
Beacon collisions
Transmission unfeasibility
Therefore, in a real network it is necessary to determine which nodes will be in charge of transmitting EBs, and assign links for EB transmission to these nodes, according to the optimal schedule.

21. Algorithm for Comparison
Random (RD)
The 𝑁 𝑏 link devoted to EB advertisement are randomly distributed within the slotframe.
Channel Offset 3 2 1
timeslot

22. Algorithm for Comparison(Cont.)
Random Vertical (RV)
EBs are sent during the first timeslot of the slotframe using all the available channel offsets.
𝑁 𝑏 ≤ 𝑁 𝑐
Channel Offset 3 2 1
timeslot

23. Algorithm for Comparison(Cont.)
Random Horizontal (RH)
EBs are transmitted during all the timeslots of the slotframe, always using the same channel offset (i.e., 0).
Channel Offset 3 2 1
timeslot

24. Simulation Setting Two scenarios: Single Joining Node Network Setup
𝑁 𝑠 =101
𝑁 𝑐 =16
It’s performed independent replications, and 95% confidence intervals.

25. Simulation Result
Single Joining Node, 𝑝 𝑙𝑜𝑠𝑠 =0%

26. Simulation Result(Cont.)
Single Joining Node, 𝑝 𝑙𝑜𝑠𝑠 =10%

27. Simulation Result(Cont.)
Single Joining Node, 𝑝 𝑙𝑜𝑠𝑠 =30%

28. Simulation Result(Cont.)
Network Setup, 20 nodes

29. Simulation Result(Cont.)
Network Setup, 40 nodes

30. Simulation Result(Cont.)
Network Setup, 60 nodes

31. Conclusion
In this paper we have investigated the network formation process in e TSCH networks.
Finally, we have defined a Model-based Beacon Scheduling (MBS) algorithm that allows network nodes to autonomously select the links to use for advertising EBs, starting from the optimal EB schedule provided by the model.
Through simulations, that MBS outperforms all the other solutions present in the literature, providing a significant reduction in the average joining time of single nodes and the average network formation time.

32. Reference
Domenico De Guglielmo; Simone Brienza; Giuseppe Anastasi, “A Model-based Beacon Scheduling algorithm for IEEE e TSCH networks” in 2016 IEEE 17th International Symposium on A World of Wireless, Mobile and Multimedia Networks (WoWMoM), June , pp1-9
Domenico De Guglielmo; Alessio Seghetti; Giuseppe Anastasi; Marco Conti, “A performance analysis of the network formation process in IEEE e TSCH wireless sensor/actuator networks” in IEEE Symposium on Computers and Communication (ISCC), June 2014, pp1-6
IETF RFC 7554

33. Thank you for listening.