SIU 2022 · 30th Signal Processing and Communications Applications Conference · IEEE

Improving Synchronization Time in 6TiSCH Networks with Smart Antennas

Akıllı Antenlere Sahip 6TiSCH Ağlarda Senkronizasyon Süresinin İyileştirilmesi
Sercan Külcü · Sedat Görmüş
Computer Engineering Department, Karadeniz Technical University, Trabzon, Türkiye
ProblemNetwork formationAntenna modelScan packet Beam alignmentSetupResultsConclusion
Paper (IEEE Xplore) · Use ← → or Space to navigate, F for full screen

Summary

At a Glance

Steerable antennas reach farther and suffer less interference, but their narrow beam makes it hard for a new node to hear the network. A new scan packet lets nodes learn their neighbours' directions before they synchronize; beacons are then sent and heard in those directions.

Improvement with scan packets≈ 25%paper, varies with density
75 nodes, 1 channel≈ 3.9 minrandom direction: ≈ 9.8 min
Omni, 75 nodes≈ 15.9 mincollisions dominate
Autonomous cells≈ 1.3 minsteerable, 75 nodes

Minute values are read from the vector plots of Figs. 3 and 4 (Cooja, 10 runs each).

I–II · Introduction and Background

Smart Antennas Help, Until a Node Has to Join

Steerable (ESPAR) antennas

  • stronger transmission in the chosen direction, longer range
  • ignore interference from unwanted directions
  • more parallel transfers on the same frequency
  • electronically steered, cheap enough for IoT nodes

The price

  • a node must know which direction to use for each neighbour
  • a narrow beam covers little: work an omni antenna does once must be repeated around the node
  • joining the network and finding neighbours takes longer
  • existing fixes for slow network formation assume omni antennas
Schedules now have three dimensions: time slot, frequency channel and antenna direction.

II.B · 6TiSCH

How a 6TiSCH Network Forms

Listena new node keeps its radio on and scans up to 16 channels
→
Enhanced Beaconfirst from the root, later from every synchronized node
→
Synchronizedthe node joins the TSCH slotframe
→
DIO messagesRPL builds the routing tree (DODAG)
→
Parent selectedthe best neighbour becomes the parent
Minimal schedule

Until it has its own schedule, every node uses one or more shared cells for EBs, join traffic and RPL messages.

Waiting costs energy

Nodes far from the root wait for nearer neighbours to join first, with the radio always on.

Dense networks

Collisions in the shared cell slow formation even with omni antennas; a narrow beam slows it further.

III.A · Fig. 1 and Eq. (1) · Interactive

Antenna Pattern and Link Budget

Beam of A off target
Beam of B off target
Distance

RSS = PTX + PPL − 10K·log10(d/d0) + G(θ12) + G(θ21), with PTX = 0 dBm, PPL = −40 dB, K = 2.5, d0 = 1 m. Pattern: smooth curve through the paper's 5, −5, −15 dBi; threshold: an omni link reaches 50 m.

III.B · Fig. 2, Algorithms 1–2

The Scan Packet

  • Synchronized nodes send Enhanced Beacons; unsynchronized nodes send the new scan packet.
  • Unsynchronized nodes keep their radio on, so they hear scan packets with high probability.
  • The receiver stores the sender's MAC ID with its own current antenna direction and RSSI: a virtual neighbour list built before synchronization.
  • After it receives an EB and synchronizes, a node stops sending scan packets and starts sending EBs.
  • In the shared cell, EBs are sent, and nodes listen, towards listed neighbours instead of random directions.
  • The radio layer must report direction and RSSI to the MAC layer, so both layers were redesigned.
Algorithm 1: neighbour discovery
id ← extract_id(packet)
n  ← get_neighbor(id, list)
if n is null:              # new neighbour
    insert(create(id, current_dir, current_rssi))
else if current_rssi > n.rssi:
    n.dir  ← current_dir    # better direction
    n.rssi ← current_rssi
Algorithm 2: direction in the shared cell
i ← random(list_size)
if i is 0:
    return random(2π)      # P = 1 / (neighbours + 1)
else:
    return list[i].dir     # towards a known neighbour

III · Why Directions Matter · Interactive

Beam Alignment: Will One Beacon Be Heard?

Distance
Listed neighbours

Illustrative model, not a result of the paper: one transmission, one listener, no collisions, link budget of the previous slide. Neighbour-aware: each side points exactly at the other with probability 1/(k+1) (Algorithm 2), otherwise at a random angle.

IV · Table I

Test Setup

ParameterValue
Slotframe length101 slots
Shared cells1
EB interval16 s to 50 s
Range, omni antenna50 m
Range, smart antenna15 m to 80 m
Area200 m × 200 m, random placement
Nodes25, 35, 45, 55, 65, 75
Test lengthuntil the network is synchronized
Cooja + Contiki

Nodes run Contiki in emulation; the radio medium was extended to include relative antenna angles, not only distance.

10 runs

Each configuration is repeated ten times and averaged.

Antenna configurations

Omni, steerable with random directions, steerable with scan packets, and both with autonomous cells (control messages copied per neighbour, as in Orchestra).

IV · Fig. 3 · Interactive

Synchronization Time with One Channel

Click a legend entry to hide a curve; hover for values. Omni is fastest in small networks but explodes with density (collisions). Steerable antennas first gain from more neighbours in reach, then lose to interference. Scan packets keep the time flat in dense networks; autonomous cells are short and scalable for both antenna types.

IV · Fig. 4 · Interactive

Scan Packets with One and Two Channels

Nodes1 ch.2 ch.

Table: change in time with scan packets (green = faster). A second channel makes joining slower, because a new node must listen on each channel in turn. The gain grows with density; in small one-channel networks scan packets cost time.

IV · Discussion

What Drives Synchronization Time?

Density

More nodes mean more neighbours in reach but more collisions in the shared cell. Omni antennas suffer most; narrow beams shield steerable nodes.

Direction knowledge

Random beams rarely meet. Scan packets turn guesses into aimed transmissions and receptions.

Channels

More frequency resources lengthen joining: an unsynchronized node must find the right channel as well as the right direction.

Desynchronization

A node that misses EBs for too long leaves the network and must join again, lengthening total formation time.

Dedicated autonomous cells per neighbour remove contention altogether and give short, size-independent synchronization times.

V · Conclusion

Conclusion

  • 6TiSCH was designed for omni antennas; steerable antennas slow down joining and neighbour discovery.
  • Scan packets let nodes discover neighbours before synchronization; shared-cell traffic then follows known directions, improving synchronization time by about 25% depending on density.
  • Copying control messages to autonomously allocated cells in each neighbour's direction gives a solution that does not depend on network size.
  • Future work: machine learning for more efficient use of shared cells.

Supported by TÜBİTAK under project 118E289.

V · Conclusion

Check Questions

  1. Which two phases must complete before a 6TiSCH network is fully functional?
  2. Why does a steerable antenna make joining harder than an omni antenna?
  3. What does a node store when it hears a scan packet, and why can it hear one before it is synchronized?
  4. Why does Algorithm 2 still pick a random direction with probability 1/(neighbours + 1)?
  5. Why is the omni antenna fastest at 25 nodes but slowest at 75?
Show answers
  1. Synchronization through Enhanced Beacons, then building the RPL tree with DIO messages.
  2. The beam covers only part of the surroundings; the node must also find the right direction, and random directions rarely line up.
  3. The sender's MAC ID, its own current antenna direction and the RSSI; unsynchronized nodes keep the radio on while they search.
  4. So that new nodes, not yet in the list, can still be discovered.
  5. In small networks it reaches everyone at once; in dense ones its wide coverage causes more collisions and lost packets.

References

Selected References

  • Kulcu, S., Gormus, S., Jin, Y. (2021). Integration of steerable smart antennas to IETF 6TiSCH protocol for high reliability wireless IoT networks. IEEE Access, 9, 147780–147790.
  • Vilajosana, X. et al. (2019). IETF 6TiSCH: A tutorial. IEEE Commun. Surveys Tuts., 22(1), 595–615.
  • Vilajosana, X., Pister, K., Watteyne, T. (2017). Minimal IPv6 over the TSCH mode of IEEE 802.15.4e (6TiSCH) configuration. RFC 8180.
  • Vallati, C. et al. (2018). Improving network formation in 6TiSCH networks. IEEE Trans. Mobile Comput., 18(1), 98–110.
  • Duquennoy, S. et al. (2015). Orchestra: Robust mesh networks through autonomously scheduled TSCH. ACM SenSys, 337–350.

See the full paper for the complete reference list. Result curves are redrawn from the paper's figures.

Thank You

Improving Synchronization Time in 6TiSCH Networks with Smart Antennas
How to cite: Külcü, S., & Görmüş, S. (2022). Improving Synchronization Time in 6TiSCH Networks with Smart Antennas. In 2022 30th Signal Processing and Communications Applications Conference (SIU), pp. 1–4. IEEE. doi:10.1109/SIU55565.2022.9864879
sercan.kulcu@giresun.edu.tr · All publications
1 / 1