Kyeong Soo Kim, "Energy-efficient time synchronization in wireless sensor networks," Invited talk, 2019 Distinguished Lecture and International Interdisciplinary Workshop, Chungnam National University (CNU), Daejeon, Korea, August 5-9, 2019.

1. Energy-Efficient Time Synchronization in
Wireless Sensor Networks
Kyeong Soo (Joseph) Kim
(With X. Huan, S. Lee, E. G. Lim, and A. Marshall)
Department of Electrical and Electronic Engineering
Xi’an Jiaotong-Liverpool University
2019 Distinguished Lecture and
International Interdisciplinary Workshop
Chungnam National University
August 5-9, 2019
1 / 62

3. Outline
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Simulation Results
Next Steps: Multi-Hop Time Synchronization
Conclusions
3 / 62

4. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Hardware and Logical Clock Models
Effect of Clock Skew on Measurement Time
Estimation
Asynchronous Source Clock Frequency Recovery at
Sensor Nodes: One-Way Clock Skew Estimation
Simulation Results
Performance of One-Way Clock Skew Estimation
Performance of Measurement Time Estimation and
Energy Efficiency
Effect of Bundling of Measurement Data
Next Steps: Multi-Hop Time Synchronization
Conclusions
4 / 62

5. WSN Applications - Smart Cities1
1
Source: Transmedia Newswire.
5 / 62

6. WSN Applications - Oil & Gas2
2
Source: OleumTech.
6 / 62

7. WSN Applications - Military3,4
3
Source: ElProCus.
4
Source: Smart Dust - Berkeley Robotics.
7 / 62

8. WSN Applications - Military3,4
3
Source: ElProCus.
4
Source: Smart Dust - Berkeley Robotics.
7 / 62

9. WSN Application - Mix with UAV5
5
Source: IEEE Pervasive Computing.
8 / 62

10. Asymmetric Wireless Sensor Networks
HeadeNode
Internet
RemoteeUser
SensoreNodes
9 / 62

11. Head Node
I A base station that serves as a
gateway between wired and
wireless networks.
I A center for fusion of data
from distributed sensors.
I Equipped with a powerful
processor and supplied power
from outlet.
10 / 62

12. Head Node
I A base station that serves as a
gateway between wired and
wireless networks.
I A center for fusion of data
from distributed sensors.
I Equipped with a powerful
processor and supplied power
from outlet.
10 / 62

13. Head Node
I A base station that serves as a
gateway between wired and
wireless networks.
I A center for fusion of data
from distributed sensors.
I Equipped with a powerful
processor and supplied power
from outlet.
10 / 62

14. Sensor Node
I Measuring data and/or detect
events with sensors and
connected to a WSN only
through wireless channels.
I Limited in processing and
battery-powered.
11 / 62

15. Sensor Node
I Measuring data and/or detect
events with sensors and
connected to a WSN only
through wireless channels.
I Limited in processing and
battery-powered.
11 / 62

16. Design Goals
I Achieving sub-microsecond time synchronization
accuracy
I Through propagation delay compensation.
I With higher energy efficiency at battery-powered
sensor nodes
I Minimize the number of packet transmissions and the
amount of computation at sensor nodes.
12 / 62

17. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Hardware and Logical Clock Models
Effect of Clock Skew on Measurement Time
Estimation
Asynchronous Source Clock Frequency Recovery at
Sensor Nodes: One-Way Clock Skew Estimation
Simulation Results
Performance of One-Way Clock Skew Estimation
Performance of Measurement Time Estimation and
Energy Efficiency
Effect of Bundling of Measurement Data
Next Steps: Multi-Hop Time Synchronization
Conclusions
13 / 62

18. Two Kinds of Synchronization
I Phase.
I Frequency.
14 / 62

19. Two Kinds of Synchronization
I Phase. I Frequency.
14 / 62

20. Effects of Time and Space
The effects of time and space are so closely related that
they cannot be easily separated from each other as in the
following examples:
I Synchronization and localization accuracies.
I In time-based localization.
I e.g. Time of arrival (TOA).
I Clock offset and propagation delay.
I In one-way synchronization.
I e.g. Flooding time synchronization protocol (FTSP).
15 / 62

21. Synchronization and Localization Accuracies
I Accuracies
I 1 ms ↔ 300 km
I 1 µs ↔ 300 m
I 1 ns ↔ 30 cm
I 1 ps ↔ 0.3 mm
I Time-based localization schemes
I Time of arrival (TOA)
I Time difference of arrival (TDOA)
I A special variation of TDOA with virtual anchors does
not require synchronization among devices.
⇒ See the next slide.
16 / 62

22. TDOA with Virtual Anchors 6
Anchor
Agent
Virtual
Anchors
6
E. Leitinger et al., IEEE J. Sel. Areas Commun., vol. 33, no. 11, pp.
2313–2328, Nov. 2015.
17 / 62

23. Clock Offset and Propagation Delay
Can the receiver distinguish between the following two
cases if θ = d?
Packet with
Timestamp T vs. Packet with
Timestamp T
TX
RX
TX
RX
• : Clock offset
• : Propagation delay
I Answer is “No”.
I Two-way message exchanges needed for delay
compensation.
18 / 62

24. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Hardware and Logical Clock Models
Effect of Clock Skew on Measurement Time
Estimation
Asynchronous Source Clock Frequency Recovery at
Sensor Nodes: One-Way Clock Skew Estimation
Simulation Results
Performance of One-Way Clock Skew Estimation
Performance of Measurement Time Estimation and
Energy Efficiency
Effect of Bundling of Measurement Data
Next Steps: Multi-Hop Time Synchronization
Conclusions
19 / 62

25. Background
Synchronous
SCFR Method
(IEEE/ACM ToN, 1995)
Periodic Asynchronous
SCFR Method
(IEEE ToC, 2000)
Aperiodic Asynchronous
SCFR Method
(IEEE CL, 2013)
Establishment of
Clock Offset and
Propagation Delay
Duality
(IEEE CL, 2014)
Energy-Efficient
Time Synchronization
Scheme
(IEEE ToC, 2017)
No Common
Network Clock
CBR to
VBR Stream
One-Way to
Two-Way
Communication
20 / 62

26. Conventional Two-Way Message Exchanges I
Master
sHead Node)
Slave
sSensor Node)
Measurement
Interval of Time Sync. si.e., 2-Way Message Exchange)
…
…
Report
Request
Response
Report
Measurement
T1
T2
T4
T3
I Sensor nodes transmit “Request” messages for
synchronization.
I In addition to measurement data packets.
21 / 62

27. Conventional Two-Way Message Exchanges
II
I The sensor node can estimate its clock offset w.r.t. the
head node and synchronize its clock to that of the
head node:
I Clock offset: θ̂ =
(T2 − T1) − (T4 − T3)
2
.
I Propagation delay: ˆ
d =
(T2 − T1) + (T4 − T3)
2
.
22 / 62

28. Reverse Two-Way Message Exchanges I
Master
sHead Node)
Slave
sSensor Node)
Beacon/
Request
sMeasurement)
Report/
Response
T1 T4
T3
T2
d
tm
I Sensor nodes do not transmit any other messages
except “Request/Response” messages.
I If there are no measurement data, sensor nodes just
receive messages.
23 / 62

29. Reverse Two-Way Message Exchanges II
I The head node can estimate the clock offset of the
sensor node, but the sensor node cannot.
I As a result, the information of all sensor node clocks
is centrally managed at the head node.
I “Response” (synchronization) and “Report”
(measurement data) messages can be combined to
save the number of message transmissions from the
sensor node.
I Optionally measurement data and corresponding
timestamps can be bundled together in a
“Report/Response” message when there are no strict
timing requirements.
24 / 62

30. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Hardware and Logical Clock Models
Effect of Clock Skew on Measurement Time
Estimation
Asynchronous Source Clock Frequency Recovery at
Sensor Nodes: One-Way Clock Skew Estimation
Simulation Results
Next Steps: Multi-Hop Time Synchronization 25 / 62

31. Hardware Clock Model
Time Ti of the hardware clock of the ith sensor node at the
reference time t is modeled as a first-order affine function:
Ti(t) = (1 + i)t + θi,
where
I (1 + i) ∈ R+: Clock frequency ratio.7
I θi ∈ R: Clock offset.
7
i is called a clock skew in the literature.
26 / 62

32. Logical Clock Model
Time Ti of the logical clock of the ith sensor node at
hardware clock time Ti(t) is modeled as a piecewise linear
function: For tkt≤tk+1 (k=0, 1, . . .),
Ti
Ti(t)
= Ti
Ti(tk)
+
Ti(t) − Ti(tk)
1 + ˆ
i,k
− θ̂i,k,
where
I tk: Reference time when a kth synchronization occurs.
I ˆ
i,k: Estimated clock skew from the kth
synchronization.
I θ̂i,k: Estimated clock offset from the kth
synchronization.
27 / 62

33. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Hardware and Logical Clock Models
Effect of Clock Skew on Measurement Time
Estimation
Asynchronous Source Clock Frequency Recovery at
Sensor Nodes: One-Way Clock Skew Estimation
Simulation Results
Next Steps: Multi-Hop Time Synchronization 28 / 62

34. Measurement Time Estimation Error:
Conventional Two-Way Message Exchanges
Master
sHead Node)
Slave
sSensor Node)
Measurement
Request
Response
Report
s1
s2≈s3
s4
d
tm
I When Tmd,
∆t̂Conv.
m ∼ Tm × ∆ˆ
i,
where ∆ˆ
i is the clock skew estimation error.
29 / 62

35. Measurement Time Estimation Error:
Reverse Two-Way Message Exchanges
Master
sHead Node)
Slave
sSensor Node)
Beacon/
Request
sMeasurement)
Report/
Response
T1 T4
T3
T2
d
tm
I When Tmd,
∆t̂Rev.
m ∼
Tm
2
× ∆ˆ
i.
30 / 62

36. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Hardware and Logical Clock Models
Effect of Clock Skew on Measurement Time
Estimation
Asynchronous Source Clock Frequency Recovery at
Sensor Nodes: One-Way Clock Skew Estimation
Simulation Results
Next Steps: Multi-Hop Time Synchronization 31 / 62

37. Message Departure and Arrival Times
I Let td(k) (k=0, 1, . . .) be the reference time for the kth
message’s departure from the head node.
I td(k) also denotes the value of the timestamp carried
by the kth message.
I Then the arrival time of the kth message with respect
to the ith sensor node’s hardware clock is given by
ta,i(k) = Ti (td(k)) + d(k) = (1 + i)td(k) + θi + d(k),
where
I d(k): One-way propagation delay in terms of the ith
sensor node’s hardware clock.
32 / 62

38. Maximum Likelihood Estimation:
Conditional Probability Density Function
Given |
Observations (not yet made)
…
33 / 62

39. Maximum Likelihood Estimation: Likelihood
Function
Find s.t.
ℒ | = ∈ ℒ( | )
ℒ |
Observations (already made)
…
34 / 62

40. Cramér-Rao Lower Bound (CRLB)
CRLB provides a lower bound on the variance of
unbiased estimators.
Var
θ̂
≥
1
I(θ)
,
where I(θ) is the Fisher information defined as
I(θ) = E






∂L(θ|x)
∂θ
!2




 .
I An unbiased estimator achieving CRLB is called
(fully) efficient and therefore is the minimum variance
unbiased (MVU) estimator.
35 / 62

41. Joint Maximum Likelihood Estimators
For a white Gaussian delay d(k) with known mean d and
variance σ2
,
θ̂ML
i (k) =
t2
d
· ta,i − td · tdta,i
t2
d
−
td
2
− d,
R̂ML
i (k) =
tdta,i − td · ta,i
t2
d
−
td
2
,
where
I x ,
Pk
j=0
x(j)
k
,
I xy ,
Pk
j=0
x(j)y(j)
k
.
36 / 62

42. Regression through The Origin (RTO) Model
The problem of asynchronous source clock frequency
recovery (SCFR) can be formulated as a linear RTO model
as follows: For k = 1, 2, . . .,
t̃a,i(k) = (1 + i)t̃d(k) + ˜
d(k),
where
I t̃a,i(k),ta,i(k)−ta,i(0),
I t̃d(k),td(k)−td(0),
I ˜
d(k),d(k)−d(0).
37 / 62

43. 38 / 62

44. Cumulative Ratio (CR) Estimator
R̂CR
i (k) =
t̃a,i(k)
t̃d(k)
= Ri +
˜
d(k)
t̃s(k)
,
where
I Ri: Ratio of the ith sensor node hardware clock
frequency to that of the reference clock (i.e., 1+i).
39 / 62

45. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Hardware and Logical Clock Models
Effect of Clock Skew on Measurement Time
Estimation
Asynchronous Source Clock Frequency Recovery at
Sensor Nodes: One-Way Clock Skew Estimation
Simulation Results
Performance of One-Way Clock Skew Estimation
Performance of Measurement Time Estimation and
Energy Efficiency
Effect of Bundling of Measurement Data
Next Steps: Multi-Hop Time Synchronization
Conclusions
40 / 62

46. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Simulation Results
Performance of One-Way Clock Skew Estimation
Performance of Measurement Time Estimation and
Energy Efficiency
Effect of Bundling of Measurement Data
Next Steps: Multi-Hop Time Synchronization
41 / 62

47. Estimated Clock Skews with Gaussian Delays: σ=1 ns
5 10 15 20 25 30 35 40 45 50
Number of Messages
10−21
10−20
10−19
10−18
10−17
10−16
10−15
MSE
RLS
CR
Joint MLE
GMLLE (Two-Way)
LB for CR
CRLB for Joint MLE
LB for GMLLE
42 / 62

48. Estimated Clock Skews with Gaussian Delays: σ=1 µs
5 10 15 20 25 30 35 40 45 50
Number of Messages
10−15
10−14
10−13
10−12
10−11
10−10
10−9
MSE
RLS
CR
Joint MLE
GMLLE (Two-Way)
LB for CR
CRLB for Joint MLE
LB for GMLLE
43 / 62

49. Estimated Clock Skews with AR(1) Delays8
: σ=1 µs
5 10 15 20 25 30 35 40 45 50
Number of Messages
10−14
10−13
10−12
10−11
10−10
MSE
RLS
CR
Joint MLE
GMLLE (Two-Way)
8
ρ = 0.6.
44 / 62

50. Estimated Clock Skews with AR(1) Delays: σ=1 ms
5 10 15 20 25 30 35 40 45 50
Number of Messages
10−8
10−7
10−6
10−5
10−4
MSE
RLS
CR
Joint MLE
GMLLE (Two-Way)
45 / 62

51. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Simulation Results
Performance of One-Way Clock Skew Estimation
Performance of Measurement Time Estimation and
Energy Efficiency
Effect of Bundling of Measurement Data
Next Steps: Multi-Hop Time Synchronization
46 / 62

52. Estimated Frequency Ratio (Sensor Node) and
Measurement Time (Head Node): SI=100 s
-4E-11
-2E-11
0E+00
2E-11
4E-11
Frequency
Difference
[ppm]
Proposed (w/ CR)
Two-Way (w/ GMLLE)
0 500 1000 1500 2000 2500 3000 3500
Time [s]
-1E-02
-8E-03
-6E-03
-4E-03
-2E-03
0E+00
2E-03
4E-03
Measurement
Time
Error
[s]
Proposed (w/ CR)
Two-Way (w/ GMLLE)
Two-Way
47 / 62

53. Estimated Frequency Ratio (Sensor Node) and
Measurement Time (Head Node): SI=1 s
-4E-11
-2E-11
0E+00
2E-11
4E-11
Frequency
Difference
[ppm]
Proposed (w/ CR)
Two-Way (w/ GMLLE)
0 500 1000 1500 2000 2500 3000 3500
Time [s]
-1E-04
-8E-05
-6E-05
-4E-05
-2E-05
0E+00
2E-05
4E-05
Measurement
Time
Error
[s]
Proposed (w/ CR)
Two-Way (w/ GMLLE)
Two-Way
48 / 62

54. Estimated Frequency Ratio (Sensor Node) and
Measurement Time (Head Node): SI=1 ms
-4E-11
-2E-11
0E+00
2E-11
4E-11
Frequency
Difference
[ppm]
Proposed (w/ CR)
Two-Way (w/ GMLLE)
0 500 1000 1500 2000 2500 3000 3500
Time [s]
-1E-06
-8E-07
-6E-07
-4E-07
-2E-07
0E+00
2E-07
4E-07
Measurement
Time
Error
[s]
Proposed (w/ CR)
Two-Way (w/ GMLLE)
Two-Way
49 / 62

55. Effect of SI on Time Synchronization and
Energy Consumption9
Synchronization Skew Estimation Measurement Time
NTX NRX
Scheme MSE Estimation MSE
Proposed
SI=100 s 8.8811E-25 5.8990E-19 100 36
SI=1 s 9.1748E-25 5.4210E-19 100 3600
SI=10 ms 1.0887E-24 4.7684E-19 100 360100
Two-Way with GMLLE
SI=100 s 1.9021E-24 4.7784E-19 136 36
SI=1 s 1.7034E-24 6.1452E-19 3700 3600
SI=10 ms 9.0992E-25 4.0485E-19 360100 360000
Two-Way
SI=100 s
N/A
3.4900E-05 136 36
SI=1 s 3.4564E-09 3700 3600
SI=10 ms 3.3638E-13 360100 360000
9
Estimations are for the samples taken after 360 s (i.e., one tenth of
the observation period) to avoid the effect of a transient period.
50 / 62

56. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Simulation Results
Performance of One-Way Clock Skew Estimation
Performance of Measurement Time Estimation and
Energy Efficiency
Effect of Bundling of Measurement Data
Next Steps: Multi-Hop Time Synchronization
51 / 62

57. Effect of Bundling on Measurement Time Estimation10
0 500 1000 1500 2000 2500 3000 3500
Time [s]
-2.0E-09
-1.0E-09
0.0E+00
1.0E-09
2.0E-09
Measurement
Time
Error
[s]
NBM=1
NBM=2
NBM=5
NBM=10
10
SI = 1 s.
52 / 62

58. Effect of Bundling on Time Synchronization and
Energy Consumption
Synchronization Scheme
Measurement Time
NTX NRX
Estimation MSE
Proposed
NBM = 1 5.4210E-19 100 3600
NBM = 2 5.1116E-19 50 3600
NBM = 5 3.7504E-19 20 3600
NBM = 10 2.6468E-19 10 3600
I In interpreting the results, the following should be
taken into account:
I The bundling increases the length of message
payload.
I The increased message payload also can affect the
frame errors and the number of retransmissions.
53 / 62

59. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Hardware and Logical Clock Models
Effect of Clock Skew on Measurement Time
Estimation
Asynchronous Source Clock Frequency Recovery at
Sensor Nodes: One-Way Clock Skew Estimation
Simulation Results
Performance of One-Way Clock Skew Estimation
Performance of Measurement Time Estimation and
Energy Efficiency
Effect of Bundling of Measurement Data
Next Steps: Multi-Hop Time Synchronization
Conclusions
54 / 62

60. Multi-Hop Extension through Gateways
55 / 62

61. Challenges and Opportunities
I Tradeoff between time-translating and
packet-relaying gateways..
I The multi-hop extension should be implemented
together with a routing protocol.
I As in LEACH protocol11 and its many variations, the
energy efficiency is also critical in the formation of a
hierarchy and the selection of cluster heads (i.e., the
gateway nodes in the multi-hop extension of the
proposed scheme).
11
W. R. Heinzelman et al., Proc. HICSS’00, Jan. 2000, pp. 1–10.
56 / 62

62. Challenges and Opportunities
I Tradeoff between time-translating and
packet-relaying gateways..
I The multi-hop extension should be implemented
together with a routing protocol.
I As in LEACH protocol11 and its many variations, the
energy efficiency is also critical in the formation of a
hierarchy and the selection of cluster heads (i.e., the
gateway nodes in the multi-hop extension of the
proposed scheme).
11
W. R. Heinzelman et al., Proc. HICSS’00, Jan. 2000, pp. 1–10.
56 / 62

63. Challenges and Opportunities
I Tradeoff between time-translating and
packet-relaying gateways..
I The multi-hop extension should be implemented
together with a routing protocol.
I As in LEACH protocol11 and its many variations, the
energy efficiency is also critical in the formation of a
hierarchy and the selection of cluster heads (i.e., the
gateway nodes in the multi-hop extension of the
proposed scheme).
11
W. R. Heinzelman et al., Proc. HICSS’00, Jan. 2000, pp. 1–10.
56 / 62

64. Next . . .
Wireless Sensor Networks
Time and Space in Synchronization
Energy-Efficient Time Synchronization for Asymmetric
Wireless Sensor Networks
Hardware and Logical Clock Models
Effect of Clock Skew on Measurement Time
Estimation
Asynchronous Source Clock Frequency Recovery at
Sensor Nodes: One-Way Clock Skew Estimation
Simulation Results
Performance of One-Way Clock Skew Estimation
Performance of Measurement Time Estimation and
Energy Efficiency
Effect of Bundling of Measurement Data
Next Steps: Multi-Hop Time Synchronization
Conclusions
57 / 62

65. Conclusions
I Propose an energy-efficient time synchronization
scheme for asymmetric wireless sensor networks
achieving sub-microsecond time synchronization
accuracy, which is based on
I Asynchronous SCFR for one-way clock skew
estimation/compensation at sensor nodes;
I Reverse two-way message exchanges for clock offset
estimation/translation at the head node.
I Also, discuss the optional bundling of measurement
data in a “Report/Response” message.
I If there are no strict timing requirements, the
bundling can further reduce the number of message
transmissions without significantly affecting the time
synchronization performance.
58 / 62

66. Conclusions
I Propose an energy-efficient time synchronization
scheme for asymmetric wireless sensor networks
achieving sub-microsecond time synchronization
accuracy, which is based on
I Asynchronous SCFR for one-way clock skew
estimation/compensation at sensor nodes;
I Reverse two-way message exchanges for clock offset
estimation/translation at the head node.
I Also, discuss the optional bundling of measurement
data in a “Report/Response” message.
I If there are no strict timing requirements, the
bundling can further reduce the number of message
transmissions without significantly affecting the time
synchronization performance.
58 / 62

67. Conclusions
I Propose an energy-efficient time synchronization
scheme for asymmetric wireless sensor networks
achieving sub-microsecond time synchronization
accuracy, which is based on
I Asynchronous SCFR for one-way clock skew
estimation/compensation at sensor nodes;
I Reverse two-way message exchanges for clock offset
estimation/translation at the head node.
I Also, discuss the optional bundling of measurement
data in a “Report/Response” message.
I If there are no strict timing requirements, the
bundling can further reduce the number of message
transmissions without significantly affecting the time
synchronization performance.
58 / 62

68. Conclusions
I Propose an energy-efficient time synchronization
scheme for asymmetric wireless sensor networks
achieving sub-microsecond time synchronization
accuracy, which is based on
I Asynchronous SCFR for one-way clock skew
estimation/compensation at sensor nodes;
I Reverse two-way message exchanges for clock offset
estimation/translation at the head node.
I Also, discuss the optional bundling of measurement
data in a “Report/Response” message.
I If there are no strict timing requirements, the
bundling can further reduce the number of message
transmissions without significantly affecting the time
synchronization performance.
58 / 62

69. Conclusions
I Propose an energy-efficient time synchronization
scheme for asymmetric wireless sensor networks
achieving sub-microsecond time synchronization
accuracy, which is based on
I Asynchronous SCFR for one-way clock skew
estimation/compensation at sensor nodes;
I Reverse two-way message exchanges for clock offset
estimation/translation at the head node.
I Also, discuss the optional bundling of measurement
data in a “Report/Response” message.
I If there are no strict timing requirements, the
bundling can further reduce the number of message
transmissions without significantly affecting the time
synchronization performance.
58 / 62

70. Topics of Ongoing and Further Studies I
I Design and implementation of hardware-oriented
multi-hop synchronization schemes.
I Demonstration of the proposed schemes through a
real testbed
I Energy-delay tradeoff and the effect of frame errors
and retransmissions in measurement data bundling.
59 / 62

71. Topics of Ongoing and Further Studies I
I Design and implementation of hardware-oriented
multi-hop synchronization schemes.
I Demonstration of the proposed schemes through a
real testbed
I Energy-delay tradeoff and the effect of frame errors
and retransmissions in measurement data bundling.
59 / 62

72. Topics of Ongoing and Further Studies I
I Design and implementation of hardware-oriented
multi-hop synchronization schemes.
I Demonstration of the proposed schemes through a
real testbed
I Energy-delay tradeoff and the effect of frame errors
and retransmissions in measurement data bundling.
59 / 62

73. Topics of Ongoing and Further Studies II
I Joint time
synchronization
and ranging.
I e.g., drone
networks.
I Indoor localization with wireless
fingerprints based on ANNs
trained by evolutionary algorithms.
I See the next slide for details.
60 / 62

74. Topics of Ongoing and Further Studies II
I Joint time
synchronization
and ranging.
I e.g., drone
networks.
I Indoor localization with wireless
fingerprints based on ANNs
trained by evolutionary algorithms.
I See the next slide for details.
60 / 62

75. Topics of Ongoing and Further Studies II
I Joint time
synchronization
and ranging.
I e.g., drone
networks.
I Indoor localization with wireless
fingerprints based on ANNs
trained by evolutionary algorithms.
I See the next slide for details.
60 / 62

76. Topics of Ongoing and Further Studies II
I Joint time
synchronization
and ranging.
I e.g., drone
networks.
I Indoor localization with wireless
fingerprints based on ANNs
trained by evolutionary algorithms.
I See the next slide for details.
60 / 62

77. (SSID, RSSI)
Building
Floor
Room
(SSID, RSSI)
=
?
Hierarchical Multiclass Classifier
with Flat Loss Function
Flat Multiclass Classifier
with Hierarchical Loss Function
Building,
Floor,
Room
…
…
…

78. Thanks for your attention!
If you have any question, please contact me at
Kyeongsoo.Kim@xjtlu.edu.cn!