Ma wcnc 2018 slides

IHEHGBDGB LDJC G JK E H @E , . , : O , O , , [email protected]@ C IBDGB O . * * , * M * , * * * , , N , * , Outl...

1 downloads 86 Views 15MB Size
IHEHGBDGB LDJC G JK E H @E ,

. ,

:

O

,

O ,

,

[email protected]@ C IBDGB O

.

*

*

, * M

* , * *

*

,

, N

,

* ,

Outline • Introduction

• Nonlinear superposition • New problem in wireless charging

• Algorithm design

• Related works • Weight-Greedy Picking

• Experiments • Conclusion and future works

Nonlinear Superposition • Radio interference occurs • Two chargers may strengthen or weaken each other

Nonlinear superposition • Charging model • !"# = '(*

%& cos(3# 4 +, ⁄-.&

:

+ 6# − 29;?|A = ? ∫[D# (4)]F G. = > ∑I+∈A : K L + > ∑I+∈A ∑IM ∈AIM NI+ * +,

:+, P:M, O cos(29 ) ; +, *M,

The difference between these two charging model

Our problem in wireless charging • Variable: scheme chargers on/off state in every charging period. • Objective: minimize charging periods. • Constraint: charge each sensor with energy no less than E.

Related work • Computation ahead Have to calculate the charging utility of each charger set at every sensor node in advance. As a result, the complexity of this step grows exponentially with the number of chargers, making it much more complex than previously thought.

Weight-Greedy Picking First find a MCS: pick charger with the maximum weight, and the remove chargers and repeat; Add new chargers: add the charger which has the most effective

A simple Example S1

S2

S3

C1

3

0

0

C2

2

3

2

C3

0

3

3

C4

0

0

2

C2,C3

2

0

1

C3,C4

0

3

5

C1,C2

4

3

2

C2,C4

4

3

0

C2,C3,C4

2

0

2

An example of a MCS

Experiments

Conclusion and Future Work • Nonlinear superposition • Adjustable initial phases • Different duration of each charging period

! [email protected]