Page 52 - ITU Journal Future and evolving technologies Volume 2 (2021), Issue 3 – Internet of Bio-Nano Things for health applications

ITU Journal on Future and Evolving Technologies, Volume 2 (2021), Issue 3

in the presence of enzymes was derived. Moreover, it was
Healthy cells
shown that ISI, as well as useful signals, reduce by using
enzymes. For analysis purposes, the number of received
Cancer cell molecules was shown to be binomial distributed and its
Red blood cell
approximation as Poisson distribution was shown to be
Fusion more accurate than the Gaussian distribution when cal‑
center culating the detection probability at the receiver.
Mobile
nanosensor Similar to [24], the work in [42] also used an OOK mod‑
Cancer
biomarker ulation scheme. However, the number of molecules re‑
ECF leased by the transmitter was modeled as a random vari‑
able whose mean was known to both transmitter and re‑
Fig. 8 – Cooperative MC system of mobile nano‑sensors and fusion cen‑ ceiver. For signal detection at the receiver, a one‑shot
ter using biomarker concentration inside blood vessels for detecting
cancer cell. detector with three different signal processing schemes
based on sampling, correlation (multiplying the received
mitted in the jth bit interval. The MAP sequence detection concentration by a correlation function which was con‑
corresponding to the transmitted bit sequence of length stant or based on MF), and ISI cancellation (the difference
N + 1 can be done according to the following equation between the peak concentration or concentration at sym‑
bol end time whichever is minimum and the concentra‑
ˆ ˆ 0 0 0 = argmax f(y y y ,b b b ), (1) tion at starting time of the received concentration) were
N N N
N N N
ˆ N N N
b b b
0 0 0
0 0 0
N
N
b b 0 0 b 0 N proposed. Each of the schemes results in a different test
statistic. In this work, the ISI cancellation approach was
ˆ ˆ
ˆ N N N
where b b b represents the estimated sequence (of length
0 0 0 found to perform better than the schemes based on sam‑
N N N
N N N
N N N
N + 1) for the transmitted sequence b b b , and f(y y y ,b b b )
0 0 0 0 0 0 0 0 0 pling and correlation techniques.
is the joint PDF of the received signal samples y y y N N N and the
0 0 0
N N N
transmitted bits b b b , which can be expressed as In [43], a transmission scheme similar to Release Time
0 0 0
Shift Keying (RTSK) was proposed to send a pattern of
N N
∏ ∏ molecules based on different time‑instants. For example,
N N N
N N N
f(y y y ,b b b ) = P(b k ) f(y k |b b b k k k k−I ), (2) the pattern δ t,1 + δ t,1.5 + 3δ t,5 can be used for transmis‑
k−I
k−I
0 0 0
0 0 0
k=0 k=0 sion of bit‑1. As shown in Fig. 11, this pattern implies
that one molecule is sent at t = 1 s and t = 1.5 s time‑
if b k ’s are independent, I is the ISI length, P(b k ) is the PMF
of the transmitted bit b k , and f(y k |b b b k k k ) denotes the con‑ instant, and three molecules are sent at t = 5 s time‑
k−Ik−I instant. In this work, transmission, diffusion of molecules,
k−I
ditional PDF of y k .
and reception were modeled using the Continuous‑Time
Further, in [40], a linear equalizer based on the Mini‑ Markov Process (CTMP). For the reception, a MAP demod‑
mum Mean Square Error (MMSE) criterion and a non‑ ulator has been employed which maximizes the posterior
linear Decision Feedback Equalizer (DFE) were also pro‑ probability that a symbol was sent given the history of
posed. Performance was evaluated for time‑invariant and ligand‑receptor complexes formed at the receiver.
3
time‑varying channels. It is shown that the maximum
Further, the bit transmission based on the release time of
likelihood‑baseddetectorperformedbetterthanDFE,and
the molecule was proposed in [44]. In [44], three differ‑
the MMSE equalizer performed the worst among all other
detectors. On the other hand, different concentration ent detection schemes based on the maximum likelihood
levels of molecules for binary and quaternary Amplitude detector, linear detector, and the detector based on the
Shift Keying (ASK) transmission schemes have been pro‑ irst arrival were considered. The propagation time of a
posed in [41]. At the receiver, a signal detection rule based particle was assumed to be Lévy distributed for a pure‑
on NP tests has been derived for both schemes. Perfor‑ diffusive channel and inverse Gaussian (IG) for a low‑
mance of both detectors was evaluated with increasing ISI induced diffusive channel. The performance of each de‑
and BER was found to be lower in the case of binary ASK tection scheme was evaluated for different propagation
than the quaternary ASK transmission. pro iles where arrival time is modeled as (i) uniformly, (ii)
exponentially, (iii) IG, and (iv) Lévy distributed random
A method of ISI mitigation using enzymes in the environ‑ variable. Further, it is shown that the detection based
ment was presented in [24]. For system setup, OOK at on the irst arrival achieved the performance close to the
the transmitter and a single sample‑based detector with maximum likelihood‑based detection if the noise density
4
a ixed threshold at the receiver were employed. In this has zero mode .
work, a lower bound on the number of received molecules
Further, for reducing ISI, a novel transmission scheme
3 Two different cases of time‑varying diffusion coef icient were selected: based on Molecule Transition Shift Keying (MTSK) was
2
i) D(t)=2.2×10 −9 +0.8×10 −9 cos(2πt) m /s, ii) D(t)=2.2×10 −9 +
2
0.8 × 10 −9 cos(10πt) m /s. The channel variation is 5 times faster in 4 The value of a random variable where the PDF is maximum is de ined
the latter case. as the mode.
40 © International Telecommunication Union, 2021