Page 55 - 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

Thus, the irst element of the observation vector was set
160
to one, and the rest of the entries were set to zero. Based
on this observation vector, a maximum likelihood estima‑
140
Number of received molecules, y(t) 100 The Reed‑Solomon (RS) coding scheme for MC was intro‑
Detection Threshold
tion was carried out at the receiver.
120
duced in [57] to enhance the error rate performance at
the receiver. In this work, the system using RS coding
80
was shown to achieve better BER than the system which
uses hamming codes. The work in [58] proposed constant
60
weight codes along with a maximum likelihood sequence
40
was non‑coherent, i.e., CSI free detection.
20
r=2 m r=2 m r=2 m r=2 m r=2 m detection scheme at the receiver. This detection scheme
A soft detection scheme using subtraction and genetic am‑
0
0 2 4 6 8 10 pli ier circuits modeled via chemical reactions was pre‑
Time (s)
sented in [59]. In this work, LLR was computed assuming
Fig. 14 – Received signal including noise and ISI for the transmitted se‑ noise as Poisson (or Gaussian) distributed random vari‑
quence [1 1 0 1 0]. The maximum signal is above the threshold in 3rd able. It was found that BER performance under both Pois‑
and 5th‑bit durations leading to incorrect detection. son and Gaussian models was identical thereby justifying
identical to MAP and MMSE detection schemes at a sig‑ the Gaussian assumption. A sequence detection scheme
ni icantly low Signal‑to‑Noise‑Ratio (SNR) range. for OOK modulation has been proposed in [60]. Two‑
layered detection was performed where the irst layer
found the number of bit‑1 in the received sequence of 7‑
In [53], OOK was used at the transmitter and an adap‑
tive threshold‑based detection in presence of noise and bits. A different group of 7‑bit patterns was divided into
ISI was proposed at the receiver. Half of the total re‑ zones based on a logarithmic metric. Zones containing
ceived molecules in the previous bit‑interval was used as more number of bit‑1 had a higher metric. Then the sec‑
the threshold for detection in the current bit‑interval. It ond layer found the locations of bit‑1 in a speci ic zone by
is shown that the ixed threshold scheme does not work using a Pearson correlation metric. This process need not
well since a varying number of molecules (depending on search all the 7‑bit states thereby reducing the complex‑
previous bit transmissions) due to ISI is added in the cur‑ ity.
rent bit‑interval. A transmission scheme similar to [53] On the other hand, a detector based on the derivative of
was also considered in [54]. However, at the receiver, the the received signal considering the OOK‑based transmis‑
authors proposed energy‑based and sampling‑based de‑ sion was proposed in [61]. More speci ically, for detection
tection schemes. For the energy‑based detection scheme, of bit b j at the receiver, all the derivative values within jth
an optimal time‑interval was determined to minimize the time slot of duration T were calculated using the received
error rate at the receiver. For sampling‑based detection, signal y(t), and then the maximum of these derivative val‑
the total number of samples and the sampling point were ues was compared against the ixed threshold ω as shown
optimized to minimize the error rate. below.
{
The Pulse Position Modulation (PPM) scheme was pro‑ ˆ 1 if max y(kT s )−y((k−1)T s ) ≥ ω,
posed in [55] where molecules were released at the be‑ b j = 0 otherwise, T s (3)
ginning and middle of the bit‑interval in case of bit‑1 and
bit‑0, respectively. A novel low complexity MAX detection where jN s ≤ k < (j + 1)N s , sampling time T s is de ined
scheme was proposed where the time at which the maxi‑ as T s = T/N s and N s is the number of samples taken in a
mum concentration occurs is determined. In the case of bit interval. Note that the threshold ω can be determined
bit‑1 and bit‑0, the maximum concentration occurs be‑ empirically, where minimum BER is achieved. Simulation
fore and after the midpoint of the bit‑interval, respec‑ resultsshowedthat theproposeddetectorperformedbet‑
tively. The performance of the MAX detector was close ter than the MAP detector at a high data rate or small bit
to the maximum likelihood detector except for scenarios duration.
that experience signi icant ISI. Further, similar to [55], the
work in [56] also employed PPM modulation scheme at Further, a preprocessing scheme based on higher order
the transmitter. However, in contrast to [55], two asyn‑ derivatives has been proposed in [62]. Note that higher
chronous detectors were proposed in [56], where the irst order derivatives can be used if a high transmission rate
detection scheme approximated the observation vector to is required compared to the irst order derivative used in
a linear weighted quantity and used a maximum likeli‑ [61]. This is because of the fact that the peak time reduces
hood estimate for detection. In the second method, only with an increase in the derivative order. Also, higher or‑
the irst arriving molecule was considered for detection. der derivative processing offers better ISI mitigation since

© International Telecommunication Union, 2021 43