Low-Complex Synchronization Method for
Intra-Body Links in the Terahertz Band

Jorge Torres Gómez, Senior Member, IEEE, Jennifer Simonjan, Member, IEEE, and Falko Dressler, Fellow, IEEE

Abstract—Precision medicine applications supported by
nanotechnologies enforce designing a communication interface
between in-body nanosensors and external gateways. Such a
communication interface will enable both a data and a control
channel between nanodevices operating within the human body
and external control units. In this direction, recent literature
focuses on deriving analytic channel models for intra-body
links through the human tissues, including the analysis of
achievable communication capacities in the terahertz band. A
yet missing component, however, is a synchronization module to
implement communication schemes in the intra-body link. Such
synchronization module will ultimately bound the communication
performance regarding the perceived signal to noise ratio (SNR)
and bit error rate (BER), for instance. This paper contributes to
the state of the art in two directions: (a) evaluating the bounds
on the communication performance with the Cramer-Rao lower
bound (CRLB) for the synchronization symbol timing offset
(STO) and (b) designing a low-complex mechanism to synchronize
communication. This analysis considers a communication link
between external gateways located on the skin and nanosensor
devices flowing in the human vessels. Using envelope and slope
detectors, we devise a low-complex solution that relies on the
received signal strength (RSS) metric to trigger data emissions.
The method estimates the peak of the received RSS metric to
ignite communication in the most favorable location, i.e., when the
nanosensor is located at the shortest distance in the communication
range with external gateways. Our findings illustrate the feasibility
of such a low-complex synchronization method. Performance
−5
illustrates a BER less than 1×10 for those nanosensors traveling
close to the upper vessel wall.
Index Terms—Intra-body terahertz communication, Synchro-

1!!\\times!10^{-5}

Index Terms—Intra-body terahertz communication, Synchronization, Localization, In-body nanosensors

I. INTRODUCTION

DVANCES in nanotechnologies progressively bridge the
Aboundaries between the human body and the digital
world. Battery-free electronics smoothly integrate with the
skin as unnoticeable patches \[1\], integrating processing units
running artificial intelligence (AI) modules, sensors \[2\], and
radio frequency (RF) interfaces \[3\]. Inside the human body,
nanotechnologies also allow radiating information from sensors to the outside through plasmonic antennas in the THz
band \[4\]. Measurements with human-phantoms demonstrate
the communication capabilities through the human tissues \[5\].
Such promising technologies advocate the paradigm of cellconnected-to-internet, where specific body regions are monitored and treated from remote locations, with impactful
Jorge Torres Gómez and Falko Dressler are with the School for Electrical

applications in precision medicine \[6\]. Making use of gateway
systems, very powerful communication and control applications
become possible \[7\], \[8\].
The THz band is becoming a promising means to convey

The THz band is becoming a promising means to convey
information exchange in the nanoscale for precision medicine
applications; see \[9\] for instance. Illustrative designs of RF
components affords miniaturization in the µm size using
graphene antennas \[10\], \[11\]. Numerous research advances
the use of THz band communication, enabling communication
among nanosensors in the human body and with external
devices for monitoring and controlling applications \[12\].
Specifically, for intra-body links (inside to outside the

devices for monitoring and controlling applications \[12\].
Specifically, for intra-body links (inside to outside the
body), recent studies on models render the propagation of
THz waveforms, allowing us to evaluate the communication
performance. These models compute the path loss when rays
undergo through the tissue and define the noise produced by
the surrounding molecules in the link \[13\]. Research also
commends low-complex modulation schemes \[14\] like on-off
keying (OOK) \[4\] and pulse position modulation (PPM) \[15\]
to convey information between the nanonodes.
However, fulfilling the communication exchange in realistic

However, fulfilling the communication exchange in realistic
scenarios requires implementing a synchronization scheme
between emitter and receiver nodes \[12\], \[16\]. In the THz
band, research reports timing acquisition algorithms detecting
the average \[17\] and power levels \[18\] of received pulses, as
well as using the maximum likelihood (ML) criterion with
a training sequence \[19\]. Also, in the spectral domain, the
phase of the incoming symbols is estimated using the Fourier
transform as in \[19\]. Although these examples illustrate the
feasibility of recovering the symbol timing in the THz band,
these solutions are not applied to the particular case of intrabody channels. This requires a more thorough analysis and
implementation when the communication happens through the
human tissues.
Aiming to frame a communication interface in the THz

To implement such a synchronization mechanism, based
Jennifer Simonjan is with Technology Innovation Institute, Masdar City,
on the spatial location of the nanosensor, we perform the
periodic emission of pulses at the gateway and evaluate

Jorge Torres Gómez and Falko Dressler are with the School for Electrical
Engineering and Computer Science, TU Berlin, Berlin, Germany, email: {torresgomez, dressler}@ccs-labs.org.
Jennifer Simonjan is with Technology Innovation Institute, Masdar City,

Jennifer Simonjan is with Technology Innovation Institute, Masdar City,
Abu Dhabi, UAE, e-mail: [Jennifer.Simonjan@tii.ae](mailto:Jennifer.Simonjan@tii.ae).
Manuscript received October XX, 2023; revised YYYY XX, 2023.

Manuscript received October XX, 2023; revised YYYY XX, 2023.

* * *

the backscattered received signal strength (RSS) from the
nanosensor. The method, running at the more powerful gateway
device, conceives a peak detector to evaluate the RSS’s
peak as an indicator of the shortest communication distance
with the nanosensor. The RSS peak detection is used as
a synchronization signal to trigger the data emission from
nanosensors to the gateway.
Following the above, our key contributions can be summa-

Fig. 1. Conceptual representation of nanosensors flowing through the HCS
and a gateway attached to the hand of the person to collect the sensor data.

nanosensors to the gateway.
Following the above, our key contributions can be summarized as follows:
• We design a low-complex synchronization method to
trigger the communication in the intra-body link.
We provide an analytic model describing the randomness

• We provide an analytic model describing the randomness
of the RSS peak detection in the intra-body link.
We evaluate the Cramer-Rao lower bound (CRLB) metric

• We evaluate the Cramer-Rao lower bound (CRLB) metric
to evaluate the minimum variance of the RSS peak
detection.
We evaluate the performance for the proposed method

• We evaluate the performance for the proposed method
with the symbol timing offset (STO) and the impact on
the achievable BER.
Our findings illustrate the feasibility of devising a synchroniza-

II. STATE OF THE ART

Our findings illustrate the feasibility of devising a synchronization signal that triggers communication when the nanosensor
travels near the vessel’s upper wall.
Elaborating on these contributions, the rest of the paper

travels near the vessel’s upper wall.
Elaborating on these contributions, the rest of the paper
is structured as follows. We discuss the state of the art in
Section II. Details on the system model introducing the mobility
of nanosensors and the communication channel model are given
in Section III. In Section IV, we provide insights on the core
idea to devise a synchronization signal with the RSS’s peak
and evaluate the minimum variance of such an estimation with
the CRLB formulation. We detail the proposed low-complex
mechanism in Section V with the corresponding block diagram
and design parameters. Section VI illustrates the performance
of the proposed design with the STO and the BER metrics.
Finally, we summarize open research directions in Section VII
and conclude the work in Section VIII

works on synchronization of emissions with the position of the
nanosensor. In this section, we therefore discuss relevant works
focusing on developing clock recovery methods. Furthermore,
we included comprehensive surveys on THz channel modeling
and synchronization for further reading.
Gupta et al. \[17\] proposed a synchronization scheme for

Gupta et al. \[17\] proposed a synchronization scheme for
pulse-based THz-band communications, which aims at jointly
determining the symbol start time and the observation window
length. This is achieved by successively narrowing down
the integration window to the true location of the pulse.
Furthermore, the authors investigated the symbol error rate, the
synchronization preamble length and the maximum number
of bits to be transmitted before requiring re-synchronization
analytically.
Han et al. \[19\] proposed two timing acquisition algorithms

Han et al. \[19\] proposed two timing acquisition algorithms
for pulse-based communications in the THz band. A low
sampling rate (LSR) and maximum-likelihood (ML) algorithm
are proposed for timing acquisition. They investigate the LSR
algorithm by considering the antenna gain, the distance, the
number of frames per symbol and the pulse width. However,
the proposed LSR works only well if the SNR is significantly
high. The ML approach instead adopts a two-step acquisition
procedure to derive the timing acquisition solutions based
on the ML criterion and works better in low SNR scenarios.
Overall, the authors provide promising angles to efficiently
and reliably solve the timing acquisition problem for pulsebased THz band wireless systems. Furthermore, this work
provides a comprehensive overview of existing synchronization
approaches.
A link-layer synchronization and medium access control

A link-layer synchronization and medium access control
(MAC) protocol for very-high-speed wireless communication
networks in the THz band has been presented by Xia et al.
\[21\]. The proposed protocol relies on a receiver-initiated
handshake as well as a sliding window flow control mechanism
to guarantee synchronization between transmitter and receiver,
maximize the channel utilization and minimize the packet
discard probability.
A novel receiver architecture for pulse-based electromagnetic

discard probability.
A novel receiver architecture for pulse-based electromagnetic
nanonetworks in the Terahertz band has been proposed by Cid-
Fuentes et al. \[18\]. The proposed architecture is based on a
Continuous-time Moving Average (CTMA) symbol detection
scheme and bases its symbol decision on the received signal
power maximum peak after the CTMA, which is implemented
with a single low-pass filter.
Furthermore, there are some recent interesting and com-

with a single low-pass filter.
Furthermore, there are some recent interesting and comprehensive surveys discussing THz technologies, channel
models, modulation schemes, topologies, synchronization and
localization techniques, which we would like to refer the reader
to. For example, Saeed et al. \[22\] presented a study of the THz
band for body-centric networks, by surveying works on THz
device technologies, channel and noise modeling, modulation
schemes, and networking topologies.
Another relevant survey has been published by Lemic et al.

* * *

Chen et al. \[23\] presented an extensive overview of existing
THz systems and localization surveys. Even if the study focuses
on macro-scale setups rather than on the in-body scenario, it
includes important discussions about THz channels and points
to further surveys and works dealing with localization. Finally,
Akyildiz et al. \[16\] reviewed the recent advancements and highlighted the open research directions of THz communications
thoroughly, discussing also synchronization and localization
specifics of the THz band.
In these reported works, a synchronization module for

In these reported works, a synchronization module for
intra-body links is still missing. To design such a module,
we rely on our prior work \[20\], which studies the intrabody communication channel between nanosensors in the
bloodstream and gateways attached to the skin using the THz
spectrum. In \[20\], we focused on developing a communication
scheme that considers the waveform to be travelling through
three layers of skin, tissue, and blood. The main conclusion of
this work was that optimal communication happens when the
nanosensor is below the gateway. This requires the development
of a module to synchronize emissions with the nanosensor’s
location. Contributing to that direction, this work now focuses
on developing and analyzing the performance of a low-complex
mechanism to ignite communications between nanosensors and
the gateway.

III. SYSTEM MODEL

III. SYSTEM MODEL
As for the system model, we consider the intra-body
communication link between nanosensors traveling through the
human vessels and an external gateway located on the skin’s
surface. This system is sketched in Fig. 2, illustrating the
local communication scenario between the nanosensor and the
gateway. The reference coordinates are centered horizontally
with the gateway location and in the ordinate with the upper
vessel wall, where the lv-coordinate defines the vessel stream.
The nanosensor passively moves straight from left to right
along the vessel stream, driven by the blood flow and at a
constant speed vlv, which defines the nanosensor position along
1
the x-axis as x = vlvt.
Besides, we implicitly assume that advection dominates

(1)

l\_{v}

l\_{v}

The nanosensor and the gateway communicate along
three layers through the blood vessels, the fat tissue, and the skin. Communication occurs in the THz
band 0.1 to 10 THz \[13\]. We perform emissions at the center
frequency of fc= 0.14 THz, with the communication bandwidth BW = 40 MHz, and with isotropic antennas. We assume
that the communication distance is at most dmax= 1.5 mm
3
to avoid that much attenuation level \[26\]. Further details on

d\_{\\mathrm{m a x}}=1.5

v\_{l\_{v}}

x=v\_{l,},.

vbegin{}{v\_{l\_{v}};\\gg;\\frac{D}{L\_{\\mathrm{c o m m}}}}\\end{array}

\\left(l\_{v}\\right)

Fig. 2. System model for the communication system between the nanosensor
and the gateway.

these physiological parameters are given in Table I and the
corresponding references \[27\]–\[29\]. As a result, the communication range \[−Lcomm(lv), Lcomm(lv)\] in the direction of
mobility, see Fig. 2, will be determined by

(2)

\\left\[-L\_{\\mathrm{c o m m}}(l\_{v}),,L\_{\\mathrm{c o m m}}(l\_{v})\\right\]

L\_{\\mathrm{c o m m}}\\big(l\_{v}\\big)=\\sqrt{d\_{\\mathrm{m a x}}^{2}-(L\_{v}+L\_{t}+l\_{v})^{2}},

P\_{\\mathrm{T x}}

which is dependent on the nanosensor’s vessel stream lv.
The received signal at the gateway is modeled with the

\\frac\\1{2}

v
The received signal at the gateway is modeled with the
channel path loss and including the impact of additive white
gaussian noise (AWGN) as4
r

r(t)=\\sqrt{\\frac{P\_{\\mathrm{A t t}}}{2}P\_{\\mathrm{T x}}}s(t)+w(t)

P\_{\\mathrm{A t t}}

where s(t) is the emitted sequence, PAttis the power attenuation introduced by the channel in one direction only, and PTx
1
is the transmission’s power. We include the factor as we
2
consider the backscatter signal, i.e., the signal travels from the
gateway to the nanosensor and is backscattered to the gateway
by the nanosensor. Consequently, the received signal will be a
Gaussian variable following the distribution
r P!

\\mathrm{a s}^{4}

r(t)\\sim\\mathcal{N}\\left(\\sqrt{\\frac{P\_{\\operatorname{A t t}}}{2}P\_{\\operatorname{T x}}}s(t),\\sigma^{2}\\right),

t=0

We assume the noise level produced by the human tissue
(molecular absorption noise) is negligible compared to the
noise produced by the receptor circuit. To illustrate this
concern, the noise produced by the human tissue will be less
than the amount kBTmolBW, where kBis the Boltzmann
−4πfc(Lv+Lt+Ls)k/c
constant, Tmol= T0(1 − e), T0 = 310 K

is the reference temperature, k = 0.0072 is the extinction coefficient, and BW is the communication bandwidth. Evaluating

ficient, and BW is the communication bandwidth. Evaluating
the minimum bandwidth (BW), as two times the sampling

the minimum bandwidth (BW), as two times the sampling
2
time, yields BW = = 40 MHz, as for T
Ts
value listed in Table I. To illustrate, with this formulation the
power of the molecular absorption noise results in ∼−140 dB.
On the other hand, the noise level in the receptor circuit
is evaluated with the noise equivalent power metric, which
1/2
is reported as 1 nW/Hz

the minimum bandwidth (BW), as two times the sampling
= 40 MHz, as for Tswe follow the
value listed in Table I. To illustrate, with this formulation the
power of the molecular absorption noise results in ∼−140 dB.
On the other hand, the noise level in the receptor circuit
is evaluated with the noise equivalent power metric, which
1/2
at 0.04 and 0.14 THz \[30\]. As-

1Implicitly, here we assume that t = 0 corresponds to the nanosensor’s
position just below the gateway.
2The nanosensor’s diffusion coefficients will be directly related its size and

\\textsf{t}k\_{B}T\_{\\mathrm{m o l}}\\mathrm{B}

2The nanosensor’s diffusion coefficients will be directly related its size and
shape, see examples in \[24, Fig. 4.5 pag. 57\]
3Distances less than 2 mm happen between on the wrist between the skin

\\sigma^{2}

4Following our previous work in \[20\] we assume transmissions are performed
along the channel coherence time, thereby avoiding the impact of mobility
with the Doppler effect.

3Distances less than 2 mm happen between on the wrist between the skin
surface and the capillaries, for instance.

* * *

the resulting noise level by the circuit is evaluated
√
1/2
as 1 nW/Hz × BW, yielding ∼−62 dB. This value is
comparatively larger than the noise level in the human tissue
(∼−140 dB), allowing us to disregard the molecular absorption
noise.
In Eq. (3), the attenuation power in the human tissues follows

\ 1\\mathrm{n W/H z}^{\\bar{\ }/2}\\times\\sqrt{\\mathrm{B W}}

(\\sim-140,\\mathrm{d B})

In Eq. (3), the attenuation power in the human tissues follows
the predefined equations in \[26, Eq. (2)\], accounting for the
spreading and molecular absorption losses, as

\\begin{aligned}{P\_{\\operatorname{A t t}}=e^{-\\mu\_{v}d\_{v}}\\times\\left(\\frac{\\lambda\_{v}}{4\\pi d\_{v}}\\right)^{2}\\times}&{{}e^{-\\mu\_{t}d\_{t}}\\times\\left(\\frac{\\lambda\_{t}}{4\\pi d\_{t}}\\right)^{2}}\ {\\times}&{{}e^{-\\mu\_{s}d\_{s}}\\times\\left(\\frac{\\lambda\_{s}}{4\\pi d\_{s}}\\right)^{2},}\ \\end{aligned}

(4)

including the attenuation in the three layers (vessel, tissue,
and skin), where µv, µt, µsare the molecular absorption
coefficients, and λv, λt, and λsare the effective wavelengths
5
of the vessel, tissue, and skin, respectively. The variables dv,
dt, and dsaccount for the communication distance between
the nanosensor and the gateway in the vessel, tissue, and skin,
respectively, see their representation in Fig. 2.
These communication distances can be evaluated with the

\\mu\_{v},\ \\mu\_{t},\ \\mu\_{s}

\\lambda\_{v},,\\lambda\_{t}

\\lambda\_{s}

d\_{v}

d\_{t}.

d\_{s}

These communication distances can be evaluated with the
time-varying position of the nanosensor along the x-axis
as given by x = vlvt. Using the cosine function for the
incidence angle of the communication signal ray (φ in Fig. 2),
straightforward calculations evaluate dv, dt, and dsas
l L L

x,=,v\_{l\_{v}}t

(\\varphi)

d\_{v},,d\_{t}

d\_{s}

\\begin{aligned}{d\_{v}=\\frac{l\_{v}}{\\operatorname{c o s}\\varphi},\\quad d\_{t}=\\frac{L\_{t}}{\\operatorname{c o s}\\varphi},:\\mathtt{a n d}:::d\_{s}=\\frac{L\_{s}}{\\operatorname{c o s}\\varphi},}\ {\\operatorname{w i t h}\\quad\\operatorname{c o s}\\varphi=\\sqrt{1-\\bigg(\\frac{v\_{l\_{v}}t}{l\_{v}+L\_{t}+L\_{s}}\\bigg)^{2}},}\ \\end{aligned}

(5)

which is ultimately parametrized with time (t) and the vessel
stream (lv). The term vlvevaluates the speed with the vessel
stream, which follows a parabolic profile as in \[24, Eq. (4.9)\
pag. 54\]
4vm 2
v = (L l − l), (6)

(l\_{v})

v\_{l\_{v}}

(6)

v\_{l\_{v}}=\\frac{4v\_{m}}{L\_{v}^{2}}(L\_{v}l\_{v}-l\_{v}^{2}),

IV. IDENTIFICATION OF OPTIMAL POSITION FOR
COMMUNICATION
As we discussed in our previous work in \[20\], the optimum

where vmdenotes the maximum speed as in the vessel center.
Finally, considering all the above expressions in Equations (5)
and (6), we remark that the path loss in Eq. (4) will ultimately
be a time varying function and dependent on the vessel stream
the nanosensor travels with the variable lv.

l\_{v}

v\_{m}

Fig. 3. Resulting BER and total of emission blocks, see \[20\].

time. In this plot, the nanosensor always emits the first block
just below the gateway, and the remaining ones at shifted
nanosensor’s positions, as driven by the blood flow. Less
transmission bandwidth implies emitting more than one block,
which implies a degraded BER.
On the other hand, increased transmission bandwidth will

(left(l\_{v}\\right)

On the other hand, increased transmission bandwidth will
increase the impact of noise and eventually degrade the
BER. Consequently, the optimum communication performance
is “spatially located” where the communication distance is
minimum, i.e., at φ = 0 in Fig. 2. Besides, this optimal
transmission strategy occurs with less possible transmission
bandwidth that allows emissions to be performed within a
single communication block.
Following this rationale, a strategy to optimize the commu-

\\varphi,=,0

(7)

Following this rationale, a strategy to optimize the communication performance is to estimate the distance between the
traveling nanosensor and the gateway. Then, synchronously
ignite the data emission when this distance is the minimum
along the nanosensor’s path. The literature reports two primary
metrics to estimate the distance with the received signal: the
RSS and the time of flight (ToF). To conceive the less complex
scheme, we analyze the RSS metric and let for future work
the ToF and the combination of both.
With the RSS metric, the minimum distance can be detected
with the maximum power of the received signal evaluated as

l\_{v}

P\_{\\operatorname{R x}}=\\frac{1}{2}P\_{\\operatorname{A t t}}(t;l\_{v})P\_{\\operatorname{T x}}.

As expected, all the peaks occur at x = 0 where the gateway
is located, see Fig. 4a). However, depending on the nanosensor’s
vessel stream, the peak is observed earlier or later with time,
see Fig. 4b). The peak appears earliest when lvapproaches
the vessel center, where the blood speed is the fastest, and it
delays with the decreasing distance to the vessel edges, where
the blood speed approaches zero. Specifically, the time instant

5The molecular absorption coefficients and the effective wavelengths are
evaluated as indicated in \[26, Eq. (3)\] and using the parameters in Table I in
the same reference. Besides, although not directly stated here, PAttin Eq. (4)
is also frequency dependent with the effective wavelength parameters.

* * *

for the RSS’ peaks can be readily evaluated as

t\_{\\mathrm{p e a k}}=\\frac{L\_{\\mathrm{c o m m}}(l\_{v})}{v\_{l\_{v}}},

(8)

as the nanosensor displaces with constant speed along the
vessel stream with Lcomm(lv) as given in Eq. (1). Estimating
the peak location with tpeakin Eq. (8) will readily provide
the synchronization signal needed to ignite the optimum
communication performance.
However, estimating the peak location will inherently en-

However, estimating the peak location will inherently encompass an error, which is lower bounded by the variance
produced by the noise component. The location and amplitude
of the RSS’s peak will be a random variable that depends on
the random vessel stream the nanosensor travels and the noise
that is part of the communication channel. The parameter lvis
a random variable of arbitrary distribution, and the amplitude
of the received signal is also impacted with the AWGN random
noise w(t) as in Eq. (2). In the next Sections, we derive such
a variance with the CRLB metric and illustrate the boundaries
on the communication performance for reported electronic
noise-powers.

l\_{v}

w(t)

A. Minimum STO to estimate the RSS peak location

We evaluate the minimum error for the RSS peak detection
with the STO metric. The STO refers to the time difference
between the detected symbol and the actual arrival time instants \[32\]. This metric evaluates the accuracy of synchronizers
and ascertains communication metrics like the signal to noise
ratio (SNR) and the BER.
As the RSS peak’s location is a random variable, the STO

ratio (SNR) and the BER.
As the RSS peak’s location is a random variable, the STO
will be lower bounded with half of the standard deviation
of tpeak, yielding

(9)

| Parameter | Variable | Value | Reference |
| --- | --- | --- | --- |
| Pulse shape | Rectangular pulse |  |  |
| Pulse duration | T | 1μs |  |
| Pulse period | Tp | 1ms |  |
| Receiver's sensitivity |  | 0.1nW/Hz$^{1/2}$ | \[30\] |
| Center Frequency | fc | 0.14THz | \[30\] |
| Sampling time | Ts | 50ns |  |
| Blood speed in the veins | v | 0.03m/s | \[31\] |
| Skin thickness | Lskin | 76μm | \[27\] |
| Tissue thickness | Ltissue | 1mm | \[28\] |
| Vessel thickness | LVessel | 200μm | \[29\] |

t\_{\\mathrm{p e a k}}

TABLE I
SIMULATION PARAMETERS

f\_{c}

0.1,\\mathrm{{n W}/z^{1/2}}

To evaluate var(tˆpeak) in Eq. (9), we resort to the CRLB
metric, as it assesses the lower bound for unbiased estimators \[33, Chap. 3\]. As tpeakdepends on the vessel stream

\\mathrm{S T O} _{\\mathrm\*{m i n}}=\\frac{1}{2}\\sqrt{\\mathrm{v a r}(\\hat{t}_{\\mathrm{p e a k}})},

\ \_{s}

v

L\_{\\mathrm{s k i n}}

76,\\upmu\\mathrm{m}

\\begin{array}{r l}&{\\boxed l{\ l\_{v}=6\ {\\mathcal{V}}{L\_{v}}\\quad l\_{v}=12\ {\\mathcal{V}}{{L} _{v}}\\quad l l l_{v}=12\ {\\mathcal{V}}{L{} _{v}}\\quad l l_{v}=19%{\\mathcal{V}}{L{L} _{v}}}\ &{\\box l{{\\longrightarrow}}{{l}_{v}}=8.1{\\mathcal{V}}{{L} _{v}}\\quad\ l l_{v}=14\ {\\mathcal{V}}{{6} _{v}}\\mathrm{-----R}c e c i v e r\ \ s{s e n s i t i v i t y}}\ &{\\longmapsto l_{v}=10%{l\_{v}}\\quad l\_{v}=17%{L\\mathcal{V}}{{L}\_{v}}}\\end{array}

200,\\upmu\\mathrm{m}

(a) Received power with the x-coordinate

t\_{\\mathrm{p e a k}}

(10)

Fig. 4. Received power at the gateway with the nanosensor’s position along
the x-direction and with time.

\\textstyle{\\big(\\sqrt{\\frac{P\_{\\mathrm{A t t}}}{2}}P\_{\\mathrm{T x}}}nonumber\\,

parameter lv(see Eq. (8)), we formulate the CRLB for tpeak
in terms of the CRLB for lvas \[33, Eq. (3.16) pag. 37\]

\\begin{array}{l}{{\\sf{n e t e r}}l\_{v}\ (\\mathrm{s e e ~~E q.~~(3)},\ \\mathrm{w e ~~f o r m u l a t e~~ t h e ~~C R L B~~ f r}}\ {{\\sf{m s o f f t t e}}\ \\mathrm{C R E B. ~~f o r}\ l\_{v}\ \\mathrm{a s}\[33\ ,\ q.\ (3.16)\ \\mathrm{p a g.~~ 37}}\ {{\\sf{v a r}}(\\hat{t} _{\\mathrm{p e a k}})\\geq\\mathrm{C R L B}_{t\_{\\mathrm{p e a k}}}}\\end{array}\
\
allowing us to focus on evaluating the lower bound for the\
lvparameter estimation instead. In this equation, the partial\
derivative is solved directly by derivating Eq. (8) after replacing\
its numerator with Eq. (1) and its denominator with Eq. (6).\
The term CRLBlin Eq. (10) refers to evaluating the\
\
The term CRLBlvin Eq. (10) refers to evaluating the\
minimum variance to estimate the vessel stream lvbased on the\
received amplitude. Its calculation becomes straightforward due\
to the dependence of the amplitude sequence with the vessel\
stream, as follows from Equations (4), (5) and (7). To evaluate q\
PAtt\
this term, the amplitude of the received signal ( PTxin\
2\
Eq. (2)) is conveniently isolated when sampling r(t) with the\
6\
emitted pulse period (Tp), and assuming the pulse’s amplitude\
\
(T\_{p}),^{6}\
\
6Without restraining the theoretical calculations for the minimum variance\
of the peak location, we assume such an ideal sampling process with the pulse\
period.\
\
* * *\
\
as one. Sampling Eq. (2) with Tp, the resulting sequence will\
follow the received amplitude level as follows\
\
T\_{p}\
\
r\[n\]=A\[n;l\_{v}\]+w\[n\],\
\
(11)\
\
where\
\
A\[n;l\_{v}\]=\\sqrt{\\frac{P\_{\\mathrm{A t t}}(n T\_{p};l\_{v})}{2}P\_{\\mathrm{T x}}}.\
\
(12)\
\
Eq. (11) yields a constant A\[n; lv\] embedded in a white\
Gaussian noise variable w\[n\], as w\[n\] is the sampling of\
the white Gaussian variable w(t) in Eq. (3). Finally, due to\
Gaussian distribution for A\[n; lv\], we use \[33, Eq. (3.14) pag.\
36\] to evaluate the CRLBlvterm as follows\
\
A\[n;l\_{v}\]\
\
w\[n\]\
\
w\[n\]\
\
A\[n;l\_{v}\]\
\
\\operatorname{C R L B} _{l_{\\tau}}\
\
\\operatorname{C R L B} _{l_{v}}=\\frac{\\sigma^{2}}{\\sum\_{n=0}^{N\_{l\_{v}}}\\left(\\frac{\\partial A\[n;l\_{v}\]}{\\partial l\_{v}}\\right)^{2}},\
\
where Nlvdenotes the total of samples in the time window\
where the received amplitude is above the receptors’ sensitivity\
level. As illustrated in Fig. 4 b), this time window will depend\
on the vessel stream lvas it decreases with lvapproaching the\
vessel center.\
Following this development, we provide a straightforward\
\
N\_{l\_{v}}\
\
l\_{v}\
\
l\_{v}\
\
Following this development, we provide a straightforward\
calculation path for the STOmin. After calculating the partial\
derivative in Eq. (10), and evaluating the CRLBlvwith Eq. (13),\
we can readily compute the right term in Eq. (10). The next\
Section illustrates its calculation with the parameters provided\
in Table I.\
\
\\mathrm{S T O\_{m i n}}\
\
\\mathrm{C R L B} _{l_{\\nu}}\
\
B. Evaluating the CRLB and the resulting SNR\
\
B. Evaluating the CRLB and the resulting SNR\
To illustrate, Fig. 5 depicts the evaluation of Eq. (9) with\
7\
the vessel stream. This plot is derived using the communication and vessels parameters listed in Table I, as well as\
electric parameters in the human body from \[26, Table I\].\
We evaluate the noise power as explained in Section III\
yielding σ2 = −− 62 dB. According to this plot, the STOmin\
results in the order of the ms. Such an STO amount can be\
only produced when the nanosensor travels quite close to the\
vessel walls (in this plot depicted in the range 6 to 10 % of the\
vessel thickness); rapidly increasing to the units of seconds for\
larger lv’s.\
As a result of the non-zero STO in this range, the commu-\
\
As Fig. 6 depicts, the received power becomes comparable to\
the noise level when the vessel stream becomes larger than 7 %,\
as a result of SNR ≈ 0. The impact of the STO rapidly\
increases the gap with the ideal receiver.\
These theoretical results indicate the need to conceive a\
\
7%\
\
\\mathrm{S T O\_{m i l}}\
\
\\mathrm{S N R}\ \\approx\ 0\
\
\\operatorname{S N R} _{\\operatorname{d}}=\\frac{P_{\\operatorname{R x}}(t\_{\\operatorname{p e a k}}+\\operatorname{S T O};l\_{v})}{\\sigma^{2}},\
\
{{{l\_{v}}}S}\
\
\\sigma^{2}=--62,\\mathrm{d B}\
\
Fig. 5. Minimum STO with the vessel stream lv when the noise power\
(13)is σ2 = −62 dB.\
\
l\_{v}\
\
Fig. 6. Perceived SNR with the vessel stream lv when the noise power\
is σ2= −62 dB.\
\
Section introduces a low-complex synchronization scheme to\
illustrate a method that can be implemented in practice. Based\
on this proposal, we later evaluate its limits on the achievable\
communication distance with the BER, as introduced in the\
Results Section.\
\
V. SYNCHRONIZATION METHOD\
\
\\sigma^{2}=-62,.\
\
We realize the synchronization method at the gateway\
due to its more extensive computational capabilities than the\
nanosensor. The gateway emits a pulse train sequence that\
travels into the tissue and measures the RSS metric for the\
bounced signal. The synchronization signal will be generated\
at the gateway with the detection of the RSS’s peak.\
Fig. 7 depicts the synchronization scheme implemented at\
\
\\sigma^{2}=-62,\\
\
7The code to evaluate the CRLB is publicly accessible at [https://github.com/](https://github.com/)\
jorge-torresgomez/Terahertz\_C\_Matlab\_codes\
\
8In this scheme, we omitted to represent the upconversion and downconversion steps. Implicitly, we assume both operations are compensating to each\
others.\
\
* * *\
\
Fig. 7. Reference coordinate system for a nanosensor flowing through a blood vessel and communicating to a gateway outside the body.\
\
The gateway detects the incoming sequence with a matched\
filter to maximize the detection performance \[34\]. The matched\
filter is implemented with a finite impulse response (FIR) filter,\
where the coefficients are the samples of the emitted rectangular\
pulse after normalizing it with the square root of its average\
power. Then, the gateway recovers the envelope at the output\
of the matched filter block exhibiting a peak close to tbelow,\
Fig. 8.\
as sketches Fig. 7 d). Finally, a peak detector module will\
locate the envelope’s peak with the zero crossing time of the\
slope, as represented in Fig. 7 e). The gateway will produce\
magnitude, introducing the least amplitude distortion on the\
the synchronization signal by detecting the zero crossing of\
recovered envelope; see \[36, Appendix B.1\].\
the differentiator’s output.\
\
magnitude, introducing the least amplitude distortion on the\
recovered envelope; see \[36, Appendix B.1\].\
We implement a 12th-order Butterworth filter in the discrete\
\
Fig. 8. Synchronization scheme.\
\
We implement a 12th-order Butterworth filter in the discrete\
domain with the 3 dB cutoff frequency at half the pulse train\
1\
periodicity (fc=), with Tp= 1 ms according to the\
2Tp\
9\
parameters in Table I. Fig. 9 depicts the filter’s magnitude\
and group delay response. As expected, the magnitude results\
flat in the bandpass, and the non-zero group delay introduces\
a time shift in the peak amplitude of less than 5 ms. This time\
shift will introduce additional STO due to the digital signal\
processing (DSP) operation.\
\
(f\_{c},=,{\\textstyle{\\frac{1}{2T\_{p}}}})\
\
\ {\\mathrm{I}}^{\ 9}\
\
9See details on Butterworth filter design in \[36, Sec. 7.1 pag. 442\]\
\
* * *\
\
(a) Magnitude response for the Butterworth filter\
\
(b) Group delay for the Butterworth filter\
\
We implement the peak detector, evaluating the slope at\
the output of the Envelope detector block and comparing it to\
zero. This low complex mechanism is depicted with the two\
blocks in Fig. 8. We implement a differentiator with a delay\
and a subtractor blocks to follow the slope of the envelope.\
The differentiator’s output is compared to zero to produce\
the synchronization signal with the nanosensor’s position. To\
reduce the variability at the output, the differentiator employs\
a delay block equivalent to 20 pulse periods, which in turn\
introduces an additional processing delay of 20 × Tp= 20 ms.\
In total, the DSP chain in Fig. 8 introduces a minimum STO\
of 25 ms.\
Fig. 10 illustrate the functioning of the synchronization\
\
B. Peak Detector Scheme\
\
20\\times T\_{p}=20,\\mathrm{m s}.\
\
Fig. 10. Recovered synchronization signal.\
\
C. Complexity of the Proposed Synchronization Scheme\
\
The proposed scheme for the synchronizer is comprised\
mainly of three blocks: the matched filter, the envelope detector,\
and the peak detector, as depicted in Fig. 8. We evaluate\
complexity regarding the multipliers and adders needed for\
each individual block as summarized in Table II. We count the\
multipliers with the total of coefficients not equal to zero or\
one. For instance, all the matched filter coefficients are equal to\
one (as the reference signal is a rectangular pulse of amplitude\
one, see Fig. 7 a)), thereby its implementation only consists\
of adders. Furthermore, we used the pulse duration T and\
sampling time Tsas given in Table I to compute the matched\
filter total of coefficients, as indicated in Table II.\
\
We remark that the complexity of the proposed scheme\
increases linearly with the symbol time and similarly with the\
sampling frequency. As follows from the “Matched Filter” entry\
in Table II, the total of adders increases linearly with T and\
with the sampling frequency (fs), as Ts=. The hardware\
f1s\
implementation for this scheme will require 23 adders and 19\
multipliers in total. We also notice that the Butterworth filter is\
the most complex block to conceive, which can be reduced by\
implementing a more efficient filter like the elliptic one \[36\]\
or decreasing the filter order, we discuss further remarks on\
this in Section VII.\
\
\\begin{array}{r}{T\_{s}=\\frac{1}{f\_{s}}}\ {\\cdots=\\frac{1}{f\_{s}}}\\end{array}\
\
TABLE II\
COMPLEXITY OF THE PROPOSED SYNCHRONIZATION SCHEME.\
\
| Block | Non-zero coefficients |  | Adders | Multipliers |\
| --- | --- | --- | --- | --- |\
| Matched Filter | $\\frac{T}{T\_{s}}=5$ |  | 4 | 0 |\
| Square block |  |  |  | 1 |\
| Butterworth Filter | 18 |  | 18 | 18 |\
| Differentiator | 2 |  | 1 | 0 |\
| Total |  |  | 23 | 19 |\
\
* * *\
\
Fig. 11. Symbol timing offset curves when detecting the peak with the zero\
interception, and the minimum STO when the noise power is σ2= −86.5 dB.\
\
\\sigma^{2}=-86.5,\\mathrm{d B}.\
\
VI. RESULTS\
\
We evaluate results for the proposed synchronization method\
illustrating the resulting STO metric and its impact on the BER.\
We simulate emissions through the channel model introduced\
in Section III using the reference coordinates in Fig. 2. As\
for the communication and physiological parameters of the\
human tissues, we use the values summarized in Table I and\
in \[26, Table I\]. Besides, we also evaluate the CRLB as the\
benchmark for the STO metric.\
Fig. 11 illustrates the STO metric with the vessel stream\
\
benchmark for the STO metric.\
Fig. 11 illustrates the STO metric with the vessel stream\
the nanosensor travels, where lv∈ (0, Lvessel). We also plot\
the minimum STO using the CRLB formulation, as given in\
Section IV-A. For this evaluation, we use the noise level at\
the input of the peak detector block; see Fig. 8. The resulting\
noise level is approximately −86.5 dB due to the lowpass filter\
operation, as implemented by the envelope detector block; see\
Fig. 9.\
The STO for the peak detector is in the order of the ms as\
\
\\sigma^{2}=-86.5,\\mathrm{d B}\
\
Fig. 9.\
The STO for the peak detector is in the order of the ms as\
the case for the minimum STO. Its declining behavior obeys\
the increased sharpness of the amplitude level with lv(see\
Fig. 4 b)), which produces a more abrupt intersection with zero\
for the slope. Although the STO for the proposed scheme is 5\
to 10 times larger than the optimum solution, its impact on\
the communication performance is quite satisfactory regarding\
the achievable BER. As depicted in Fig. 12, the BER is less\
−5\
than 1×10 units and the SNR larger than 13 dB. These results\
are attainable by the nanosensors traveling at less than 8 % of\
the vessel thickness from the vessel wall.\
To compute the BER, we assume the simple communication\
\
:mathrm B B E Q\\left(\\sqrt{\\mathrm{S N R}\_{\\mathrm{d}}}\\right),\
\
l\_{v},\\in,(0,,L\_{\\mathrm{v e s s e l}})\
\
where we evaluate the noise power as σ2 = −86.5 dB.\
The resulting STO for the proposed method is produced by\
the delay introduced in the processing chain, mainly by two\
\
1!\\times!10^{-5}\
\
Fig. 12. Comparative BER curves versus SNR for the ideal receiver and the\
synchronization scheme with zero-crossing detector when the noise power is\
σ2= −86.5 dB.\
\
{dot\\sigma}^{2}=-86.5,\\mathrm{d B}\
\
8,%\
\
blocks. On the one hand, by the lowpass filter, and on the other\
hand, the delay introduced by the differentiator scheme; see\
Fig. 8. Overcoming this delay, the STO can be further reduced\
when anticipating the RSS’s peak. That can be accomplished\
when comparing the estimated envelope with a positive value\
instead of zero.\
To illustrate the room for performance improvements, Fig. 11\
\
l\_{v}\
\
instead of zero.\
To illustrate the room for performance improvements, Fig. 11\
depicts the resulting STO when increasing the threshold to a\
−17\
positive value; 3×10 in this case. This threshold value was\
manually adjusted after visualizing the computed slope (see\
Fig. 10). Although this positive threshold value is not optimal,\
it still reduces the STO along the vessel stream compared to\
the case where the threshold is zero. A real system application\
will require an adaptive or learning mechanism to anticipate the\
RSS peak with the vessel stream lv; in this way compensating\
for the processing-chain delays.\
In contrast to the CRLB formulation, both mechanisms\
\
VII. FURTHER RESEARCH DIRECTIONS\
\
for the processing-chain delays.\
In contrast to the CRLB formulation, both mechanisms\
sketch a declining behavior for the STO in the range 6.8 to 8 %\
with lv, as depicted in Fig. 11. This behavior obeys the\
increased sharpness with lvfor the RSS amplitude, as illustrated in Fig. 4 b). The narrower the peak, the more abrupt\
the transition with zero, thereby, a more accurate estimate\
of the symbol arrival. Although this is the case with the\
proposed method, more accurate solutions will exhibit the\
same increasing behavior of the CRLB with lv.\
Finally, we also illustrate results in Fig. 13 for the BER\
\
3{\\times}1\\bar{0}^{-17}\
\
During this research, we also identified several open research\
directions. Aiming to advance the receiver’s design further, we\
elaborate on the following items\
Designing the synchronization scheme with the optimal\
\
• Designing the synchronization scheme with the optimal\
complexity: According to the discussion in Section V-C,\
\
l\_{v}.\
\
10Using this formulation, we implicitly assume there is an ideal clock\
recovery mechanism; performance degradation is only produced due to\
transmissions occurring at shifted positions from below the gateway location.\
\
* * *\
\
Fig. 13. Comparative BER curves versus SNR for the ideal receiver and the\
synchronization scheme when using a positive threshold.\
\
the lowpass filter in the Envelope Detector block (see\
Fig. 8) results in the most complex block to implement\
due to the number of multipliers and adders needed. This\
complexity can be reduced at the expense of increased\
variability for the filter’s output signal, which eventually\
will increase the STO when detecting the envelope’s peak.\
However, the envelope’s variability can be reduced by\
increasing the buffer size for the differentiator block,\
although it will also increase the STO. For a given STO,\
it is still pending to minimize the complexity, looking\
at the balance between the lowpass filter’s order and the\
differentiator block’s buffer size.\
• Reducing further the STO adapting the threshold value:\
\
• Reducing further the STO adapting the threshold value:\
Aiming to reduce the STO, the received RSS’s peak\
can be anticipated with a positive threshold value; see\
the discussion in Section VI. This value will depend\
on the vessel stream the nanosensor is traveling in,\
which is generally unknown. However, this threshold can\
be adjusted heuristically by implementing an adaptive\
mechanism with the slope of the received RSS sequence.\
For instance, as the slope increases, the threshold can\
increase to anticipate the RSS’s peak, and it can stop\
increasing with the this slope approaching zero. This way,\
the estimated RSS’s peak can be closer to the actual one.\
• Implementing RADAR signal processing solutions to\
further reduce the STO: Another direction to reduce the\
\
frequency-modulated pulses in \[37, Sec. 4.6\].\
• More realistic models: The antenna beam pattern will\
also impact the perceived power and the synchronization\
performance. This work assumes an isotropic antenna;\
however, in practice, the gateway and nanosensor antennas\
\
implement a radiation pattern with a given communication\
19.5\
direction. With the nanosensor’s mobility in the blood\
vessels, the antennas will not always be oriented to each19\
other, thereby preventing a direct communication link\
18.5\
with the gateway. A model accounting for the impact of\
antennas orientation in the synchronization performance\
will account for more realistic results. Furthermore, with\
smaller nanosensor sizes, it will also vary its radial\
position when traveling in the blood; see simulation\
results in \[38\]. In such a case, diffusion might dominate\
advection, causing the received signal strength variability.\
Such variability will also impact the detection for the\
RSS’s peak, thereby the synchronization performance. A\
model for the impact of nanosensor’s radial mobility in\
the synchronization performance is missing.\
• Synchronization of a cluster of nanosensors with the\
\
the synchronization performance is missing.\
• Synchronization of a cluster of nanosensors with the\
gateway: When considering a cluster of nanosensors,\
some of them will travel closer to the gateway, enabling\
communication, and some others will be limited. In\
this regard, there is a need to account for models that\
evaluate the amount of nanosensors that might travel in\
the communication range of the gateway. It is necessary to\
model their distribution across the vessel’s cross-sectional\
area and devise communication schemes between the\
nanosensors at the farthest distance and those closer\
to the gateway. Recent studies can be used to include\
realistic transport models for the traveling nanosensors in\
the capillaries. As indicated in \[38\], the radial transport\
in the capillary flow highly depends on the nanosensor\
size, which will ultimately determine their communication\
capabilities with the gateway.\
\
peak\
Going well beyond the state of the art, we introduced\
can be anticipated with a positive threshold value; see\
a low-complex synchronization method enabling terahertz\
the discussion in Section VI. This value will depend\
communications via intra-body links. The proposed scheme\
in,\
becomes relevant for interfaces between external gateways and\
which is generally unknown. However, this threshold can\
nanosensors operating in the human body. In addition to the\
be adjusted heuristically by implementing an adaptive\
conceptual design, we analyzed the communication performechanism with the slope of the received RSS sequence.\
mance’s theoretical bounds. Using a Cramer-Rao lower bound\
For instance, as the slope increases, the threshold can\
formulation, the minimum achievable symbol timing offset\
increase to anticipate the RSS’s peak, and it can stop\
(STO) is in the order of the ms when nanosensors travel in the\
increasing with the this slope approaching zero. This way,\
proximity of the vessel walls and for typical noise power values\
the estimated RSS’s peak can be closer to the actual one.\
in the terahertz band. Our findings illustrate the feasibility of\
to\
such a low-complex synchronization method achieving a BER\
further reduce the STO: Another direction to reduce the−5\
less than 1×10 for those nanosensors traveling close to the\
STO is looking at RADAR-based solutions like pulse\
upper vessel wall. Next research direction is to integrate this\
compression techniques. These signal-processing methods\
synchronization method with a realistic communication scheme,\
narrow the envelope of the received signal after detection,\
where a clock synchronization method needs to be implemented.\
thereby improving the peak location with time. Instead of\
In addition, this paper also summarized challenging research\
emitting a sequence of pulses, as illustrated in Fig. 7 b), a\
directions to advance the receiver’s design further and introduce\
frequency-modulated pulse can be emitted to later decode\
more realistic scenarios.\
it with reduced time-variability; see, for instance, linear\
\
The reported research was supported by the project NaBo-\
Com, funded by the German Research Foundation (DFG) under\
grant DR 639/21-2 as well as by the project IoBNT, funded\
\
1!\\times!10^{-5} by the Federal Ministry of Education and Research (BMBF,\
Germany) under grant 16KIS1986K.\
\
REFERENCES\
\[1\] D.-H. Kim, N. Lu, R. Ma, Y.-S. Kim, R.-H. Kim, S. Wang, J. Wu,\
S. M. Won, H. Tao, A. Islam, K. J. Yu, T.-i. Kim, R. Chowdhury,\
M. Ying, L. Xu, M. Li, H.-J. Chung, H. Keum, M. McCormick, P. Liu,\
Y.-W. Zhang, F. G. Omenetto, Y. Huang, T. Coleman, and J. A. Rogers,\
“Epidermal Electronics,” Science, vol. 333, no. 6044, pp. 838–843, Aug.\
2011.\
\[2\] X. T. Zheng, Z. Yang, L. Sutarlie, M. Thangaveloo, Y. Yu, N. A. B. M.\
Salleh, J. S. Chin, Z. Xiong, D. L. Becker, X. J. Loh, B. C. K. Tee,\
and X. Su, “Battery-free and AI-enabled multiplexed sensor patches\
for wound monitoring,” Science Advances, vol. 9, no. 24, Jun. 2023.\
\[3\] Y. Yu, J. Nassar, C. Xu, J. Min, Y. Yang, A. Dai, R. Doshi, A. Huang,\
Y. Song, R. Gehlhar, A. D. Ames, and W. Gao, “Biofuel-powered\
soft electronic skin with multiplexed and wireless sensing for humanmachine interfaces,” Science Robotics, vol. 5, no. 41, Apr. 2020.\
\[4\] J. M. Jornet and I. F. Akyildiz, “Femtosecond-Long Pulse-Based\
Modulation for Terahertz Band Communication in Nanonetworks,”\
IEEE Transactions on Communications, vol. 62, no. 5, pp. 1742–1754,\
May 2014.\
\[5\] P. K. B. Rangaiah, R. L. Karlsson, A. S. Chezhian, L. Joseph, B. Mandal,\
B. Augustine, M. Mani, M. D. Perez, T. Voigt, and R. Augustine,\
“Realization of a Portable Semi-Shielded Chamber for Evaluation of Fat-\
Intrabody Communication,” IEEE Access, vol. 11, pp. 72 743–72 755,\
2023.\
\[6\] I. F. Akyildiz, M. Ghovanloo, U. Guler, T. Ozkaya-Ahmadov, A. F. Sarioglu, and B. D. Unluturk, “PANACEA: An Internet of Bio-NanoThings\
Application for Early Detection and Mitigation of Infectious Diseases,”\
IEEE Access, vol. 8, pp. 140 512–140 523, Jan. 2020.\
\[7\] F. Dressler and S. Fischer, “Connecting In-Body Nano Communication\
with Body Area Networks: Challenges and Opportunities of the Internet\
of Nano Things,” Elsevier Nano Communication Networks, vol. 6,\
pp. 29–38, Jun. 2015.\
\[8\] J. Torres Gómez, J. Angjo, and F. Dressler, “Age of Information-based\
Performance of Ultrasonic Communication Channels for Nanosensor-to-\
Gateway Communication,” IEEE Transactions on Molecular, Biological\
and Multi-Scale Communications, vol. 9, no. 2, pp. 112–123, Jun. 2023.\
\[9\] J. M. Jornet and A. Sangwan, “Nanonetworking in the Terahertz Band\
and Beyond,” IEEE Nanotechnology Magazine, vol. 17, no. 3, pp. 21–31,\
Jun. 2023.\
\[10\] A. Sangwan and J. M. Jornet, “Joint Communication and Bio-Sensing\
With Plasmonic Nano-Systems to Prevent the Spread of Infectious\
Diseases in the Internet of Nano-Bio Things,” IEEE Journal on Selected\
Areas in Communications, vol. 40, no. 11, pp. 3271–3284, Nov. 2022.\
\[11\] A. Abohmra, F. Jilani, H. T. Abbas, A. Alomainy, M. Ur-Rehman,\
M. A. Imran, and Q. H. Abbasi, “Flexible and Wearable Graphene-based\
Terahertz Antenna for Body-Centric Applications,” in International\
Workshop on Mobile Terahertz Systems (IWMTS 2019), Bad Neuenahr-\
Ahrweiler, Germany: IEEE, Jul. 2019.\
\[12\] F. Lemic, S. Abadal, W. Tavernier, P. Stroobant, D. Colle, E. Alarcon,\
J. M. Marquez-Barja, and J. Famaey, “Survey on Terahertz Nanocommunication and Networking: A Top-Down Perspective,” IEEE Journal\
on Selected Areas in Communications, vol. 39, no. 6, pp. 1506–1543,\
Jun. 2021.\
\[13\] J. M. Jornet and I. F. Akyildiz, “Channel Modeling and Capacity\
Analysis for Electromagnetic Wireless Nanonetworks in the Terahertz\
Band,” IEEE Transactions on Wireless Communications, vol. 10, no. 10,\
pp. 3211–3221, Oct. 2011.\
\[14\] S. Tarboush, H. Sarieddeen, M.-S. Alouini, and T. Y. Al-Naffouri,\
“Single- Versus Multicarrier Terahertz-Band Communications: A Comparative Study,” IEEE Open Journal of the Communications Society,\
vol. 3, pp. 1466–1486, 2022.\
\[15\] P. Singh and S.-Y. Jung, “Multilevel Pulse Position Modulation With\
Level Trimming for Electromagnetic Nanocommunications in the\
Terahertz Band,” IEEE Access, vol. 10, pp. 94 158–94 168, 2022.\
\[16\] I. F. Akyildiz, C. Han, Z. Hu, S. Nie, and J. M. Jornet, “Terahertz Band\
Communication: An Old Problem Revisited and Research Directions\
for the Next Decade,” IEEE Transactions on Communications, vol. 70,\
no. 6, pp. 4250–4285, Jun. 2022.\
\[17\] A. Gupta, M. Medley, and J. M. Jornet, “Joint Synchronization and\
Symbol Detection Design for Pulse-Based Communications in the THz\
Band,” in IEEE Global Communications Conference (GLOBECOM\
2015), San Diego, CA: IEEE, Dec. 2015.\
\
by the Federal Ministry of Education and Research (BMBF, \[18\] R. G. Cid-Fuentes, J. M. Jornet, I. F. Akyildiz, and E. Alarcon, “A receiver architecture for pulse-based electromagnetic nanonetworks in the\
Terahertz Band,” in IEEE International Conference on Communications\
(ICC 2012), Ottawa, Canada: IEEE, Jun. 2012.\
\[19\] C. Han, I. F. Akyildiz, and W. H. Gerstacker, “Timing Acquisition\
and Error Analysis for Pulse-Based Terahertz Band Wireless Systems,”\
D.-H. Kim, N. Lu, R. Ma, Y.-S. Kim, R.-H. Kim, S. Wang, J. Wu,\
IEEE Transactions on Vehicular Technology, vol. 66, no. 11, pp. 10 102–\
S. M. Won, H. Tao, A. Islam, K. J. Yu, T.-i. Kim, R. Chowdhury,\
10 113, Nov. 2017.\
M. Ying, L. Xu, M. Li, H.-J. Chung, H. Keum, M. McCormick, P. Liu,\
\[20\] J. Torres Gómez, J. Simonjan, J. M. Jornet, and F. Dressler, “Opti-\
Y.-W. Zhang, F. G. Omenetto, Y. Huang, T. Coleman, and J. A. Rogers,\
mizing Terahertz Communication Between Nanosensors in the Human\
“Epidermal Electronics,” Science, vol. 333, no. 6044, pp. 838–843, Aug.\
Cardiovascular System and External Gateways,” IEEE Communications\
Letters, vol. 27, no. 9, pp. 2318–2322, Sep. 2023.\
X. T. Zheng, Z. Yang, L. Sutarlie, M. Thangaveloo, Y. Yu, N. A. B. M.\
\[21\] Q. Xia, Z. Hossain, M. Medley, and J. M. Jornet, “A Link-Layer Syn-\
Salleh, J. S. Chin, Z. Xiong, D. L. Becker, X. J. Loh, B. C. K. Tee,\
chronization and Medium Access Control Protocol for Terahertz-Band\
and X. Su, “Battery-free and AI-enabled multiplexed sensor patches\
Communication Networks,” IEEE Transactions on Mobile Computing,\
vol. 20, no. 1, pp. 2–18, Jan. 2021.\
Y. Yu, J. Nassar, C. Xu, J. Min, Y. Yang, A. Dai, R. Doshi, A. Huang,\
\[22\] N. Saeed, M. H. Loukil, H. Sarieddeen, T. Y. Al-Naffouri, and\
Y. Song, R. Gehlhar, A. D. Ames, and W. Gao, “Biofuel-powered\
M. S. Alouini, “Body-Centric Terahertz Networks: Prospects and\
soft electronic skin with multiplexed and wireless sensing for human-\
Challenges,” IEEE Transactions on Molecular, Biological and Multi-\
Scale Communications, vol. 8, no. 3, pp. 138–157, Sep. 2022.\
Pulse-Based\
\[23\] H. Chen, H. Sarieddeen, T. Ballal, H. Wymeersch, M.-S. Alouini, and\
Modulation for Terahertz Band Communication in Nanonetworks,”\
T. Y. Al-Naffouri, “A Tutorial on Terahertz-Band Localization for 6G\
IEEE Transactions on Communications, vol. 62, no. 5, pp. 1742–1754,\
Communication Systems,” IEEE Communications Surveys & Tutorials,\
vol. 24, no. 3, pp. 1780–1815, 2022.\
P. K. B. Rangaiah, R. L. Karlsson, A. S. Chezhian, L. Joseph, B. Mandal,\
\[24\] H. C. Berg, Random Walks in Biology. Princeton University Press,\
B. Augustine, M. Mani, M. D. Perez, T. Voigt, and R. Augustine,\
1993.\
“Realization of a Portable Semi-Shielded Chamber for Evaluation of Fat-\
\[25\] V. Jamali, A. Ahmadzadeh, W. Wicke, A. Noel, and R. Schober,\
Intrabody Communication,” IEEE Access, vol. 11, pp. 72 743–72 755,\
“Channel Modeling for Diffusive Molecular Communication - A Tutorial\
Review,” Proceedings of the IEEE, vol. 107, no. 7, pp. 1256–1301, Jul.\
I. F. Akyildiz, M. Ghovanloo, U. Guler, T. Ozkaya-Ahmadov, A. F. Sar-\
2019.\
ioglu, and B. D. Unluturk, “PANACEA: An Internet of Bio-NanoThings\
\[26\] J. Simonjan, B. D. Unluturk, and I. F. Akyildiz, “In-body Bionanosensor\
Application for Early Detection and Mitigation of Infectious Diseases,”\
Localization for Anomaly Detection via Inertial Positioning and THz\
Backscattering Communication,” IEEE Transactions on NanoBioscience,\
F. Dressler and S. Fischer, “Connecting In-Body Nano Communication\
vol. 21, no. 2, pp. 216–225, Apr. 2022.\
with Body Area Networks: Challenges and Opportunities of the Internet\
\[27\] G. Piro, P. Bia, G. Boggia, D. Caratelli, L. A. Grieco, and L. Mescia,\
of Nano Things,” Elsevier Nano Communication Networks, vol. 6,\
“Terahertz electromagnetic field propagation in human tissues: A\
study on communication capabilities,” Nano Communication Networks,\
J. Torres Gómez, J. Angjo, and F. Dressler, “Age of Information-based\
vol. 10, pp. 51–59, Dec. 2016.\
Performance of Ultrasonic Communication Channels for Nanosensor-to-\
\[28\] P. Oltulu, B. Ince, N. Kokbudak, S. Findik, and F. Kilinc, “Measurement\
Gateway Communication,” IEEE Transactions on Molecular, Biological\
of epidermis, dermis, and total skin thicknesses from six different body\
and Multi-Scale Communications, vol. 9, no. 2, pp. 112–123, Jun. 2023.\
regions with a new ethical histometric technique,” Turkish Journal of\
J. M. Jornet and A. Sangwan, “Nanonetworking in the Terahertz Band\
Plastic Surgery, vol. 26, no. 2, p. 56, 2018.\
and Beyond,” IEEE Nanotechnology Magazine, vol. 17, no. 3, pp. 21–31,\
\[29\] M. Fruchard, L. Arcese, and E. Courtial, “Estimation of the Blood\
Velocity for Nanorobotics,” IEEE Transactions on Robotics and\
A. Sangwan and J. M. Jornet, “Joint Communication and Bio-Sensing\
Automation, vol. 30, no. 1, pp. 93–102, Feb. 2014.\
With Plasmonic Nano-Systems to Prevent the Spread of Infectious\
\[30\] A. Rogalski, “Graphene-based materials in the infrared and terahertz\
Diseases in the Internet of Nano-Bio Things,” IEEE Journal on Selected\
detector families: a tutorial,” Advances in Optics and Photonics, vol. 11,\
Areas in Communications, vol. 40, no. 11, pp. 3271–3284, Nov. 2022.\
no. 2, p. 314, May 2019.\
A. Abohmra, F. Jilani, H. T. Abbas, A. Alomainy, M. Ur-Rehman,\
\[31\] A. C. Guyton and M. E. Hall, Guyton and Hall Textbook of Medical\
M. A. Imran, and Q. H. Abbasi, “Flexible and Wearable Graphene-based\
Physiology, 14th ed. Elsevier, 2015.\
Terahertz Antenna for Body-Centric Applications,” in International\
\[32\] M. S. Chaudhari and S. Majhi, “STO Estimation for OFDM System\
Workshop on Mobile Terahertz Systems (IWMTS 2019), Bad Neuenahr-\
Using CDM,” IEEE Communications Letters, vol. 26, no. 11, pp. 2651–\
2655, Nov. 2022.\
F. Lemic, S. Abadal, W. Tavernier, P. Stroobant, D. Colle, E. Alarcon,\
\[33\] S. Kay, Fundamentals of Statistical Signal Processing: Estimation\
J. M. Marquez-Barja, and J. Famaey, “Survey on Terahertz Nanocom-\
Theory. Upper Saddle River, NJ: Prentice Hall, 1998.\
munication and Networking: A Top-Down Perspective,” IEEE Journal\
\[34\] B. Sklar, Digital Communications: Fundamentals and Applications,\
on Selected Areas in Communications, vol. 39, no. 6, pp. 1506–1543,\
2nd ed. Upper Saddle River, NJ: Prentice Hall, 2001.\
\[35\] A. B. Carlson, P. B. Crilly, and J. C. Rutledge, Communication Systems:\
J. M. Jornet and I. F. Akyildiz, “Channel Modeling and Capacity\
An Introduction to Signals and Noise in Electrical Communication,\
Analysis for Electromagnetic Wireless Nanonetworks in the Terahertz\
4th ed. New York City, NY: McGraw-Hill, 2002, p. 850.\
Band,” IEEE Transactions on Wireless Communications, vol. 10, no. 10,\
\[36\] A. V. Oppenheim and R. W. Schafer, Discrete-time signal processing,\
3rd ed. Upper Saddle River, NJ: Prentice Hall, 2010.\
S. Tarboush, H. Sarieddeen, M.-S. Alouini, and T. Y. Al-Naffouri,\
\[37\] M. A. Richards, Fundamentals of Radar Signal Processing, 2nd ed.\
“Single- Versus Multicarrier Terahertz-Band Communications: A Com-\
New York City, NY: McGraw-Hill Education, 2014, p. 894.\
parative Study,” IEEE Open Journal of the Communications Society,\
\[38\] T.-R. Lee, M. Choi, A. M. Kopacz, S.-H. Yun, W. K. Liu, and P.\
Decuzzi, “On the near-wall accumulation of injectable particles in the\
P. Singh and S.-Y. Jung, “Multilevel Pulse Position Modulation With\
microcirculation: smaller is not better,” Scientific Reports (Sci. Rep.),\
the\
vol. 3, no. 1, Jun. 2013.\
\
* * *\
\
Jorge Torres Gómez received the B.Sc., M.Sc., and Ph.D. degrees from the Technological Uni- versity of Havana, CUJAE, Cuba, in 2008, 2010, and 2015, respectively. He is currently with the Telecommunication Networks Group, Department of Telecommunication Systems, Technical University of Berlin. He holds a position as head of educational activities at the IEEE Germany Section, supporting teaching activities and chairing conference events in education. Since 2008, he has been with the School of Telecommunications and Electronics, CUJAE University, where he was a Lecturer from 2008 to 2018. He has been with the Department of Signal Theory and Communications, Carlos III University of Madrid, Leganés Campus, Madrid, Spain, as a guest lecturer, and with the Department of Digital Signal Processing and Circuit Technology, Chemnitz University of Technology as a postdoc. His research interests include molecular communications, nanonetworks, the age of information, and digital signal processing.\
\
Jennifer Simonjan received her Ph.D. degree in In- formation and Communication Engineering from the University of Klagenfurt, Austria in 2019. After her Ph.D, she gathered experience working as a post doc at the Georgia Institute of Technology in Atlanta, US and as a researcher in Austria. Currently, she works as a lead researcher in the Autonomous Robotics Re- search Center at the Technology Innovation Institute in Abu Dhabi, UAE. Her research interest comprise self-organization and communication in nanoscale and macroscale sensor networks as well as robotic networks.\
\
Falko Dressler is full professor and Chair for Data Communications and Networking at the School of Electrical Engineering and Computer Science, TU Berlin. He received his M.Sc. and Ph.D. degrees from the Dept. of Computer Science, University of Erlangen in 1998 and 2003, respectively. Dr. Dressler has been associate editor-in-chief for IEEE Trans. on Mobile Computing and Elsevier Computer Communications as well as an editor for journals such as IEEE/ACM Trans. on Networking, IEEE Trans. on Network Science and Engineering, Elsevier Ad Hoc Networks, and Elsevier Nano Communication Networks. He has been chairing conferences such as IEEE INFOCOM, ACM MobiSys, ACM MobiHoc, IEEE VNC, IEEE GLOBECOM. He authored the textbooks Self- Organization in Sensor and Actor Networks published by Wiley & Sons and Vehicular Networking published by Cambridge University Press. He has been an IEEE Distinguished Lecturer as well as an ACM Distinguished Speaker. Dr. Dressler is an IEEE Fellow as well as an ACM Distinguished Member. He is a member of the German National Academy of Science and Engineering (acatech). He has been serving on the IEEE COMSOC Conference Council and the ACM SIGMOBILE Executive Committee. His research objectives include adaptive wireless networking (sub-6GHz, mmWave, visible light, molecular communication) and wireless-based sensing with applications in ad hoc and sensor networks, the Internet of Things, and Cyber-Physical Systems.