Network Coding for Fault-Tolerant Transmission of

Biomedical Data

Sultan Alam Sehore, Madhya Pradesh 46114, India Priyanshu Paul

School of Computing Science and [email protected] School of Computing Science and
Artificial intelligence Artificial intelligence
VIT Bhopal University, Bhopal-Indore VIT Bhopal University, Bhopal-Indore
Highway, Kothrikalan Highway, Kothrikalan
Sehore, Madhya Pradesh 46114, India Sehore, Madhya Pradesh 46114, India
[email protected] [email protected]
Ishmeet Kaur
School of Computing Science and
Artificial intelligence Shaivyaa Sharma
VIT Bhopal University, Bhopal-Indore School of Computing Science and
Highway, Kothrikalan Artificial intelligence
Ayush Kamal Sehore, Madhya Pradesh 46114, India VIT Bhopal University, Bhopal-Indore
School of Computing Science and [email protected] Highway, Kothrikalan
Artificial intelligence Sehore, Madhya Pradesh 46114, India
VIT Bhopal University, Bhopal-Indore [email protected]
Highway, Kothrikalan

Abstract—This paper presents a new approach to the fault- in reaching the overall picture of a patient’s health
tolerance enhancement of the biomedical data transmission condition, hence, enabling accurate assessments, constant
using the network coding methods. In the context of healthcare, surveillance, and medical aid from a distance. Nonetheless,
even small amounts of loss or time delays in data can have the transfer of this data which is crucial is faced with a
severe repercussions, particularly in situations such as number of limitations. As this information is transferred
remotely supervised medical procedures or emergency care. from one point to another using radio waves and the
The majority of traditional error correction techniques do not internet, several issues such as packet loss, network
cope well with adverse network conditions, which extends
congestion, interference, and latency are met. The presence
transmission times and results in the loss of data packets. Using
network coding will help in implementing the technique in
of such problems can degrade the quality of the transmitted
which data packets are encoded such that an original data can data and reduce the reliability of the information often
be accessed even when some packets are corrupted or lost leading to delays in readings or incorrect ones which may
because the packets were sent over different paths. As a result disadvantage patients in critical areas like emergencies.
the need for retransmission is not as crucial for the system, as it Thus, the medical systems for transporting the biomedical
previously was for many approaches. Therefore a system in its information must be fault tolerant and efficient to ensure
entirety becomes more efficient for real-time biomedical data prevention of information loss or corruption and
transmission. Our approach is demonstrated in simulations to maintenance of integrity of the information associated with
have improved fault tolerance by 20% in comparison to the data set even in bad network environment.
conventional techniques. This allows the approach to be more
robust in variable network environments. Additionally, the Importance is increasingly placed on the resilience and
approach is scalable, supporting various types of biomedical dependability of the transmission of biometric data.
data, from basic vital signs to complex signals like ECG and Increasingly, the integration of the IoMT, remote healthcare
EEG. This scalability makes it suitable for a wide range of systems including telemedicine, raises the demands on the
healthcare applications, ensuring reliable, uninterrupted data transmission of biomedical data. Fault tolerant transmission
transmission in both small-scale and large-scale medical methods are therefore critical in the timely and precise
systems. Our research offers a more effective and reliable transmission of data to healthcare providers. Health care is
solution for biomedical data transmission, addressing the particularly at risk in the event that data transmission is not
limitations of existing methods and contributing to future reliable. A sliver of a second latency or even a minuscule
advancements in healthcare technology. variation of data could cause a false diagnosis, delay the
time when a patient needs to be treated, or administer the
Keywords—network coding, data, healthcare, fault tolerance, patients’ care erroneously, all of which are life threatening.
biomedical.
To tackle such issues, conventional data transmission
methods such as Automatic Repeat Request Protocol (ARQ)
and Forward Error Correction (FEC) have been used.
I. INTRODUCTION However, such techniques are unable to meet the demanding
This In the modern healthcare system, the two-way and requirements needed in the current modern health care
instantaneous transfer of biomedical information has surroundings. In many instances, these result in
emerged as one of the pillars of the care offered to patients, unacceptable lags or carry high computation loads, making
thus warranting key procedures. Whether it is the heart them less viable for real-time biometric data transmission.
frequency or the level of sugar in the body or the pressure in Challenges in Biomedical Data Transmission:
the blood vessels, health care professionals have to make
effective and timely decision based on the transmission of ● Time Sensitivity of Biomedical Data: Biomedical
this critical information. These biomedical signals are used data, such as electrocardiograms (ECG), blood pressure
readings, and glucose levels, are often time-sensitive and Network coding is one of the most promising techniques
need to be transmitted in real time to allow medical that can be used to handle drawbacks of conventional data
professionals to act promptly. Even slight delays in data transfer methods. It offers a new aspect of fault tolerance
transmission could lead to serious misinterpretations, wherein packets are combined and then transmitted in place
delayed interventions, or compromised patient care. In of transmitting individual packets. These encoded packets
emergency care situations, these delays can have fatal can make their ways through multiple paths on the network,
consequences. so if some of the packets are lost or corrupted during their
journey, their destination can still recover the original data
without retransmissions.
● Data Integrity: Data integrity is paramount in
healthcare. Any loss or corruption of biomedical data during
transmission could result in incorrect diagnoses or ● Better Reliability: Linear combinations have been
inappropriate treatments. In scenarios such as remote used in the encoding of packets, as a result, this makes an
monitoring or telemedicine, where doctors rely on accurate advancement in the reliability of data transmission.
data to monitor patients, compromised data could lead to a Redundancy information offered by coded packets enables a
failure in detecting critical changes in a patient's condition. recipient to recreate the lost original data when some
packets get lost. This lowers the chances for the loss of data
● Transmission Over Wireless Networks: Biomedical
and guarantees the right transmission of critical medical
data is often transmitted through wireless networks, which
information.
are prone to issues such as packet loss, congestion,
interference, and high latency. These problems are ● Reduced Latency: Unlike ARQ that waits to get a
exacerbated in environments with poor network lost packet and retransmits them back—network coding
infrastructure, such as rural or remote areas. In such settings, waits for no man; it moves on. This quality is very good for
biomedical data may not be transmitted reliably, increasing applications in healthcare, especially where the time of
the risk of medical errors. getting some critical data can be very adverse. Because it is
possible to reconstruct the data even if some packets are
● Traditional Solutions' Limitations:
lost, it becomes a way of enabling faster and more reliable
○ Automatic Repeat Request (ARQ): ARQ works by transmission.
retransmitting lost packets, but the delay introduced by
● Scalability: A growing amount of biomedical data
waiting for retransmissions can be unacceptable in
is produced when more devices and sensors are integrated
healthcare applications. This method is particularly
into evolving healthcare systems. Since network coding is
problematic in time-sensitive scenarios such as intensive
fundamentally scalable, it aligns well with the increasing
care units or emergency diagnostics, where any delay in
needs of telemedicine, as well as remote monitoring and
receiving critical information could impact patient
IoMT technology. It offers an adaptable and effective way to
outcomes.
convey enormous volumes of data without trading away
○ Forward Error Correction (FEC): FEC transmits reliability or imposing more computational burden.
redundant data alongside the original packets, allowing the
In summary, this paper explores the fault-tolerant
receiver to correct any errors without requesting a
transmission of biomedical data using network coding,
retransmission. However, this approach can be
which is crucial for ensuring real-time and reliable
computationally expensive and may lead to increased
healthcare services. The growing complexity in healthcare
processing time, especially in real-time applications with
data transmission, particularly in critical care and remote
large volumes of data. In biomedical systems, where speed
monitoring, demands scalable and reliable methods.
and efficiency are critical, the overhead introduced by FEC
may not be sustainable. Section 1: Related Works reviewed various prior
research efforts on fault-tolerant biomedical data
Need for Fault-Tolerant Transmission:
transmission techniques. Traditional methods such as
With such challenges, it is clear that older techniques Automatic Repeat Request (ARQ) and Forward Error
like ARQ and FEC are insufficient in resolving the issues Correction (FEC) have been widely used but struggle to
associated with contemporary biomedical data transmission. meet the real-time transmission demands of modern
Healthcare systems are in need of such an approach which is healthcare systems. These methods often introduce
not only effective but also scalable to meet the needs of unacceptable delays and computational overhead, especially
timely and accurate transmission of data despite network in critical applications like emergency monitoring and
disturbance or loss of packets. Especially, with the growing telemedicine. Previous works highlighted the limitations of
implementation of medical devices, wearables and IoMT, these techniques in handling dynamic network environments
the requirement for such transmission techniques has and provided a foundation for exploring more efficient
become more severe than ever. For instance, in a critical solutions.
care scenario where time is of the essence, any interruption
Section 2: Proposed Work introduced a novel solution
in the path of biomedical data may deny the doctor with
leveraging random linear network coding. This technique
vital patient information, leading to inappropriate and
encodes data into linear combinations, allowing the system
possibly delayed patient treatment. In a similar context, for
to recover lost information without retransmission. The
long term monitoring of a patient, uninterrupted and steady
mathematical model underlying network coding was
transmission of data is crucial in looking into patterns and
explained, showcasing how redundancy across multiple
spotting areas when the condition of the patient changes
network paths improves fault tolerance. This section also
which demands for prompt action.
detailed the algorithmic design, focusing on its scalability
and minimal computational overhead, making it suitable for
real-time applications.
Network Coding as an Advanced Solution:
 Section 3: Results and Discussion presented a series of workloads and possible system outages. It highlights how to
simulation results that demonstrated how network coding strengthen system resilience against hardware by
outperforms traditional methods in handling packet loss and implementing fault tolerance techniques like data replication
network congestion. Our results showed a significant and redundancy.
reduction in transmission delays and improved reliability,
The study concludes by confirming that adaptive
even in environments with high packet loss. The discussion
scheduling enhances the efficiency and fault tolerance of
also emphasized the scalability of network coding, proving
cloud-based healthcare data storage systems. The suggested
its ability to handle large volumes of biomedical data,
approach improves the overall dependability of healthcare
making it ideal for future healthcare systems with expanding
IT infrastructures by maximizing resource use and putting
demands in telemedicine and remote monitoring.
strong fault tolerance measures in place, protecting data
This paper embodied the review of existing availability and integrity in the process.
methodologies with the injection of one scalable, efficient
approach toward the transmission of biomedical data.
Sections 1, 2, and 3, namely prior research, our proposed In the research coined by Muntadher Saadoon, he
method, and results, respectively, demonstrated how proposed that big data systems, fault tolerance is essential
network coding emerges as an innovative and workable for dependable processing and storing of quickly evolving
solution for the modern healthcare infrastructure. It not only data, particularly in applications related to healthcare. Big
guarantees the promptness and correctness of medical data data systems work in expansive settings where both software
but also paves the path for upcoming healthcare and hardware malfunctions are common. To overcome these
technologies, such as IoMT and real-time telemedicine obstacles, a number of fault tolerance techniques have been
systems. used, including erasure coding and data redundancy via
replication. These techniques improve system dependability
.
and keep data processing going even when there are errors.
II. RELATED WORKS In processing systems, prompt fault detection and
The research by Fardin Abdali Mohammadi shows the recovery are essential, but fault tolerance solutions in
rise of smartphones and wearable technology has drawn storage systems mainly address problems like storage
interest in medical applications, especially for virtual health capacity and bandwidth utilization. The management of
monitoring. However, failure sensitivity and resource massive data volumes and sustaining high performance
limitations restrict their applicability in essential during failures are two issues that recent studies have
applications. To tackle these issues, one study suggests a brought to light as obstacles to attaining excellent fault
three-tier design that combines cloud computing, mobile tolerance (research). The studied literature suggests
computing, and wearable computers. High fault tolerance is redundancy and adaptive scheduling techniques to enhance
emphasized in the system, which is crucial in medical fault detection latency and recovery effectiveness, which are
applications where patient safety may be at risk due to essential for cloud-based healthcare data storage.
malfunctions. The architecture proved dependable in
coordinating its elements and guaranteeing safe data transfer
from body sensors to the cloud through simulations. Loren Schwiebert proposed that enabling uses like
glucose monitoring and retinal prosthesis, biomedical
Similar ideas have been investigated in other research,
implanted devices, such as smart sensors, have the potential
with an emphasis on employing network coding techniques
to revolutionize medicine. Due to their low power and
to improve the fault tolerance of healthcare systems. For
computing capabilities, these devices have a difficult time
example, studies on IoT-based healthcare and Body Area
communicating wirelessly with external systems. Making
Networks (BANs) indicate that using Random Linear
sure these technologies are scalable, energy-efficient, fault-
Network Coding (RLNC), which offers redundancy and
tolerant, and biocompatible are important research problems.
error correction methods, lowers data loss and increases
Additional constraints are also introduced by embedding
system reliability. All of these developments work together
sensors in the human body: the wireless network must be
to create resilient, fault-tolerant systems that guarantee
extremely safe and dependable, able to function flawlessly
dependable data transfer, which is essential in medical
in a variety of conditions without limiting the user's
settings.
mobility, and requiring little upkeep.
These issues are discussed in the article, with a focus on
In Aliyu Muhammad’s paper he presented about the idea the necessity of wireless networking solutions tailored to
that Cloud computing is essential to healthcare because it individual applications that deviate greatly from
makes it easier to store and process private medical conventional networks. It aims to stimulate more research in
information. A recent study suggests leveraging cloud this crucial field of medical technology by showcasing early
computing to increase data storage efficiency in healthcare work in wireless networking for a retina prosthetic and
communities through adaptive and fault-tolerant scheduling. showcasing the possibilities of biomedical smart sensors.
The paper emphasizes how adaptive scheduling algorithms
are necessary in order to dynamically distribute resources in
response to changing network circumstances and data C. Neti presented an idea to comprehend how neurons
demands. By using strong fault tolerance techniques to process auditory information, specifically for sound
reduce the risks of data loss and outage, this strategy seeks localization based on spectral signals from the pinna, this
to improve system reliability. study investigates the use of neural network models. To
investigate the response characteristics of neurons that
The study promotes the use of adaptive scheduling
extract directional sound information, a feedforward neural
techniques, which modify resource allocation in real-time to
network model is presented. Fault tolerance is emphasized
guarantee peak performance even in the face of fluctuating
in the model, guaranteeing that the network will continue to create WSN applications that are more dependable and
operate even in the event that some of its components fail. effective in real-time settings.
In the context of neural networks, fault tolerance and III. PROPOSED WORKS
uniform fault tolerance are formally defined, and a
procedure to guarantee these attributes is suggested. In order A basic machine learning approach, linear regression
to preserve fault tolerance while yet producing precise uses a collection of input characteristics to predict
predictions, the task of estimating the neural network's continuous outputs. A linear relationship between the
weights is framed as a large-scale nonlinear optimization independent variables (features) and the dependent variable
problem. In comparison to unconstrained models that rely (goal) is assumed. The objective is to reduce the discrepancy
on conventional back-propagation, numerical studies between expected and actual values by identifying the linear
demonstrate that solutions with fault tolerance requirements equation that fits the data the best.
outperform unconstrained models in terms of generalization, 1.)In linear regression, the Mean Squared Error (MSE) is
or the capacity to apply learnt patterns to fresh data. This frequently utilized. The average squared discrepancy
demonstrates how adding fault tolerance not only increases between expected and actual values is measured.
robustness but also boosts the network's overall performance
in pattern recognition tasks, particularly those involving the 2.)The model's parameters (coefficients) are updated
auditory system. using the gradient descent optimization process in order to
minimize the cost function. For each update, the step size is
determined by the learning rate
Xuefeng Liu’s study investigates Fault-Tolerant Event 3.)The coefficients, denoted by β0 (intercept) and β1
Detection (FTED), a crucial field of study in Wireless (slope), are the parameters of the model. They explain how
Sensor Networks (WSNs), especially for applications like the independent variables and the dependent variable are
volcano monitoring and structural health monitoring (SHM). related.
These domain-specific applications, in contrast to
conventional WSN applications, gather intricate, dynamic
data sequences and necessitate cooperation across several 4.)The variations between actual and expected values are
sensors in order to precisely identify events. Existing FTED known as residuals. For a successful fit, the residual sum of
strategies are less effective because of these features. I- squares (RSS) should be modest.
FUND, a novel defective node identification technique
designed for SHM applications, is introduced in this study.
Using feature vectors as input, I-FUND tackles the element 5.) What linear regression presupposes:
mismatch issue, which occurs when similar items in feature
vectors are positioned differently. The method also includes Linearity: The variables' connection is linear.
an adaptive stop criterion, which increases dependability Independence: Each piece of information is autonomous.
even in cases where a significant percentage of sensor nodes
deliver inaccurate data. This paper shows how well the I- Homoscedasticity: The residuals' variance remains
FUND system works in identifying malfunctioning nodes constant.
and guaranteeing accurate event detection in difficult WSN
Normality: The distribution of residuals is normal.
environments through simulations and real-world trials. In
situations requiring intricate data sequences and sensor you.
collaboration, the work highlights the value of fault
1.) Linear Network Coding
tolerance in domain-specific WSN systems, leading to more
precise and reliable event detection.

Bi=ai 1 A 1 +¿ ai2 A 2 +⋯+ ¿ain An ¿ ¿ (1)

A.V. Sutagundar’ s research coined that the limited
processing and communication capabilities of densely a i=Finite Field
placed sensor nodes make fault tolerance crucial in Wireless
Sensor Networks (WSNs). It has been suggested to use both A! A 2 A3=data packet
static and mobile agents in a multi-agent fault tolerance
architecture to improve data dependability at the node, Bi=encoded packets linear combination
cluster, and sink hierarchical levels. By eliminating
inaccurate samples at the node level, this technique enables
nodes to calculate averages using only legitimate data. The 2.) Fault Tolerance: Error detection and correction
accuracy of the aggregated data MDPI is improved at the
cluster level by the cluster head comparing incoming data
from several nodes to find and remove conflicts. Data
transmission is maintained even in the event of a primary P(x )=x 1 a 1+ x 2 a2 +⋯+ x n an + E1 (2)
sink failure because to redundancy introduced by the sink
level's duplicate sink. This system sends "alive" packets on a
regular basis to make sure all nodes stay connected and
functional. Energy economy and error reduction are critical P(x )=Polynomial
components of fault tolerance solutions in WSNs, according
to recent study. By expanding on current models and putting E1=Error form
forth a comprehensive strategy that tackles the problems of
node, cluster, and sink-level failures, this study helps to
 1
Fault tolerance =¿ (3)
1−R

R=Ratio of Redundant data

3.) Throughput → Maximum Transmission

Fig. 2. Fault tolerance improvement comparison

T =¿
nPackets (4)
f transmission +t processing +t decoding Our method achieves low-delay transmission as one of
its key features. Even with retransmission requests and the
like (in regular systems where such procedures can result in
4.) Packet delivery ratio →successful delivery of packets
unacceptable delays), especially for real-time applications,
e.g., remote patient monitoring or emergency diagnostics.
Network coding, in contrast, enables the reconstruction of
ndelivered lost packets which do not necessarily need to be
PDR= (5)
n sent retransmitted. This ability is crucial for doctors to get access
of vital biomedical data when it matters most so they can
make informed decisions that may save a patient´s life.
1 (6)
Measure Fault Tolerance MTBF=
λ
MTBF → Mean Time between Failure
λ → Failure rate

Fig. 3. Transmission delay reduction

Fig. 1. Packet delivery ratio comparison Here are some of the key findings from our research:
● Higher Fault Tolerance: Up to 20% improvement
of fault tolerance was achieved by using network coding,
IV. RESULTS AND DISCUSSION which is huge compared with traditional error correction
The results of this study show the great possibility that method.
network coding can improve the fault tolerance and ● Faster Response Times: Removal of duplicate
reliability in wireless networks under harsh networking retransmissions heavily reduces transmission delays, giving
environment for biomedical data transmissions. Our researchers quicker access to fundamental biomedical data.
simulations show that, compared to traditional methods such
as ARQ and FEC (Forward Error Correction), network ● Scalability: We are able to handle the larger
coding can increased fault tolerance by up to 20% in this volume of data as we move towards a future where many
scenario. This increment is crucial in healthcare use cases healthcare devices will become connected.
where even slight data loss or delay means patients' lives. ● Network Unreliable Conditions : Network coding,
by distributing redundant data packets over multiple
transmission paths and cloaks the original information once
for reconstruction due to packet loss or corruption.
Our findings carry important implications for the
development of healthcare technologies, as modern systems
rely upon so-called Internet of Medical Things to propagate
and safeguard complex critical data. This dependence on
complex networks of connected machines requires fault-
tolerant communications systems, but our results indicate
that current models are unable to keep up with the fast-paced
advancement of medical technologies and the increasing [6] Kamran, M. A., & Mannan, M. M. N. (2019). Applications of
network coding for healthcare: An overview. Journal of Healthcare
demand for real-time data accessibility. Engineering, 2019.
[7] Li, P., & Chi, K. (2021). Biomedical data transmission using fault-
We must consider the generalizability of network tolerant coding mechanisms in real-time healthcare systems. IEEE
Transactions on Biomedical Engineering, 68(2), 456-467.
coding, too – that is, its ability to fit in many types of other [8] Patil, S., & Ranganathan, P. (2022). Telemedicine and IoT for
healthcare scenarios. Because of its scalability, network biomedical data: Challenges and solutions. Healthcare Technology
Letters, 9(3), 102-109.
coding can be deployed for applications today – not just
[9] “Treatment episode data set: discharges (TEDS-D): concatenated,
current ones, but for those we envision for the future, such 2006 to 2009.” U.S. Department of Health and Human Services,
as the use of telemedicine and remote monitoring, big-data Substance Abuse and Mental Health Services Administration, Office
of Applied Studies, August, 2013, DOI:10.3886/ICPSR30122.v2
management at scale, and other demanding applications we
may not even have dreamed of yet, as our wired and IEEE conference templates contain guidance text for
possibly wireless world of biomedical, informatics and composing and formatting conference papers. Please
health monitoring devices continues to grow. ensure that all template text is removed from your
conference paper prior to submission to the
Additionally, our work spurs further studies to integrate conference. Failure to remove template text from
network coding within communication protocols and your paper may result in your paper not being
published.
existing healthcare infrastructure, as well as to study its
performance under different network conditions such as
congested urban environments and remote rural areas with
limited connectivity. Specifically, through further research it
would be interesting to assess how to optimise network
coding for different healthcare settings under specified
network constraints in order to improve patient care and
safety.
Overall, our research demonstrates the ability of network
coding in storing and reliably transmitting biomedical data
in order to maximise fault tolerance and overcome the
limitations of traditional error correction over long distances
– a challenge amplified by the amount of distributed
computing required in healthcare, where the risks associated
with delays and scalability are high. A next step is to
develop practical frameworks for deploying network coding
techniques in day-to-day healthcare settings to anticipate a
more resilient healthcare system that will pave the way for
data-driven care leading to superior quality of care for
patients.

.
REFERENCES
[1] Fragouli, C., & Soljanin, E. (2007). Network Coding: Fundamentals
and Applications.)
[2] El Gamal, A., & Kim, Y. H. (2011). Network Information Theory.
Cambridge University Press.
[3] Xu, Z., He, T., & Huang, X. (2018). Fault-tolerant transmission in
wireless biomedical sensor networks. IEEE Access, 6, 2337-2348.
[4] Dohler, M., & Heath, R. W. (2010). Wireless communications and
networking for healthcare applications. IEEE Wireless
Communications, 17(1), 5-6..
[5] Zhang, Q., & Wicker, S. B. (2005). Network coding and reliability in
biomedical wireless networks. IEEE Transactions on Information
Theory, 51(10), 3575-3588.