A Pilot Study on Tissue Deformation Using an Integrated Sensor–Actuator Blood Collection Setup in Aquaculture (Salmo salar)

Abstract

This pilot study presents a sensor–actuator setup designed to evaluate tissue deformation in Atlantic Salmon (Salmo salar) during needle insertion. The system integrates three types of low-cost, commercially available force sensors to capture force profiles and identify biomechanical events associated with tissue layer transitions. Controlled insertions were performed on a deceased specimen, and the resulting force data were analyzed to quantify insertion dynamics and estimate tissue deformation. A simulation model based on the recorded force values was developed to calculate stress distribution and deformation, which ranged from 0.001 µm to 8.4 µm and from 0.3 N/m² to 4.9 N/m², respectively. The results indicate minimal biomechanical disruption, supporting the feasibility of using sensor–actuator systems for minimally invasive procedures in aquaculture. Although biological responses such as inflammation and healing could not be assessed, this study lays the basis for future research involving live specimens and blood sampling. The findings carry important implications for fish welfare, offering a pathway toward ethical and automated sampling techniques in aquaculture and livestock applications.

1. Introduction

The integration of sensory feedback systems with electronics and actuators has opened new possibilities for bioanalytical procedures in both clinical and livestock environments [1,2,3,4]. This integration demands careful consideration for animal welfare and the reliability of analytical outcomes. Such cross-disciplinary initiatives create new application areas, where insights from one domain can enhance practices in the other. In this context, needle-based procedures remain a fundamental technique for collecting and analyzing samples. By embedding sensors, actuators, and electronic controls into these procedures, novel applications can be developed. One promising application of this integration is in monitoring the health of farmed fish and advancing aquaculture research. However, needle-based procedures inherently involve mechanical interaction with soft tissue. The interaction between the needle and the tissue layers may result in tissue deformation, which causes discomfort to experimental subjects [5]. Deformation beyond the threshold may compromise the texture by damaging muscle fibers. Moreover, uncontrolled tissue deformation can hamper the sampling accuracy and reduce the reliability of collected data, specifically when sampling small areas where the caudal vein is the target for fish health monitoring. Tissue deformation due to needle insertion is an unavoidable consequence. The magnitude of tissue deformation is an important metric that must be considered for designing a sample collection setup. In addition, quantifying stress due to tissue deformation during needle insertion [6] plays a pivotal role in determining fish welfare, procedural accuracy, and the integrity of the analytical outcome.

Previous studies have investigated the mechanics of needle–tissue interaction, emphasizing insertion velocity, tissue characteristics, and force feedback. These studies have contributed to minimizing tissue deformation during needle insertion. Other investigations have assessed force variation during penetration [7,8]; however, quantification of stress and deformation specifically for the aquaculture research has not been delved into. A few attempts have investigated the mechanical properties of fish tissue despite the growing demand for automated, ethical sample collection methods in aquatic bioanalysis [9,10].

The insertion of a needle into the tissue may contribute to needle deflection. The factors behind that are the mechanical properties of soft tissue, contact force, and frictional force [11]. In addition, several factors affect the amount of tissue deformation. These are boundary conditions, tissue geometry, biomechanical tissue properties, and applied force, which are important in simulation and modeling applications [12]. Therefore, key issues for accurate insertion depend on knowing the interaction forces and developing a tissue deformation model during needle insertion. Previous studies have determined that the knowledge of interactive forces during needle insertion may give useful insights into identifying different tissue types. Moreover, this may help precisely control automated insertion while reducing tissue deformation. In the absence of real-time imaging, force feedback, and an accurate mechanical model, identifying tissue deformation and target movement would be valuable [13,14]. Such models could be updated with force feedback during needle insertion.

This article presents a low-cost, easy-to-assemble test setup, integrating sensors, actuators, and electronics to generate force profiles of needle insertion similar to the kind of needle–tissue contact mechanics model. A comparative overview is presented in Table 1. The test setup is conceptualized and developed with the help of force measurement sensors, actuators, and microcontrollers. Moreover, accurately determining the precise position of the needle within the tissue during insertion is challenging without imaging tools. To address this limitation, force sensors were utilized to gather insertion force profiles. Variations in force readings were used to infer transitions between tissue layers, providing a general understanding of tissue behavior during penetration.

Furthermore, analysis of the force profile during needle insertion reveals a relationship between rupture force and insertion velocity. The rupture decreases for faster insertion under certain conditions. The previous studies observed that increasing needle velocity effectively reduces tissue deformation and minimizes damage [15,16]. Therefore, the experimental setup was repeatedly tested to ensure that it achieved the maximum attainable needle insertion velocity.

This study adopted an approach in which force data were collected and the resulting profiles were compared with findings from previous research conducted by other groups. The force values from the sensors were used as input variables in a multi-physics-based finite element analysis to measure the resulting tissue stress and deformation [17]. The stress and deformation during needle insertion into soft tissue were reported based on experimental data.

Table 1. Overview of the existing literature and this work’s novelty.

Study	Title	Methodology	Performance Metrics	Novelty of This Work
Mahvash et al.	Mechanics of dynamic needle insertion into a biological material [7]	Using the J integral method from fracture mechanics, rupture events are modeled as sudden crack extension that occurs.	The work shows that increasing the velocity of needle insertion will reduce the force of the rupture event. The model predicts the rupture deformation.	The mechanics of rupture events, the effect of insertion velocity on needle force, and tissue deformation. A nonlinear viscoelastic Kelvin model is used to predict the relationship between tissue deformation and the rupture force at different velocities.
Smita et al.	Assessment of Tissue Damage due to Mechanical Stresses [5]	An experimental setup for stresses mimicking minimally invasive surgery (MIS) using a motorized endoscopic grasper. Finite element analysis was used to estimate stress distribution. Statistical Analysis to detect stress magnitude.	Apoptosis Levels Neutrophil infiltration Relative tissue resistance to injury	Identifying stress threshold for safe tissue manipulation in MIS using biological and computational techniques. An initial study for a surgical robot that provides real-time force feedback to avoid tissue damage. Using surgical simulators to train MIS practitioners by developing a tissue response model.
Frank C.P. Yin et al.	An approach to quantification of biaxial tissue stress–strain data [6]	Simultaneous biaxial stretching under various combinations Fit data using a five-parameter exponential strain energy function with minimal free parameters Test the validity of the standard statistical model A non-parametric method to quantify coefficient variability.	Variability of strain energy across experimental trials Assessment of stress–strain data in tissue behavior	A systematic and statistically based work on the variability of biomechanical data First to apply bootstrapping to evaluate uncertainty in biaxial stress–strain coefficient estimates
Okamura et al.	Force modeling for needle insertion into soft tissue [9]	Used a one-degree-of-freedom robot with a load cell and needle attachment to insert needles into bovine liver tissue. Used Computer tomography (CT) for data acquisition during contact between tissue and needle. Used force modeling principles based on stiffness, friction, and cutting.	Sensitivity of insertions forces changes in tissue density and needle dimension Insertion force measurements across different needle designs.	A comprehensive multi-phase force model for needle insertion into soft tissue. Phase segmentation of needle–tissue interaction using CT imaging Quantify tip geometry and diameter effects on insertion force and mechanical behavior during needle-based procedures.
Fauver M. E. et al.	Microfabricated cantilevers for measurement of subcellular and molecular forces [17]	Microfabricated cantilever-beam force transducers using thin silicon-nitride films	Force sensitivity and precision Transducer stiffness Reproducibility	Micro-scale force transducers for ultra-sensitive biomechanical applications.
This Work	A pilot study of the tissue deformation with an integrated sensor–actuator-based blood collecting setup for aquaculture (Salmo salar)	Sensor–actuator and electronics setup for needle control and data acquisition Simulation was conducted using the force data	Insertion force profile Tissue deformation range Stress Range	First-of-its-kind study of force-induced tissue deformation during needle insertion in Salmo salar using a multi-sensor approach. Using force sensory data in a simulation to measure tissue deformation and stress during needle insertion.

Table 1. Overview of the existing literature and this work’s novelty.

Study	Title	Methodology	Performance Metrics	Novelty of This Work
Mahvash et al.	Mechanics of dynamic needle insertion into a biological material [7]	Using the J integral method from fracture mechanics, rupture events are modeled as sudden crack extension that occurs.	The work shows that increasing the velocity of needle insertion will reduce the force of the rupture event. The model predicts the rupture deformation.	The mechanics of rupture events, the effect of insertion velocity on needle force, and tissue deformation. A nonlinear viscoelastic Kelvin model is used to predict the relationship between tissue deformation and the rupture force at different velocities.
Smita et al.	Assessment of Tissue Damage due to Mechanical Stresses [5]	An experimental setup for stresses mimicking minimally invasive surgery (MIS) using a motorized endoscopic grasper. Finite element analysis was used to estimate stress distribution. Statistical Analysis to detect stress magnitude.	Apoptosis Levels Neutrophil infiltration Relative tissue resistance to injury	Identifying stress threshold for safe tissue manipulation in MIS using biological and computational techniques. An initial study for a surgical robot that provides real-time force feedback to avoid tissue damage. Using surgical simulators to train MIS practitioners by developing a tissue response model.
Frank C.P. Yin et al.	An approach to quantification of biaxial tissue stress–strain data [6]	Simultaneous biaxial stretching under various combinations Fit data using a five-parameter exponential strain energy function with minimal free parameters Test the validity of the standard statistical model A non-parametric method to quantify coefficient variability.	Variability of strain energy across experimental trials Assessment of stress–strain data in tissue behavior	A systematic and statistically based work on the variability of biomechanical data First to apply bootstrapping to evaluate uncertainty in biaxial stress–strain coefficient estimates
Okamura et al.	Force modeling for needle insertion into soft tissue [9]	Used a one-degree-of-freedom robot with a load cell and needle attachment to insert needles into bovine liver tissue. Used Computer tomography (CT) for data acquisition during contact between tissue and needle. Used force modeling principles based on stiffness, friction, and cutting.	Sensitivity of insertions forces changes in tissue density and needle dimension Insertion force measurements across different needle designs.	A comprehensive multi-phase force model for needle insertion into soft tissue. Phase segmentation of needle–tissue interaction using CT imaging Quantify tip geometry and diameter effects on insertion force and mechanical behavior during needle-based procedures.
Fauver M. E. et al.	Microfabricated cantilevers for measurement of subcellular and molecular forces [17]	Microfabricated cantilever-beam force transducers using thin silicon-nitride films	Force sensitivity and precision Transducer stiffness Reproducibility	Micro-scale force transducers for ultra-sensitive biomechanical applications.
This Work	A pilot study of the tissue deformation with an integrated sensor–actuator-based blood collecting setup for aquaculture (Salmo salar)	Sensor–actuator and electronics setup for needle control and data acquisition Simulation was conducted using the force data	Insertion force profile Tissue deformation range Stress Range	First-of-its-kind study of force-induced tissue deformation during needle insertion in Salmo salar using a multi-sensor approach. Using force sensory data in a simulation to measure tissue deformation and stress during needle insertion.

According to ethical standards, experiments on animals should minimize stress [18]. The mentioned setup was tested on a dead salmon to measure tissue stress and deformation due to needle insertion. The literature on tissue stress due to needle insertion in aquaculture or livestock research remains limited. Needle-based procedures during transitions between tissue layers often involve puncture events that produce deformation. Deformation causes stress on the application area and the experimental subject. Therefore, minimizing deformation for a setup that involves needle insertion is an important aspect for ethical and effective fish handling, especially in commercial fish farming and aquaculture research. In this context, the results of this research are the first of their kind that have investigated an approach for stress measurement in clinical animal experimentation. Moreover, acquiring sensory data and then using the data in a simulation to measure the stress of muscles during needle insertion for aquatic specimens is a novel approach. The outcome of this study is a reference indicating minimal disruption, aligning with best practices and accurate sampling. Beyond aquaculture, this study offers broader applications in clinical research, robotic insertion, and training simulations, where it is important to document tissue or muscle stress. While real-time imaging feedback was beyond the scope of this study, further research is essential for improving procedural safety and supporting comprehensive validation of the setup.

2. Materials

2.1. Actuator

The Actuonix L16 micro linear actuators are constructed using an anodized metal shaft, gearbox, and steel ball bearings. Depending on the stroke length, it has a mass of between 56 and 84 g. The speed is dependent on the load applied. The actuators were purchased from RS Components. There are three pins for the RC input signal, ground, and power. The white, red, and black wires are used for RC (radio control) input signal, power, and ground, respectively. The controller transmits a 1.0 ms pulse to retract wholly and 2.0 ms pulse signals to extend fully. The actuators have a gear ratio of 35:1, a peak efficiency of 24 N at 24 mm/s, and a peak power point of 50 N at 16 mm/s.

2.2. Force Sensitive Resistor (FSR)

Tekscan’s Flexiforce sensor (FSR) was purchased from Sparkfun Electronics, Boulder, CO, USA. The sensor resistance changes when the force is applied to the sensing area. The resistance of the FSR depends on whether the applied force increases or decreases. If no pressure or force is applied, the resistance of the FSR is larger than 1 megaohm (mΩ). When the force passes a certain value, the resistance will not change. A static resistor is connected in series with the FSR to operate as a voltage divider. This way, for different forces/pressures, values will generate different voltages. The circuit creates a variable voltage output that can be read by the microcontroller’s ADC (analog-to-digital converter) input. From this voltage variation, the applied force to the sensor was calculated using the following equation.

V₀ = V_cc[(R_fixed) ÷ (R_fixed + R_fsr)]

(1)

Here, V₀ is the output voltage, V_cc is the supply voltage, and R_fixed is the fixed resistor value. A 3.3 kΩ resistor was connected to the Arduino microcontroller. In a breadboard, the FSR’s three terminals were connected. The middle terminal of FSR was idle. One terminal of the FSR was connected to the 3.3 kΩ resistor in series. The other terminal of the FSR was connected to the 5 V pin of the Arduino. An A₀ microcontroller pin was connected between the FSR terminal and the resistor. The other end of the resistor was connected to the microcontroller’s ground pin. When the connection was made, the reference code for the FSR was uploaded to the Arduino-integrated development environment. The Arduino serial monitor gave the values for the applied force.

2.3. Force-Sensitive Capacitor (FSC)

Singletact’s force-sensitive capacitor (FSC) is a sensor that detects capacitance changes when force is applied. The capacitive force sensor is 0.35 mm thick. The connection is made with an analog 3-wire interface for immediate data collection. Singletact capacitive force sensors were connected to the interface board flat flex connector (FFC), and the sensing area was 8 mm. Singletact digital output (PSI) can be converted to a pressure value (N/mm²). For conversion, the values in Newtons (N) were inserted into the following equation, where F_n represents the total applied force in Newtons, Output is the sensor’s measured response, 512 is the scaling factor, and the Full-Scale Range denotes the sensor’s maximum capacity. The full-scale range of the model “8 mm, 1 N” force-sensitive capacitor sensor is 1 N [6]. The applied force was calculated by following the equation:

F_n = (Output_psi ÷ 512) × full scale range (n)

(2)

2.4. Load Cell

A load cell measures force changes, and one was purchased from Sparkfun Electronics. It can convert up to 100 g of force into an electrical signal. A load cell employs strain gauges that record the minimal resistance change to the force applied. In a strain gauge-based load cell, compression and tension are measured. It converts force, pressure, and weight into a change in electrical resistance, which can be measured. Strain gauges usually consist of four strain gauge elements that are electrically connected to form a Wheatstone bridge circuit. The change in resistance was calculated by following the equation:

V_out = [{R₃ ÷ (R₃ + R₄)} − {R₁ ÷ (R₁ + R₂)} × V_in]

(3)

R₁, R₂, R₃, and R₄ are the active strain gauges that change resistance when mechanical force is applied, V_in is the input voltage, and V_out is the output voltage. To detect even a small change in resistance, an auxiliary amplifier HX711 with five connecting pins was purchased from SparkFun Electronics. Pins are labeled as red, black, white, green, and yellow. This color-coding matches the wires coming from the strain gauges of the load cells. V_DD and V_cc are connected to 5 V, DAT and CLK are connected to pins 3 and 2, respectively, of the microcontroller, and the other pin is grounded. The breakout board HX711 was soldered to the load cell. After soldering, V_DD, V_CC, DAT, CLK, and GND were connected to the Arduino. V_CC is the analog voltage to power the load cell, and V_DD is the digital supply voltage used to set the logic level. Once the connection was established, the reference code was uploaded to the Arduino.

For each sensor, a series of known static loads was applied within its operating range, and the corresponding sensory feedback was recorded. A calibration curve was then generated against the applied force. The calibration curve was validated by comparing it with the reference calibration data provided in the sensor’s datasheet. To ensure consistency and accuracy, manufacturer-recommended calibration procedures were followed.

2.5. Microcontrollers

The STM 32 Nucleo-F303RE microcontroller was collected from the University of South-Eastern Norway Electronics Lab. This board has Arduino connectivity support and ST morpho headers. STM 32 Nucleo boards work with an open development platform (Mbed). This board integrates the ST-LINK programmer. The Arduino Nano was purchased from Digi-Key Electronics, Thief River Falls, MN, USA. The Arduino Nano has an ATmega328 microchip. The operating voltage is 5 volts with 8 analog input pins. Each of these analog pins has an inbuilt analog-to-digital converter (ADC) with a resolution of 1024 bits. The microcontroller-embedded board has 22 digital inputs/outputs, including 6 are pulse-width modulation (PWM) pins, consumes 19 mA, and weighs 7 g. The flash memory is 32 (KB), static random-access memory is 2 KB, and the clock speed is 16 (MHz). Moreover, as two microcontrollers were involved in the system (one for driving the actuators and another for collecting sensory data), rather than using multiple power sources, the two microcontrollers shared a common source and completed the circuit.

3. Method

3.1. Needle–Tissue Contact Mechanics Model

Understanding how needles interact with tissue is key to recognizing why force measurement matters during insertion. This theoretical insight is vital for developing systems that prioritize minimal invasiveness. One directional force of a needle during insertion into soft tissue is the summation of different forces distributed along the needle, such as stiffness, frictional, and cutting forces. Needle insertion is categorized into pre-puncture and post-puncture phases. At pre-puncture, the force increases steadily, and a sharp drop or a peak identifies the puncture event. During post-puncture, the amount of force is variable due to friction, cutting, and collisions with the interior structure. Stiffness force belongs to the pre-puncture, while friction and cutting forces occur post-puncture. The force data generated by sensors in this research are the total force during tissue penetration.

F_total = f_stiffness + f_friction + f_cutting

(4)

Key insertion events are another terminology in needle–tissue contact theory that suggests the presence of a new tissue layer. This helps to understand and gives an overview of the internal scenario of the presence of a new tissue layer, vein, muscle, and so on in response to a sudden drop and rise in the force curve. It occurs when the needle is inserted into soft tissue. At the needle tip, interaction forces are developed and change accordingly due to tissue cutting and friction between the needle and tissue layers. For example, firstly, a needle penetrates the skin layer and enters the relatively dense muscle. Along the path, the needle may or may not encounter another stiffer muscle. Due to this change in muscle layers, the force profile witnesses a sudden drop and rise [19,20,21,22,23]. A general scenario can be pictured in terms of different tissue layer densities and sudden drops and rises in force while penetrating different muscles/tissues. Section 5 discusses the force profile, sudden drops and rises in force, and the presence of tissue layers.

3.2. Tissue Stress and Deformation Measurement with Simulation

The needle insertion into soft tissue tends to generate stress in the application area. Therefore, needle penetration may help in measuring tissue deformation by simulation. Since force sensors achieved profiles that follow the needle–tissue contact mechanics model, a multi-physics-based simulation (COMSOL v6.0) for finite element analysis was carried out using the force profiles to measure tissue stress and deformation. To replicate a chunk of salmon tissue, in the simulation, the mechanical properties of the tissue were taken from

Table 2. Mechanical properties of salmon skin, tissue (fat and muscle), and needles used in the COMSOL simulation [24,25,26,27,28,29].

Medium	Young’s Modulus (Pascal)	Poisson’s Ratio	Density (g/m3)
Needle	200 × 108	0.27	0.0078
Salmon skin	1.9 × 108	0.30	1.15
Salmon tissue	10,000	0.47	1.06
Salmon fat	10,000	0.47	0.90

. Table 2 represents the skin, fat, and tissue characteristics of salmon fish. For some of the data in Table 2, other fish species (i.e., tilapia, yellowfin) were considered as reference points.

4. Experimental Setup

4.1. Actuator-Based Control System

As shown in Figure 1, the experimental setup consisted of a linear actuator integrated with three different commercially available force sensors. As shown in Figure 2, these sensors were connected individually at a time to measure the force during needle insertion into the salmon tissue. To control the needle deflection during insertion [30], a fixed penetration depth was programmed into the microcontroller governing the actuator that would be responsible for needle movement. Based on anatomical estimates, which are typically between 2 and 5 mm depending on the fish species, the movement was programmed to reach a maximum of 5 cm; much of this distance did not make ideal contact. Consequently, between the initial contact and the static phase, the needle penetrated no more than 2 cm. To maintain alignment and minimize bending, both the needle and the actuator were supported by a rigid platform. Although real-time deflection monitoring was not implemented, the combination of controlled movement and mechanical constraints was followed to ensure a consistent needle trajectory. Using the Mbed IDE, the microcontroller was programmed. The program controlled the actuator responsible for needle insertion, enabling it to move in a steady, linear trajectory toward the sample.

The force data was collected during each insertion for each sensor and recorded for analysis. The force profiles generated by different sensors were compared for sensitivity, consistency, and characteristics. Statistical and visual analyses were performed to identify key insertion events, quantify force values, and measure deformation.

Force measurements were recorded and compiled in an Excel spreadsheet. Based on this dataset, force–time graphs were generated for each sensor to visualize insertion dynamics and identify tissue layer transitions.

This pilot study was conducted using a single specimen of Atlantic salmon (Salmo salar), which was deceased at the time of experimentation. Therefore, no anesthesia protocols were required, and no blood sampling procedures were performed. The primary objective of this study was to gather force feedback during needle insertion and evaluate tissue deformation resulting from needle insertion using a sensor–actuator setup.

4.2. Simulation

A random model was tested in the simulation to check whether the hypothesis generated a satisfactory outcome. Later, force sensory data were utilized as input variables in the simulation. In Figure 3, a needle profile was introduced into a 3 cm × 3 cm block. The material for the needle was chosen as stainless steel 405. Due to practical and anatomical considerations, a standard syringe needle with a length of 50 mm and an outer diameter of 0.819 mm (approximately 21 gauge) was selected. The 50 mm length ideally provides sufficient reach to access the caudal vein in a range of fish sizes. This prevents excessive penetration and reduces the risk of internal damage. Shorter needles may not reach the target area, and longer needles are more prone to deflection if not supported properly. The diameter of 0.819 mm offered a balance between structural rigidity and minimal invasiveness. Thinner needles may bend more easily or clog, while thicker needles could cause trauma.

A random range of forces between 0.1 N and 1 N with an interval of 0.1 N was applied to the needle. Tissue stress and displacement were plotted on the y-axis against the applied force.

5. Results and Discussion

5.1. Controlling Actuator Speed for Needle Insertion

Needle insertion with a high velocity tends to cause less tissue deformation. Therefore, two experiments were executed to understand the programming of the actuators so that they produced the highest optimum velocity shown in Figure 4. In the first experiment (Figure 4A), actuators were connected to an external power source of 6 V and the 5 V power source pin of the microcontroller. The reason for carrying out these experiments was to determine whether connecting actuators with the microcontroller power source reduces the velocity of the actuator. In both cases, the collected data observed an increment until it reached a saturation point. The maximum speed (Figure 4A) with an external power source (blue line) was 2.22 cm/s (22.2 mm/s) at 6 V and 2.12 cm/s at 5 V (green line). The actuator connected to a 6 V power supply had an average speed of 1.68 cm/s and a standard deviation (std) of 0.63 cm/s from seven different experiments. With the same number of experiments, the actuator connected to a 5 V power supply had an average speed of 1.61 cm/s, std 0.56 cm/s. In Figure 4B, Student’s t-test was performed on step and speed when the voltage was connected to 6 V.

The paired test yielded a score of 0.052, while the two-sample tests produced scores of 0.0534 (unequal variance) and 0.0336 (equal variance). The coefficients of determination (R²) were 0.834 for speed and 0.881 for step, both indicating statistically significant relationships. In Figure 3C, Student’s t-test was performed on step and speed while the actuator was connected to a 5 V power supply, which gave a paired test score of 0.0526, a two-sample unequal variance score of 0.0533, and sample equal variance scores of 0.0335, with the R²-values for speed being 0.8 and the step being 0.881. The Pearson coefficient and level of significance were 0.72 and 0.05, respectively. These analyses proved that there is an association between speed and step. The speed reached 2.13 cm/s at “STEP = 100” in both power sources. Thus, the function ‘STEP = 100’ was taken as a reference value in the programming code to determine the velocity of actuators.

5.2. Sensory Data and Force Profile Analysis

The Force Sensitive Resistor (FSR) recorded the perpendicular force distributed during the needle insertion. Figure 5A presents key insights into the force profile pattern. At point 1, the needle contacts the tissue and penetrates. The tissue has different layers, each with individual density, Young’s modulus, and thickness. Therefore, a sudden drop and rise in the force curve at points 2, 3, and 4 are detectable, presumably due to the presence of new tissue layers. At the static phase of the device (between points 4 and 5), the force profile indicates a steep drop as the needle is extracted.

Although FSR recorded the insertion force, the important insertion events are not understandable in Figure 5A. Furthermore, the plateau in this graph is not exactly appropriate. Since the plateau is a static phase, a slight drop in force was expected. To investigate the limitations of ‘test setup-1’, the FSR sensor was setup in a cantilever beam position in “test setup-2”. FSR recorded (Figure 5B) a sudden drop and rise in force during needle insertion, as seen in points 2, 3, and 4. This time, a straight line is visible at the static phase between points 4 and 5. The insertion forces in Figure 5B were recorded at 0.4 N. There is a drop in force between points 2 and 3. The insertion force range (0.4–0.7 N) in Figure 5B has increased slightly at the same spot, compared to Figure 5A. Figure 5C is acquired with a Force Sensitive Capacitor (FSC). Compared to the preceding figures, the force values from the FSC profile indicate key insertion events occurring within the 0–2 N range, significantly higher than those observed in profiles Figure 5A,B. The sudden drop and rise in Figure 5C at points 2, 3, 4, and 5 have a difference of about 1 N. The profile of Figure 5C recorded four major drops and rises between points 2, 3, 4, and 5. This indicates that the possibility of interchanging tissue layers is more visible in this force profile analogy. Despite the promising profile acquired from FSC, the sensor had a subtle noise.

FSR and FSC are both useful for tactile sensing. These sensors have a small sensing area. This is a good option for sensing force distributed by a needle in a small cross-sectional area. Understandably, FSC produced better results than FSR when put into the cantilever position. For example, the insertion force was much higher (~2 N) with FSC. The force range of sudden drops and rises was ~1 N. However, tactile sensors have some drawbacks for this application. Firstly, positioning them in the experimental setup was difficult. Secondly, the repetitive measurement produced different force profiles, although Figure 5B,C have a similar trend.

To overcome the problems associated with FSR and FSC, a load cell was used. The load cell is a type of sensor that is specifically designed for measuring force from a cantilever position. The load cell generated a better force profile scenario. The insertion range (0–1.2) is almost similar to the force values found with FSC. Which is more or less a difference of 1 N. Compared to Figure 5C, a significant drop and rise are also present in Figure 5D. Not only that, three key insertion events were found in Figure 5D, suggesting the presence of two muscle layers in points 2, 3, 4, and 5. Notably, during the static phase in points 5–6, the force curve saw a slow decrement, which was expected but not present in the FSR and FSC profiles. From point 6, the force decreased straight as a result of the needle extraction. The higher force sensitivity of FSC is likely due to its design characteristics, such as thinner sensing elements and swift responsiveness to minute pressure changes. However, this sensitivity also introduced noise and less clarity in the force curve. Therefore, it was difficult to identify important insertion events, specifically tissue layer transition and needle static phase. In contrast, the load cell detected a clearer and more interpretable force profile.

5.3. Force, Stress, and Deformation Measurement

A parametric sweep study quantified stress and tissue deformation resulting from variations in applied force and tissue geometry. Values from Table 2 were used in the simulation to replicate a chunk of salmon. In the simulation, different force profiles from Figure 5 were used as the penetration force (input variables) of the needle. Force was applied at the tip of the needle. The contact area between the needle and the salmon chunk deformed due to the applied force. The area around the tip of the needle observed maximum stress. The stress developed steadily and linearly against the applied force.

Stress and tissue deformation gradually decreased with increasing distance from the contact point. Each force profile in Figure 5 was divided into three approximate layers (skin, fat, and solid muscle). For example, in Figure 5A, the skin region was approximated for the force range between 0.15 and 0.3 N, fat 0.3–0.35 N, and solid muscle 0.25–0.4 N, respectively. In the simulation, when the needle was in the skin region, the force values in the skin were applied as the penetration force. The process was repeated for fat and muscle layers. Based on this analogy, tissue stress and deformation shown in Figure 6A were calculated using the force profile presented in Figure 5A. ‘Stress vs. Force’ and ‘Deformation vs. Force’ were plotted.

Stress acts linearly to the applied force. In Figure 6A, the skin was penetrated at 0.15 N. At that applied force, the resulting stress was recorded at 0.3 N/m², and tissue displaced about 5 µm. The force values of the fat layer (0.3–0.35) N were used as input variables in the simulation, and the results were plotted in Figure 6B. It was assumed that at 0.3 N, the needle penetrated the fat layer. At that force, the resulting stress and deformation were found to be 2.95 N/m² and ~1.18 µm, respectively. Since fat is embedded within the solid muscle structure, the fat layer is considered part of the solid muscle and thus calculated accordingly. Figure 6C represents the region of the tissue layer (0.25–0.4) N. The needle entered the muscle layer at 0.25 N.

At that force, the recorded stress and displacement were ~5.5 N/m² and 2.1 µm, respectively. The analogy of Figure 6 applies to Figure 7, Figure 8 and Figure 9, as only the sensors and force values have changed in the experimentation. Figure 7 shows the results of a finite element analysis based on the force profile from Figure 5B, using data collected with a force-sensitive resistor and test setup 2. In Figure 7A, the skin region was considered for force ranges between 0.4 and 0.6 N.

In that applied force, the tissue went through the stress of about N/m² (1.2 − 0.8) = 0.4, and the deformation is approximately (3 − 2) µm = 1 µm. Figure 7B gives measurements of the salmon fat layer (0.5–0.6 N), where the tissue had a stress of about (5.9 − 4.9) = 1 N/m² and deformation (2.4 − 1.95) = 0.45 µm. Figure 7C deducted measurements of solid muscle excluding the fat layer (0.5–0.7 N), which are as follows: stress N/m² (2.9 – 1.45) = 1. 45 and displacement (5.9 − 4.2) µm = 1.7 µm. The transition from fat to tissue layer observed deformation of about (1.7 − 0.4) µm = 1.3 µm.

Figure 8 recorded stress and deformation measurements from Figure 5C. The given skin region in Figure 8A is between (0–2 N) = 2 N. The resulting stress and deformation are = 4 N/m² (4 − 0) and (0.001 − 0.0001) µm = 0.0009 µm, respectively. Figure 8B (1–2 N) measured stress and deformation for the fat layer, which is = 1 N/m² (2 − 1) and (8 − 4) µm = 4 µm. Figure 8C is the measurement for salmon tissue (1–2 N). Stress is = 3 N/m² (5.8 − 2.8), and deformation (16.5 − 8.5) µm = 8 µm.

Figure 9 was deduced from Figure 5D. Figure 9A is the skin layer (0–0.8) N where stress is (1. 6 − 0) = 1.6 N/m² and deformation is (40 − 0) µm = 40 µm. Figure 9B recorded stress and deformation for the fat layer (0.4–0.7) N. The respective results are = 3 N/m² (6.8 − 3.8) and (2.8 − 1.5) = 1.3 µm. Figure 9C measured stress and deformation for the tissue layer (0.4–1) N. The resulting stress is (2.9 − 1.2) = 1. 7 N/m² and deformation is (8.8 − 3) µm = 5.8 µm. The results acquired by the sensors and simulation infer that during the needle insertion into salmon tissue, the range of deformation is as low as 0.001 μm up to 8.4 μm, and stress is 0.3 N/m² up to 4.9 N/m².

The calculated coefficient of variation (CV) for peak force ranged from 4.6% to 12.6%, depending on the sensor type and tissue layer. The highest variability was observed in the fat layer. The values of tissue deformation for the coefficient of variation were between 6.8% and 15%. The muscle layer exhibited the lowest deformation. CVs were calculated in

Table 3. Calculation of the coefficient of variation for peak force and deformation.

Tissue Layer	Metric	Mean Value (N)	Standard Deviation (N)	Coefficient of Variation (%)
Skin	Peak Force	0.25	0.02	8.0
Fat	Peak Force	0.55	0.07	12.7
Muscle	Peak Force	0.65	0.03	4.6
Skin	Deformation	1.0	0.15	15.0
Fat	Deformation	2.8	0.19	6.8
Muscle	Deformation	5.8	0.41	7.1

using data derived from Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8.

These coefficient variation (CV) values show that absolute values of force and deformation varied across trials, influenced by sensor sensitivity, biological heterogeneity, and minor experimental inconsistencies. Lower CV (4.6 for peak force and 6.8 for deformation) indicates consistency and repeatability in measurements, reflecting stable tissue response and reliable sensor performance. Higher CV is due to variability in the data.

The results of this pilot study demonstrate that needle insertion force profiles can be used to identify transitions between tissue layers and estimate deformation [31]. The measured deformation range (0.001–8.4 µm) and stress values (0.3–4.9 N/m²) indicate minimal mechanical disruption, demonstrating that the test setup effectively achieved its primary goal of quantifying insertion force and tissue deformation during needle insertion. While the measured deformation ranges (0.001–8.4 µm) are relatively small, it is a significant step in terms of mapping subtle mechanical disruptions that may affect fish welfare. Such microscale deformation can potentially trigger stress response, inflammation, or behavioral changes in fish. For example, needle-induced injury to the epidermis may compromise the mucosal barrier, increasing susceptibility to pathogens such as salmon lice. However, in this study, the deformation values observed with the test setup fall below thresholds. Since the fish health monitoring procedure aims to be minimally invasive, the results of this study align with welfare standards that prioritize low-impact handling.

The reported deformation (0.001 μm to 8.4 μm) and stress (0.3 N/m² to 4.9 N/m²) were derived from a simulation focused on the needle–tissue interface, specifically at the tip region during penetration. The simulation was designed to capture a small region of tissue around the needle area with high spatial resolution. The deformation values reflect localized deformation rather than the whole tissue. The simulated tissue was modeled using mechanical properties from salmon and other fish species as a reference.

These findings support the feasibility of using sensor–actuator systems for controlled needle-based procedures in aquaculture applications. By generating force profiles and mapping tissue deformation within acceptable mechanical limits, the study proposes a pilot sample collection technique that aligns with ethical standards and minimally invasive practices. The setup is intended for practical use in fish farms and livestock facilities, where it can assist in collecting biological samples for health monitoring and condition assessment of farmed animals. However, biological responses such as inflammation, healing, and stress behavior could not be assessed due to the use of a deceased salmon specimen. Additionally, the absence of blood sampling limits the immediate functionality in operational aquaculture settings. Future research should focus on increasing the sample size with live specimens and testing blood sampling. This setup is potentially applicable for both livestock and aquaculture; therefore, implications for fisheries and livestock are significant. Minimizing tissue deformation during handling and sampling is critical to reducing injury, preserving the mucosal barrier, and preventing contamination. This study presents a custom-designed and tested setup that effectively limits mechanical disruption, laying the foundation for future development of automated sampling systems.

6. Conclusions

In this study, a low-cost and easily assembled test setup was developed to capture force data and quantify tissue deformation during needle insertion. Real-time imaging was intentionally excluded to preserve the simplicity of the design [32]. Instead, commercially available force sensors were employed to record insertion forces. These force measurements were then used as input for finite element analysis to model and map tissue deformation. The experimental setup played a critical role in the data collection. The design of this setup was carefully evaluated to ensure that the applied forces remained within acceptable thresholds of tissue deformation, aligning with welfare considerations. Mapping tissue deformation during needle insertion is challenging across various applications. In this study, a novel approach was implemented to quantify deformation resulting from needle penetration. Initially, insertion forces were recorded using an actuator-driven needle insertion and three different commercially available force sensors. The force profile gave an overview of when the tissue deformation occurs, followed by a force change due to the presence of a vital tissue layer/vessel. These force profiles were then used as input variables in a simulation model to estimate stress distribution and tissue displacement.

Repetitive measurements taken with different force sensors generated a similar pattern. These low-cost sensors detected important insertion events. However, performance-wise, each of the force sensors had different remarks. The study showed that the load cell observed three important sudden drops and rises in force. These observation numbers are higher than the other two sensors (Force Sensitive Resistor and Force Sensitive Capacitor). A sudden drop and rise in the force profile is an important event indicating the presence of tissue layers. The load-cell sensor recorded a decrease in force when the needle was not in motion. At that interval, the force dropped 0.2 N in 5 s, decreasing 0.04 N every second. According to load cell sensory data, the simulation recorded stress as low as 1.15 N/m² up to 3.9 N/m² and deformation from 1.5 μm up to 3.4 μm. As a result, the load cell provided the most reliable data for measuring stress and tissue deformation among the three force sensors evaluated.

The deformation values observed in this study remain below established thresholds, supporting the goal of a pilot study for a blood collection setup that involves minimally invasive fish health and livestock monitoring. These findings align with welfare standards that emphasize low-impact handling and reduced tissue disruption.

This sensor–actuator integrated simulation-based approach can be applied as a model for similar kinds of studies [33] in the future. Specifically, where tissue stress and deformation are involved. Furthermore, the scope of this study can provide feedback for precise control of automatic insertion while reducing deformation and tissue stress.