Design and Analysis of a Dual-Screw Propelled Robot for Underwater and Muddy Substrate Operations in Agricultural Ponds

Abstract

Conventional underwater vehicles, which are typically equipped with oscillating fins or standard propellers, are incapable of effective locomotion within the viscous, high-resistance environment of muddy substrates common in agricultural ponds. To address this operational limitation, this paper presents a compact dual-screw propelled robot capable of traversing both the water column and soft substrate layers. The robot’s locomotion is driven by two optimized helical screw propellers, while depth control and roll stability are actively managed by a control fin. A dynamic model of the robot–fluid interaction was developed to optimize the screw configuration that achieves a maximum theoretical thrust of 40 N with a calculated 16% slippage rate in mud. Computational fluid dynamics simulations were employed to determine the optimal angle for the control fin, which was found to be 9°, maximizing the lift-to-drag ratio at 12.09 for efficient depth maneuvering. A cable-free remote control system with a response time of less than 0.5 s governs all operations. Experimental validation in a controlled tank environment confirmed the robot’s performance, demonstrating stable locomotion at 0.4 m/s in water and 0.3 m/s in a simulated mud substrate. This dual-screw propelled robot represents a promising technological solution for comprehensive monitoring and operational tasks in agricultural pond environments.

1. Introduction

Pond-based planting and aquaculture [1,2] are vital components of modern agricultural production, requiring consistent monitoring of plant and animal health, water quality, and substrate conditions [3]. Traditionally, these tasks have been performed manually, a method that is not only labor-intensive and costly but also poses safety risks to personnel. The deployment of robotic systems equipped with sensors offers a promising alternative for automated monitoring and intervention in these challenging environments.

The field of underwater robotics has seen significant advancements, with a wide range of applications in geological surveys, aquaculture, and underwater sampling [4,5]. He Jianhui et al. designed an underwater machine fish with multiple rudder tail fins combined with a center of gravity adjustment mechanism, which can effectively reduce the intensity of manual operation. When the steering gear mechanism and the center of gravity regulating mechanism are applied, the machine runs smoothly underwater [6]. Li Hailong designed a modular snake robot that can make the robot move flexibly under water [7]. Wu Chao designed a cabled underwater robot that uses machine vision technology to recognize marine organisms, and this research provides a new idea for underwater biomonitoring devices [8]. Cheng Liangliang designed an underwater robot that imitates the way a squid swims and verified the feasibility of the tubular origami structure of the underwater robot [9]. Danie et al. studied a new unmanned underwater vehicle simulator, Simu2VITA, which has the characteristics of a fast setup, ease of use, and simple parameter modification [10]. Su Bo et al. investigated an event-triggered integral sliding mode timing control method for the trajectory tracking problem of an autonomous underwater vehicle (AUV) with perturbation. By introducing a fixed-time integral sliding mode manifold, a fixed-time control strategy is expressed for the AUV, which can effectively eliminate the singularity [11]. Hirpara Ravish H et al. implemented nonlinear filtering via the Kolmogorov backward equation and the evolution of the conditional characteristic function. This paper introduces nonlinear filtering theory into an underwater vehicle stochastic system by constructing a lemma and a theorem for the underwater vehicle stochastic differential equation that are not available in the literature. This approach is highly important for the development of pond robots [12]. Zhao Wanting et al. designed a convolutional neural network accelerator for real-time underwater image recognition of autonomous underwater vehicles. By sending control signals to autonomous underwater vehicles, this accelerator can avoid dangerous areas such as rocks and algae in real time [13]. Michał P. used two classical controllers, namely, proportional integral-differential (PID) and sliding mode (SM) controllers, in the process of optimizing the depth controller settings to increase the precision of the depth control [14]. To obtain clear fishing net detection images, Hoosang L. proposed an attitude control strategy for underwater robots on the basis of the average gradient features of local or overall images. This method keeps the fishing net at a relatively constant position, which is sufficient for obtaining clear images of the fishing net in cloudy water and determining whether there is partial damage to the fishing net [15].

However, a critical limitation of existing AUVs is their reliance on propellers or oscillating fins, which are designed for operation in low-viscosity fluids like water. These propulsion methods become ineffective in the soft, viscous mud and substrate layers at the bottom of agricultural ponds. The high resistance and poor fluidity of mud cause conventional propellers to cavitate or clog, resulting in a complete loss of thrust. This inability to penetrate and traverse muddy substrates prevents comprehensive monitoring of benthic organisms, root systems of aquatic plants, and substrate conditions, which are crucial for pond ecosystem health.

Separately, screw-propulsion technology has been extensively studied for vehicles operating on soft, challenging terrains such as sand, snow, and swamps. The fundamental principle involves using rotating helical blades to generate thrust through shearing and displacement of the medium. Research has provided theoretical support for the steering and propulsion dynamics of these vehicles and has led to the development of amphibious vehicles capable of navigating swampy land by combining screw reels and propellers. Specialized agricultural machinery, such as lotus root diggers, has also successfully employed dual-screw propulsion chassis to operate in flooded, muddy fields.

Guo Xiaolin et al. combined the research progress on screw propulsion vehicles and studied the steering theory of helical propulsion vehicles, which provided theoretical support for research on screw propulsion vehicles [16,17]. Liu Tianao designed a screw-propelled vehicle that uses a screw propeller as the walking mechanism and implements forward movement through the rotation of the inner and outer sides of screw rollers, which can effectively solve the transportation problem of wet and soft ground [18]. Wan Yuchun et al. designed a new type of catamaran adopting the double propulsion mode of screw reel and propeller, which solved the problem of existing ships being difficult to adapt to the driving of swampy land [19]. Feng Chuangchuang and Liu Yan carried out research on double-screw propulsion-type lotus root diggers and designed different forms of double-screw drive chassis, which has reference significance for the form of double-screw propulsion in this paper [20,21]. Zhang Hongwei et al. proposed an underwater vehicle thruster program that uses pitch change to realize lateral force output. The thruster has a main thrust motor, a pitch control motor, and a gear set, and the pitch control shaft is coaxially mounted with the propulsion main shaft so that the pitch angle of the paddle can be periodically changed by controlling the motor velocity, thus realizing the lateral force output [22]. Pan Shixuan used the 6-DOF model of FLUENT software to simulate the motion position and pose changes of underwater vehicles during driving, providing a reference for the optimization of the external structure strength, internal mass structure distribution, and attitude control of underwater vehicles [23]. V. K. and Prabhu R. optimized the steering performance of a continuous split shell underwater vehicle and proved the existence of an optimum turning configuration, i.e., minimum demand for lateral thrust compared with axial thrust [24]. Savu T. et al. designed a system composed of encoders for parameter measurement of a screw propeller, which can provide the most accurate conclusion on the subsequent repair of the screw propeller [25]. Przemysław K. discussed the design and operation of rotor–stator thruster systems. The results obtained by installing energy-saving devices on specific ships were summarized to obtain higher efficiency gains at higher propeller loads by installing ESD [26]. Marko G. et al. developed and analyzed a new screw-propelled excavation robot, CASPER, to determine its effectiveness as a planetary excavator [27]. Wencan Z. et al. used a new method, OpenProp, in conjunction with computational fluid dynamics to design a propeller for an Explorer100 AUV, which increased the cruising velocity by 21.8% compared with the original propeller of the AUV self-propelled experiment [28]. Dongqi G. et al. developed an AUV equipped with a hybrid thruster that maintains fast driving capability while greatly improving maneuverability [29]. The Archimedes screw principle has been applied to amphibious platforms for many years. This has led to the development of powerful commercial systems, such as Copperstone Technologies’ Hessel series. It can operate in water, mud, and snow and is mainly used for industrial site surveys [30,31,32,33].

While the principle of amphibious screw propulsion is established, a specific gap remains for compact, low-cost systems tailored to the unique constraints of agricultural pond environments. Existing commercial platforms are typically large-scale and designed for industrial applications, whereas conventional AUVs lack substrate-traversal capabilities. This paper addresses this gap by presenting a novel amphibious robot whose contribution is threefold: (1) a specific focus on agricultural ponds, requiring a design that is compact, maneuverable in cluttered spaces, and cost-effective for this application; (2) a miniaturized form factor, featuring a 50 mm diameter fuselage, which is substantially smaller than existing industrial amphibious robots; and (3) the numerical optimization of an active control fin, a critical innovation for enabling efficient depth control and stability in such a small-scale vehicle. This paper details the robot’s structural design, the dynamic modeling of its unique propulsion system, the hydrodynamic optimization of its control surfaces via simulation, and the experimental validation of its performance, presenting a viable solution for the multifaceted challenges of agricultural pond monitoring.

2. Design of the Dual-Screw Pond Robot

2.1. Design of Overall Structure

Based on the actual pond conditions and the measured physical properties of the mud, the maximum working depth of the robot’s mud-layer operation is set to 20 cm. Following the above analysis of the robot’s motion mode and adhering to the principle of minimizing the overall mechanical dimensions, the pond robot’s body is initially designed with an outer diameter of 50 mm and a length of 400 mm.

According to the predefined dimensions of each component necessary for the robot to perform its target functions, the maximum linear velocity of the robot in water is set at 0.4 m/s, while the maximum steering, surfacing, and diving velocity is 0.3 m/s. In the mud layer, the maximum linear velocity is limited to 0.3 m/s, and the maximum velocity for steering, surfacing, and diving is set to 0.1 m/s. The parameters of the robot are shown in

Table 1. Parameters of the pond robot.

Parameters	Values
Robot body diameter (mm)	50
Robot length (mm)	400
Maximum operating depth in mud layer (cm)	20
Maximum straight-line speed in water layer (m/s)	0.4
Maximum steering, surfacing and diving speed in water layer (m/s)	0.3
Maximum straight-line speed in mud layer (m/s)	0.3
Maximum speed of steering, surfacing, and diving in mud layer (m/s)	0.1

.

To ensure that the pond robot can maneuver flexibly within the mud layer while maintaining adequate compressive strength, a cylindrical fuselage is adopted. A segmented modular design is selected for the overall scheme, enabling functional upgrades or reconfigurations by adding, removing, or rearranging modules according to operational requirements.

Based on the analysis of the robot’s locomotion modes, DC motors are selected to drive dual-screw propellers. By manipulating the rotational direction and speed of the DC motors, the robot’s movement direction is regulated to achieve steering, forward, and reverse functions. A control fin is installed to control the robot’s surfacing and diving motions through adjustment of its yaw angle. Steering maneuvers are accomplished by combining yaw angle adjustments of the control fin with directional and speed control of the screw propellers. The overall architecture of the pond robot is shown in Figure 1. During operation, power is activated, and control signals are transmitted from the transmitter terminal. These signals undergo processing within the robot’s circuitry before being routed to the screw propeller and control fin control systems, enabling the pond robot to execute targeted actions.

The justification for selecting a dual-screw propulsor stems from a critical analysis of conventional underwater propulsion systems and their limitations in the complex, multi-phase environment of agricultural ponds. Existing propulsors exhibit significant trade-offs: propellers offer high efficiency and velocity in open water but suffer from severe clogging, loss of traction, and substrate disturbance in muddy substrates; oscillating fins provide good maneuverability in water yet generate negligible thrust in dense mud due to the lack of solid reaction forces. This environmental dichotomy necessitates a hybrid solution. The proposed dual-screw propelled robot offers practical improvements over existing solutions for the specific challenge of agricultural pond monitoring, which requires operation both in the water column and within soft, muddy substrates.

2.2. Design of Screw Propeller

The propulsion system consists of two counter-rotating helical screw propellers. The cross-sectional shape of the screw blades is a vertical screw surface, which ensures that the propulsive effect is symmetrical for both forward and reverse motion. To minimize hydrodynamic and substrate resistance during forward motion, the head cone of each propeller was optimized. The shape of the head cone follows the von Kármen ogive, a curve known to produce minimal drag for a given length and diameter. Its expression is shown in Equation (1):

$$\begin{array}{l}\theta ={\cos }^{-1}\left(1-{\displaystyle \frac{2x}{L}}\right)\\ y={\displaystyle \frac{R}{\sqrt{\pi }}}\sqrt{\theta -{\displaystyle \frac{\sin (2\theta )}{2}}+C{\sin }^3\theta }\end{array}$$

(1)

To meet the robot size requirements, the head cone length is set to 40 mm, and the diameter is 50 mm. The resulting head cone curve obtained for planning and solving is shown in Figure 2a, and the final design of the head cone three-dimensional diagram is shown in Figure 2b.

Following the optimization of screw propeller cross-sectional profiles and head cone curve calculations, the pond robot employs a dual-screw propulsion system where directional control including forward/reverse motion and steering is achieved through coordinated motor rotation reversal. The propeller configuration was determined through comprehensive analysis of total motion resistance and operational requirements, requiring that the resultant force vectors converge at the robot’s center of gravity for optimal control responsiveness, while minimizing body roll induced by propeller torque, reducing propeller count to the functional minimum, and aligning propeller axes parallel to the motion coordinate system for maximum efficiency. As shown in Figure 3, clockwise rotation of propeller 1 with simultaneous counterclockwise rotation of propeller 4 generates positive x-axis locomotion, while reversed rotations produce negative x-axis displacement, enabling precise bidirectional operation.

2.3. Design of Control Fin

To enable the pond robot to perform surfacing, diving, and steering maneuvers, a control fin is mounted on the fuselage. The operational principle governing the robot’s upward surfacing and downward diving is illustrated in Figure 4, where G is the gravity of the robot, V_m is the running velocity, F_b is the buoyancy, F_t is the thrust, R is the total resistance, and L is the distance between the center of gravity and the center of the control fin.

Thrust from the dual-screw propellers imparts forward velocity to the body, creating relative motion between the mud and the body. As mud flows over the adjustment plate in its initial position (0° angle relative to the direction of motion), the motion resistance between the body and plate does not generate a component force along the negative z-axis. However, when the plate rotates counterclockwise to a specific angle, this resistance resolves into a force component fz along the positive z-axis. This disrupts the initial equilibrium, creating a counterclockwise torque about the robot’s center of gravity due to fz. This torque enables the surfacing function. Consequently, the body undergoes clockwise rotation about an axis parallel to the y-axis, completing the surfacing maneuver.

Another key function of the control fin is to prevent the pond robot from rolling over. However, during turning maneuvers, the robot first executes a controlled rolling motion, the principle of which is illustrated in Figure 5, where $f$ is the resistance of control fin by the mud–water, $\tau$ is the tangential drag between the rotating blades of the propeller and mud–water, and $F$ is the resistance perpendicular to the helical blades of the propeller rotating blades by the mud–water.

The screw propeller is connected to the motor via a positive hexagonal shaft. During turning maneuvers, vertical stability is maintained through the balance of gravity and buoyancy. As the propeller rotates, a tangential resistance (τ) is generated between the rotating blades and the surrounding mud–water mixture. In addition, the resistance perpendicular to the screw blades, together with the resistance (f) generated by the interaction between the control fin and the mud–water, achieves radial equilibrium, enabling axial overturning. After the overturning process, the control fin is reoriented in the vertical plane, and the forces involved in left and right steering are analogous to those governing surfacing and diving. The final configuration of the control fin is illustrated in Figure 6.

2.4. Design of Control System

According to the hardware equipment required for control system control and motion decomposition, the overall scheme of the control system can be obtained. The control program is the program executed after the two control circuit boards of the robot are energized. This program runs through the entire control system. The flowchart of the control procedure is shown in Figure 7.

The TX-2 and RX-2 chips were selected for the transmitter and receiver circuits, respectively, following a rigorous evaluation of performance and cost against the essential requirements for reliable robotic control in demanding experimental conditions. These cost-effective chips provide the necessary rapid response for precise actuator control and integrate multiple functions required for operation in high-substrate environments, all achieved with minimal implementation complexity. Their compact footprint satisfies the spatial constraints within the robotic chassis. As illustrated in Figure 8, this solution maintains operational stability and efficiency in rigorous muddy experimenting scenarios, concurrently meeting all critical criteria for response velocity, multifunctional control, cost-effectiveness, and miniaturization essential for both the internal circuitry and remote control system.

To achieve unrestricted mud penetration capability for the pond robot, this study employs standardized power components selected against operational requirements. Critical elements include two screw-drive motors, one control fin actuator, and dedicated power supplies. All operational components are summarized in

Table 2. Component specification summary.

Component	Mass (g)	Dimensions (mm)	Function
F130 motor	14	38 × 20 × 15	Dual-screw propulsion drive
Drive battery	8	38 × 19 × 8	F130 motor power supply
Control battery	10	Ø14 × 51	Transmitter circuit power
N20 motor	8	30 × 15 × 13	Control fin actuation

.

3. Dynamic Analysis of the Dual-Screw Propeller

3.1. Dynamic Analysis of the Double-Screw Propeller in Water or Mud

When the robot operates in a single water and mud environment, the force situation is consistent, with only a difference in the force size. After analysis, the interaction between the double-screw propeller and the mud robot in the water layer or the mud layer is shown in Figure 9, where $M$ is torque, $A$ is the long axis distance of the ellipse changes with the helical angle of the double helical propellant, $N$ is the resistance generated by the screw propeller blade, $V$ is the velocity, and $\omega$ is the rotational velocity of the screw propeller.

When the robot is running at a uniform velocity in the water layer or mud layer, the motor rotates and drives through the hexagonal shaft to finally form a force balance relationship, which can be expressed as Equation (2):

$$\left\{\begin{array}{c}{F}_{\mathrm{t}}=N\cdot \sin e+{\mathrm{F}}_f+R\\ F_{\mathrm{b}}=G+N\cdot \cos e\end{array}\right.$$

(2)

3.2. Dynamic Analysis of the Crossing Process of the Double-Screw Propeller

During the crossing of the robot in the water into the mud layer, with the compaction and subsidence of the soil in the mud, the body experiences a slip phenomenon, and the slip ratio can be expressed as Equation (3):

$$i_x={\displaystyle \frac{Sn/(2\pi )-v_x}{Sn/(2\pi )}}$$

(3)

where $v_x$ is the velocity in the x-axis direction as the pond robot moves, $S$ is the pitch, and $n$ is the rotational velocity.

When the robot is in the driving state, 0 ≤ ix ≤ 1. In the braking state, −1 ≤ ix ≤ 0. According to the ground mechanics theory of engineering machinery, the interaction between a double-screw propeller and a robot in the process of crossing from the water to the mud layer is shown in Figure 10, where $e_1$ is the effective unearthed angle of the double-screw propeller (<0), rad; $e_2$ is the drum theory unearthed angle (<0), rad; $e_3$ is the drum into the soil angle (>0), rad; $e$ is the double-screw propeller helical angle, where e₁ ≤ e ≤ e₃, rad; $h$ is the screw thruster subsidence depth, m; $\sigma$ is the positive stress of mud, Pa; and $\tau$ is the shear stress of the mud, Pa.

The mud-generated positive stress σ distribution of the double-screw propeller during operation is shown in Equation (4):

$$\sigma =\left({\displaystyle \frac{k_1}{l_s}}+k_2\right)A^{k_3}\cdot \left(\cos e-\cos e_3\right),$$

(4)

where $k_1$ is the soil cohesion modulus, taken as $k_1$ = 0.9; $k_2$ is the machine internal friction modulus, taken as $k_2$ = 1.2; and $k_3$ is the subsidence coefficient, taken as $k_3$ = 1.1.

According to Figure 11, the following geometric relationship can be obtained:

$$e_3={\displaystyle \frac{1}{\cos \left(1-h/\frac{D}{2}\right)}}$$

(5)

According to the theory of ground mechanics, the double-helix propeller satisfies the relationship that the absolute value of the effective unearthed angle is less than the absolute value of the drum relative to the soil angle. At this time, the effective unearthed angle can be expressed as Equation (6):

$$e_1=-k_4{e}_3\left(0<k_4<1\right)$$

(6)

The slip ratio under experiment conditions is definite, so for a double-screw propeller, the amount of subsidence can be expressed as $h=h_0+k_5{i}_x$ ($k_5$ > 0). The specific value of $k_5$ depends on the motion between the shell and the mud, which is related to the physical properties of the shell. The calculated ix = 0.16.

When the double-screw propeller is running, the screw blade has a shear effect on the mud, the shear trajectory can be simplified to an ellipse, and the simplified shear trajectory is shown in Figure 11.

The thrust of the robot during the process of diving from the water layer to the mud layer is mainly generated by the shear stress of the helix blade shearing the mud. The shear stress of the screw blade on mud can be expressed by the Cullen Equation (7) and the Janosi Equation (8):

$$\tau ={\tau }_{\max }\left(1-e^{-s_1/k_6}\right),$$

(7)

$${\tau }_{\mathit{max}}=F_i+\sigma \tan \alpha ,$$

(8)

where $F_i$ is the cohesive force of the mud, N; $\alpha$ is the friction angle in mud, rad; $S_1$ is the mud shear displacement, m; and $k_6$ is the mud shear modulus, Pa.

According to the shear trajectory diagram of the screw blade, the geometric relationship between the various parameters can be obtained, and the shear displacement Equation (9) of the double-screw propeller can be obtained:

$$S_1={\int }_{s \mathrm{total}}v\cdot dt={\displaystyle {\int }_0^e\sqrt{S_x^2+S_y^2+S_z^2}}de$$

(9)

Displacement along the three coordinate axes $S_x$, $S_y$, and $S_z$ can be measured from the shear trajectory diagram, where the shear trajectory is determined by the helical rise angle of the screw propeller blade. The screw rise angle also satisfies Equation (10):

$$S={\displaystyle \frac{\pi D\cdot \tan e}{2}}$$

(10)

In conclusion, the effective thrust of the final double-screw propeller is the result of the positive stress and shear force of the propeller in the x-axis direction, and the effective thrust can be expressed by Equation (11):

$$F_t={\displaystyle \iint \left(\tau \cdot \cos e-\sigma \cdot \sin e\right)}dAde,$$

(11)

where e is the angle between the tangent line of the shear trajectory of the screw blade and the x-axis, rad.

After analysis and calculation, the slip ratio of the double-screw propeller is 16%. The relationships between the positive stress and subsidence depth, shear force, and shear displacement of the double-screw propeller are verified, and the maximum effective thrust of the screw propeller is 40 N.

4. Robot–Fluid Interaction Simulation

To analyze the hydrodynamic performance of the robot during surfacing and diving maneuvers, particularly the relationship between the control fin angle and the lift-drag ratio for efficient surfacing, computational fluid dynamics (CFD) simulations were conducted using the ANSYS Fluent module (version 19.0). The simplified full 3D CAD model of the robot was imported into ANSYS, and the computational domain for the surfacing/diving mechanism was established as shown in Figure 12a. A tetrahedral structured mesh was generated for the domain, with local refinement applied specifically around the control fin to ensure accuracy in capturing critical flow features. The results indicated that the mesh with 1.6 million elements provided a stable solution as further refinement yielded negligible changes (<1%) in the calculated lift-drag coefficients. The final mesh consisted of approximately 1.6 million elements, achieving an average mesh quality metric of 0.82, as depicted in Figure 12b.

For the fluid domain, water was modeled as an incompressible Newtonian fluid. The physical properties of water were defined at a constant temperature of 25 °C under standard atmospheric pressure, reflecting typical conditions in agricultural ponds. The pressure-based solver was employed for the simulations. To accurately model the turbulent flow encountered by the robot, the Realizable k-epsilon (k − ε) turbulence model with standard wall functions was selected, chosen for its established robustness and reliability in predicting the types of attached and moderately separated flows relevant to this underwater application. The boundary conditions were set as follows: the inlet was defined as a velocity inlet with the magnitude set to the robot’s preset maximum underwater operational velocity of 0.4 m/s, the outlet was specified as a pressure outlet with gauge pressure set to 0 Pa, and the surfaces of the robot model were treated as no-slip walls. The boundaries of the computational domain itself were configured with symmetry or slip-wall conditions as appropriate for the setup. To simulate the rotation of the control fin, a dynamic mesh approach was implemented. The robot structure was modeled as a rigid body made of ABS engineering plastic, with material properties including a density of 1.05 × 10³ kg/m³, an elastic modulus of 2 × 10⁹ Pa, and a Poisson’s ratio of 0.4 [34]. A total of eight distinct simulation cases were executed, corresponding to discrete control fin deflection angles ranging from 1° to 15°, specifically 1°, 3°, 5°, 7°, 9°, 11°, 13°, and 15°.

According to the preset corresponding parameters, when the deflection angle of the control fin is 9° and the forward velocity is 0.4 m/s, the water velocity around the control fin and the pressure of the control fin in the pond robot running in the water are shown in Figure 13.

Figure 13 shows that the control fin upper and lower surface velocity distributions are reasonable, and the upper surface velocity overall is greater than the velocity of the lower surface. The pressure on the upper surface is less than that on the lower surface, so the maximum velocity difference occurs at the very front of the control fin, indicating that most of the lift is provided to achieve the expected design effect.

After eight simulated 3D models of this machine were built via SolidWorks software (2016), the results obtained by changing the deflection angle of the control fin are displayed in Figure 14.

The figure shows that both the lift coefficient and the drag coefficient increase with deflection angle of the control fin; the lift coefficient and the drag coefficient satisfy Equations (12) and (13), respectively, and the lift-drag ratio is the ratio of the lift coefficient to the drag coefficient.

$$Y=C_y\rho v^{2}S/2,$$

(12)

$$f=C\rho A{v}^2/2,$$

(13)

where $C_y$ is the lift coefficient and where $C$ is the drag coefficient.

According to Figure 14, the lift coefficient increases more gently, and the resistance coefficient increases more violently, which satisfies the general motion law of the pond robot. When the deflection angle is 9°, the lift-resistance ratio is set to the maximum value of 12.09. During the operation of the robot, the range of the machine and the lift-drag ratio are positively proportional to each other, so the effect of surfacing is maximum when the deflection angle of the control fin is 9°.

5. Experiments

A series of experiments were conducted with a physical prototype to validate the robot’s performance against the design specifications and simulation results. All experiments were performed in a laboratory water tank (0.3 m × 0.07 m × 0.1 m) (Figure 15) containing a section with a 30 cm deep layer of simulated mud. The simulated mud was prepared using a mixture of bentonite clay and water to achieve physical properties representative of agricultural pond substrate, with a measured density of approximately 1750 kg/m³ and an average dynamic viscosity of 50 Pa·s. A high-velocity camera was used to record the robot’s motion for subsequent analysis.

(1) Response time experiment of the dual-screw pond robot

The response time experiment of the control circuit is conducted to evaluate the performance of the control circuit. The design requirement stipulates that the robot must respond to commands within 0.5 s after the launch button is pressed. The primary parameters measured are the thruster start-up time and the transition time between forward and reverse rotation. As shown in Figure 16, the response times of the control circuit are all ≤0.5 s, demonstrating that the circuit meets the design control requirements.

(2) Straight-line driving experiment in water environment

A primary functional objective of the robot is to maintain straight-line driving at constant velocity. This experiment evaluates trajectory straightness and velocity stability during linear motion, as shown in Figure 17. A high-velocity camera records the motion trajectory, specifically measuring deviations from the predetermined straight path.

The relationship between the robot driving velocity in water and the yaw angle is shown in Figure 18, where negative values indicate leftward yaw and positive values indicate rightward yaw. As shown in Figure 18, the yaw angle of the robot fluctuates around 0°; however, it remains below 10°, indicating that the robot maintains stable forward motion and satisfies the design requirements.

(3) Turning maneuver experiment in water environment

The secondary functional objective involves executing turning maneuvers while maintaining constant velocity. This experiment assesses achievable turning radius and velocity stability during directional changes, as illustrated in Figure 19. From a designated starting position, the robot executes three turning operations upon command activation. High-velocity videography captures each trajectory, enabling calculation of turning radius and measurement of maneuver duration.

As illustrated in Figure 20, with increasing forward distance, both the turning time and turning radius exhibit variations; however, these changes remain minimal, indicating that the robot maintains relatively smooth performance throughout the turning process.

(4) Surfacing and diving experiment in benthic substrate

The tertiary functional objective requires vertical mobility through substrate layers while maintaining forward velocity. This experiment quantifies vertical displacement magnitude and velocity stability during depth transitions, as depicted in Figure 21. The robot initiates motion from a starting position with attitude adjustment plates angled between 0° and 9° relative to horizontal. Due to optical limitations within the substrate, reference markers mounted on the chassis are tracked via high-velocity videography. Trajectory analysis determines the vertical displacement depth.

Figure 22 demonstrates the effectiveness of the control fin for depth control. By varying the deflection angle, the robot achieves controlled changes in diving depth, transforming the horizontal screw thrust into a vertical force component. This confirms the fin’s critical role in enabling diving operation.

6. Conclusions

This study successfully designed, analyzed, and experimentally validated a novel dual-screw propelled robot specifically for navigating the challenging underwater and muddy substrate environments of agricultural ponds, overcoming the limitations of conventional propeller or fin-based underwater vehicles. The robot incorporates an optimized dual-screw propulsion system, generating a maximum thrust of 40 N with 16% slippage. It further utilizes a control fin set at a 9° angle, maximizing the lift-to-drag ratio to 12.09 for effective depth control and roll stability. Experimental validation in controlled tank experiments demonstrated effective locomotion at velocities of 0.4 m/s in water and 0.3 m/s in mud, governed by a cable-free control system achieving a response time below 0.5 s. These results confirm the robot’s capability for comprehensive pond monitoring tasks requiring operation within muddy substrates. To address practical challenges and move beyond controlled lab conditions, future work will focus on real-world validation. This entails evaluating driving stability and sampling performance in working agricultural ponds and aquaculture environments, under varied natural substrate conditions and in obstacle- and vegetation-rich settings. It will also assess long-term reliability under potential biofouling and corrosion, paving the way for autonomous monitoring via integrated advanced sensors.