PuffyBot: A Robot That Walks, Swims, and Floats

by

Author: Denis Avetisyan

Researchers have developed an untethered, shape-morphing robot capable of seamlessly transitioning between land, underwater, and surface locomotion with improved energy efficiency.

The robot transitions seamlessly between underwater states—from resting on the seabed at $0$ seconds, to initiating body expansion and lift-off by $5$ seconds, achieving positive buoyancy and ascent at $45$ seconds, and finally surfacing completely at $46$ seconds—demonstrating shape-morphing driven buoyancy control.
The robot transitions seamlessly between underwater states—from resting on the seabed at $0$ seconds, to initiating body expansion and lift-off by $5$ seconds, achieving positive buoyancy and ascent at $45$ seconds, and finally surfacing completely at $46$ seconds—demonstrating shape-morphing driven buoyancy control.

This work presents PuffyBot, an amphibious robot leveraging buoyancy control and mechanical coupling for versatile multi-environment navigation.

Existing amphibious robots often struggle with efficient transitions between drastically different environments, demanding complex and energy-intensive mechanisms. This limitation motivates the work presented in ‘PuffyBot: An Untethered Shape Morphing Robot for Multi-environment Locomotion’, which introduces a novel robot capable of dynamically altering its morphology to navigate land, underwater, and surface terrains. By leveraging a scissor-lift mechanism and buoyancy control, PuffyBot achieves seamless locomotion and demonstrates significant potential for energy savings compared to conventional designs. Could this approach pave the way for a new generation of versatile and adaptable robots for exploration and intervention in diverse and challenging environments?

Navigating Complexity: The Challenge of Adaptive Locomotion

Conventional robotic designs frequently encounter difficulties when traversing uneven or mixed terrains, a limitation particularly pronounced when attempting the transition between land and water. These machines, typically optimized for a single environment, lack the morphological and control flexibility required to maintain stability and efficiency across drastically different mediums. Rigid chassis and locomotion systems – whether wheeled, legged, or tracked – often struggle with the unpredictable forces and dynamic changes in buoyancy encountered during amphibious maneuvers. This results in decreased speed, increased energy consumption, and a high risk of instability or complete failure, highlighting the need for fundamentally new approaches to robotic body design and control algorithms that prioritize adaptability over specialized performance.

Current robotic designs frequently exhibit a specialization toward either land or water traversal, a limitation stemming from the inherent difficulties in optimizing a single morphology for such disparate environments. Robots built for terrestrial efficiency often struggle with buoyancy and hydrodynamic drag in water, necessitating inefficient and energy-intensive adaptations for aquatic locomotion. Conversely, designs prioritizing underwater maneuverability typically suffer from reduced ground clearance, poor traction, and limited power transmission for effective movement across uneven terrain. This trade-off results in diminished versatility and operational efficiency, hindering their ability to seamlessly transition between land and water and restricting their usefulness in dynamic, real-world scenarios where both capabilities are crucial.

Effective amphibious robotics hinges on a locomotion strategy capable of real-time environmental assessment and adjustment. A robot encountering a shifting landscape – transitioning from solid ground to shallow water, or navigating uneven, muddy terrain – cannot rely on pre-programmed sequences. Instead, it requires sophisticated sensing capabilities to analyze surface properties, water depth, and frictional forces. This data informs a control system that dynamically modifies gait, body posture, and propulsion method; perhaps alternating between legged walking, rolling, and paddling. Such adaptability isn’t simply about switching between modes, but rather a continuous, nuanced optimization of movement parameters to maintain stability, minimize energy expenditure, and maximize efficiency across a spectrum of conditions. The goal is a system that doesn’t just react to the environment, but anticipates changes and proactively adjusts its locomotion to maintain seamless and robust movement.

Achieving truly seamless amphibious movement requires a departure from conventional robotic designs, demanding innovative approaches to both body configuration and control systems. Researchers are investigating variable geometry robots – machines capable of altering their physical shape – to optimize performance across land and water. This includes exploring designs featuring modular components that reconfigure for different gaits, flexible spines that enable undulatory swimming, and dynamically adjusting limb arrangements. Simultaneously, advanced control algorithms are being developed to coordinate these morphological changes with real-time sensory input, allowing the robot to instantaneously transition between walking, swimming, and other locomotion modes. The goal isn’t simply to create a robot that can operate in both environments, but one that does so with comparable efficiency and agility, blurring the lines between terrestrial and aquatic robotics and opening possibilities for exploration and operation in previously inaccessible regions.

The robot successfully demonstrated versatile locomotion across land, underwater on a pebbled floor, and on the water's surface, maintaining consistent performance over three 60-second trials in each environment.
The robot successfully demonstrated versatile locomotion across land, underwater on a pebbled floor, and on the water’s surface, maintaining consistent performance over three 60-second trials in each environment.

PuffyBot: A Morphing Solution for Multi-Terrain Navigation

PuffyBot is a fully autonomous, mobile robot designed for operation in terrestrial, aquatic, and surface environments. Its untethered design allows for unconstrained movement, and its core functionality centers on adaptive locomotion achieved through dynamic body reconfiguration. This capability enables PuffyBot to navigate varied terrains and fluid dynamics without the need for specialized designs for each environment. The robot’s morphology is not fixed, allowing it to transition between walking, swimming, and floating modes as required by the surrounding conditions, effectively functioning as a multi-modal robotic platform.

PuffyBot’s adaptability stems from its capacity to actively reconfigure its physical structure to suit varying environmental demands. This dynamic alteration of body configuration involves adjusting both overall volume and the orientation of integrated fins. By changing its shape, PuffyBot optimizes its interaction with the surrounding medium – land, water, or air-water interface – to maximize propulsive efficiency and stability. This contrasts with traditional robots that maintain a fixed morphology, requiring separate designs for different terrains or aquatic conditions. The system allows for optimized locomotion by minimizing drag and maximizing thrust or surface area as needed for the current environment.

PuffyBot’s shape-morphing capability is mechanically realized through a Bell-Crank Linkage system coupled with Linear Actuators. The Bell-Crank Linkage provides the structural framework for body reconfiguration, enabling changes in overall volume and the articulation of flexible fins. Linear Actuators, strategically integrated with this linkage, provide the precise and controlled movements necessary to adjust fin orientation and body geometry. This combined system allows for dynamic alterations in shape, facilitating optimized locomotion and buoyancy control across diverse environments. The actuators operate by extending or retracting, directly manipulating the linkage and, consequently, the robot’s physical configuration with a high degree of accuracy.

PuffyBot’s environmental adaptability is achieved through a shape-morphing process requiring less than 7 Joules of energy. This low energy consumption is critical for untethered operation and extended deployment durations. The robot dynamically adjusts its body volume and fin configuration to optimize locomotion and maintain buoyancy across land, underwater, and on water surfaces. This seamless transition between environments is enabled by the integrated Bell-Crank Linkage and Linear Actuator system, which minimizes the energy expenditure associated with adapting to different physical properties and resistive forces.

The robot successfully transitioned between terrestrial and aquatic locomotion, beginning with a ramp descent over 7.5 seconds, followed by submersion and underwater crawling for one body length over 25 seconds, and concluding with buoyancy-controlled ascent and stabilization at 105 seconds.
The robot successfully transitioned between terrestrial and aquatic locomotion, beginning with a ramp descent over 7.5 seconds, followed by submersion and underwater crawling for one body length over 25 seconds, and concluding with buoyancy-controlled ascent and stabilization at 105 seconds.

Synchronized Control: Precision in Movement and Buoyancy

PuffyBot utilizes a Proportional-Integral-Derivative (PID) controller to regulate the actuation of each fin independently, enabling precise shape morphing for locomotion. The PID controller continuously calculates an error value as the difference between a desired fin angle and the actual measured angle, then applies a corrective action to minimize this error. Proportional, integral, and derivative terms within the controller respond to the current error, the accumulated past error, and the rate of change of the error, respectively. This feedback loop ensures synchronized fin patterns, critical for generating thrust and directing movement; deviations are rapidly corrected to maintain the desired shape and optimize propulsive efficiency. The controller parameters are tuned to balance responsiveness and stability, preventing oscillations and ensuring accurate tracking of the desired fin configurations.

PuffyBot utilizes buoyancy control as a key mechanism for aquatic locomotion and vertical positioning. This is achieved by modulating the robot’s internal volume, thereby altering its overall density relative to the surrounding water. Based on $Archimedes’ Principle$, the buoyant force acting on PuffyBot is equal to the weight of the water displaced. By increasing internal volume, PuffyBot decreases its density and rises; conversely, decreasing volume increases density and causes descent. Precise control of this volume allows for stable hovering and efficient navigation through varying water depths, supplementing propulsive forces generated by fin movements.

Open-loop control was implemented in PuffyBot’s actuation system to minimize computational load and energy expenditure. This approach bypasses the need for real-time sensory feedback and complex calculations typically required by closed-loop systems. Instead, pre-programmed actuation sequences, determined through empirical testing and modeling, directly drive the robot’s fin movements. While sacrificing some adaptive capacity, this simplification significantly reduces the processing requirements, allowing for operation with limited onboard resources and extended operational duration. The demonstrated locomotion speeds of 0.70 cm/s on land, 0.75 cm/s swimming, and 0.24 cm/s underwater crawling validate the effectiveness of this open-loop strategy for PuffyBot’s intended applications.

Quantitative analysis of PuffyBot’s locomotion was performed using ArUco marker-based tracking to determine speeds across various terrains. Results indicate a land-based locomotion speed of 0.70 cm/s, a swimming speed of 0.75 cm/s, and an underwater crawling speed of 0.24 cm/s. These measurements were obtained through repeated trials and represent the average velocity achieved by the robot while utilizing its shape-morphing locomotion method. The ArUco marker system provided a precise method for quantifying robot position and movement, enabling accurate speed calculations.

The robot consistently sank to the tank bottom within seven seconds and resurfaced approximately sixty seconds later through shape morphing, with observed depth variations attributed to pebbles on the tank floor.
The robot consistently sank to the tank bottom within seven seconds and resurfaced approximately sixty seconds later through shape morphing, with observed depth variations attributed to pebbles on the tank floor.

Durability and Efficiency: Design for Real-World Deployment

PuffyBot’s capacity to function reliably in aquatic environments stems from a design prioritizing both flexibility and impermeability. The robot’s exterior employs a thermoplastic polyurethane (TPU) material, chosen for its inherent waterproof qualities and remarkable elasticity. This allows PuffyBot to not only repel water, safeguarding internal components, but also to maintain operational capacity even when subjected to bending and compression underwater. The flexible nature of the TPU enables effective sealing around critical areas, preventing water ingress during dynamic movements and ensuring sustained performance in challenging, wet conditions. This robust waterproofing is crucial for applications ranging from environmental monitoring in waterways to underwater inspection and exploration.

PuffyBot’s extended operational lifespan stems from a deliberate focus on minimizing energy expenditure. The robot utilizes highly efficient actuators, carefully selected to deliver precise movements with minimal power draw. Complementing this hardware optimization is a streamlined control system; algorithms have been refined to eliminate unnecessary computations and prioritize energy-conscious path planning. This synergistic approach—combining efficient mechanics with intelligent software—allows PuffyBot to maintain functionality for significantly longer durations than comparable robots, opening possibilities for prolonged data collection and sustained performance in remote or resource-limited environments. The resulting low power consumption isn’t merely a design feature, but a core element enabling practical, real-world applications.

Fabrication of the PuffyBot relies entirely on additive manufacturing, more commonly known as 3D printing. This approach grants researchers and engineers an unprecedented level of design freedom and accelerates the development cycle. Complex geometries, previously difficult or impossible to achieve with traditional manufacturing methods, become readily attainable. More importantly, 3D printing allows for highly customized iterations; components can be rapidly prototyped, tested, and refined, or even tailored to specific environmental demands or task requirements without the significant tooling costs associated with conventional production. This adaptability is crucial for exploring a wide range of applications and optimizing the robot’s performance in diverse, real-world scenarios, ultimately lowering barriers to innovation and deployment.

The design of PuffyBot prioritizes complete autonomy through untethered operation, fundamentally expanding its potential applications beyond the constraints of wired systems. This freedom of movement allows the robot to navigate complex and previously inaccessible environments, such as conducting underwater inspections, performing search and rescue operations in disaster zones, or autonomously monitoring remote agricultural fields. By eliminating the need for a physical connection, PuffyBot isn’t limited by cable length or susceptible to entanglement, offering a level of versatility crucial for real-world deployments. This capability shifts the focus from power and signal transmission to pure robotic function, opening doors for innovative uses in environments where wired robots are impractical or impossible to operate effectively.

The robot exhibits varying buoyancy states depending on its volume, transitioning from negative buoyancy when the weight force exceeds the buoyant force, to neutral buoyancy when they are equal, and to positive buoyancy when the buoyant force dominates the weight.
The robot exhibits varying buoyancy states depending on its volume, transitioning from negative buoyancy when the weight force exceeds the buoyant force, to neutral buoyancy when they are equal, and to positive buoyancy when the buoyant force dominates the weight.

The design philosophy behind PuffyBot exemplifies a commitment to holistic systems thinking. The robot’s ability to seamlessly transition between locomotion modes—land, surface, and underwater—isn’t achieved through isolated innovations, but rather a unified approach to mechanical coupling and buoyancy control. This interconnectedness mirrors the principle that structure dictates behavior. As David Hilbert noted, “One must be able to say at any time exactly what is known and what is not.” PuffyBot’s predictable and adaptable locomotion arises from a clearly defined structure, allowing for efficient energy expenditure and a reduction in complexity. The robot’s simplicity scales, unlike designs reliant on intricate, specialized mechanisms for each environment.

Where the Air Meets the Water

The elegance of PuffyBot lies in its simplicity – a rejection of brute force in favor of adaptive form. However, true multi-environment locomotion isn’t merely a question of transitioning between domains, but of integrating them. Current designs, even this one, treat land, sea, and surface as discrete challenges, solved sequentially. A more holistic approach would necessitate a fundamental rethinking of structural mechanics – a morphology that isn’t optimized for any single environment, but possesses inherent resilience across all. The robot’s reliance on pre-defined morphological states, while efficient, suggests a limitation in its capacity to respond to truly novel or unpredictable conditions.

The presented buoyancy control, though effective, remains mechanically coupled to the shape-morphing system. Disentangling these functions—allowing for independent adjustments of volume and form—could unlock a degree of freedom currently unavailable. This decoupling would allow for more nuanced responses to dynamic currents and complex terrain. Moreover, the energy gains, while significant, are predicated on relatively benign environments. The true test will be performance in turbulent conditions or highly variable densities – situations where the system’s structural integrity and responsiveness will be most severely challenged.

Ultimately, the path forward isn’t simply about building more sophisticated robots, but about understanding the fundamental principles governing adaptation and resilience in natural systems. PuffyBot offers a promising starting point, a demonstration that less can indeed be more. But the ocean, and the world beyond, demands not just clever engineering, but a deeper appreciation for the interconnectedness of form and function.

Original article: https://arxiv.org/pdf/2511.09885.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/

See also:

2025-11-15 16:26