Formation control dictates how a group of robots maintains a specified spatial arrangement while moving collectively towards a designated goal. Traditional methods often assume static reference signals; however, as real-world scenarios involve fluctuating conditions and requirements, the necessity for time-varying references is paramount. The approach proposed by the researchers effectively addresses this need, ensuring that robots can adapt instantaneously to changes in their operating environment, leading to enhanced performance and reliability.

Directed graphs are utilized as a framework for modeling the interaction and communication between networked robots. Each node represents a robot, while the edges signify the directed relationships, showcasing how information is transmitted within the network. The introduction of directed graphs in their optimization approach allows the robots to achieve distributed control, meaning each robot can take autonomous actions based on local information, thus boosting overall system robustness and adaptability.

Moreover, the finite-time aspect of their control strategy distinguishes this work from traditional infinite-time approaches that often require prolonged periods for settling into the desired formation. This finite-time convergence guarantees that robots reach their target configurations swiftly, an essential feature when dealing with real-time applications such as disaster response, search and rescue missions, and collaborative exploration, where every second counts and rapid decision-making is crucial.

To ensure the effectiveness of the proposed control mechanism, the authors provide a comprehensive mathematical framework. Their process begins with the establishment of dynamic models that define how each robot should adjust its position in relation to others based on the prevailing conditions dictated by the time-varying reference signals. The development of algorithms to solve these dynamic equations forms the backbone of their research, catering to the complexities that directed graphs introduce.

In addition to the formulation of the control strategy, the researchers conducted extensive simulation experiments to validate their theoretical findings. By modeling various scenarios with multiple robots and diverse reference signal conditions, they were able to assess the efficacy of their approach robustly. The results demonstrated not only the feasibility of their method but also its superiority when compared to existing models—highlighting faster convergence times and greater resilience to network disruptions.

This advancement is especially pertinent when considering how robotic formations can be utilized in challenging and unpredictable environments. Whether it’s swarming drones for environmental monitoring or coordinated underwater robots for ocean exploration, the ability to maintain an optimal formation quickly and efficiently under variable conditions is a game-changer. Zhao, Chen, and Ding’s research has laid the groundwork for these applications by presenting a model that can continually adapt, ensuring that robotic teams function seamlessly.

Beyond practical applications, the implications of this research extend into theoretical domains within areas such as control theory and networked systems. The engagement with directed graphs offers fresh insights into how more complex relationships among robots can be harnessed to improve performance metrics. The work also sheds light on potential future explorations into multi-robot systems, suggesting pathways for further investigation into collaborative efforts with heterogeneous robotic units.

Additionally, the interplay of optimization with formation control may inspire further research into diverse fields such as autonomous vehicles and smart infrastructure. By leveraging the core principles detailed in Zhao, Chen, and Ding’s study, new methodologies could be developed that improve cooperative behavior amongst various automated systems, fostering a future where sophisticated robotics become integral to our daily lives and crucial infrastructures.

In conclusion, the contributions of Zhao, Chen, and Ding not only enrich the existing literature on robotic control systems but also provide a practical framework ready for implementation in real-world scenarios. Their finite-time distributed optimization mechanism sets a new paradigm for formation control, ensuring that autonomous robots can work in concert, adapting to real-time challenges. This research signifies a crucial step toward enhancing the collaborative nature of robotics in increasingly dynamic environments.

As we venture further into uncharted territories of robotics and artificial intelligence, studies like this will undoubtedly inspire a new generation of innovations, redefining our expectations and capabilities in countless fields.

Subject of Research: Distributed optimization formation control of networked robots with time-varying reference signals under directed graphs.

Article Title: Finite-time distributed optimization formation control of networked robots with time-varying reference signals under directed graphs.

Article References:

Zhao, W., Chen, Q. & Ding, L. Finite-time distributed optimization formation control of networked robots with time-varying reference signals under directed graphs.
Discov Artif Intell 5, 184 (2025). https://doi.org/10.1007/s44163-025-00415-5

Image Credits: AI Generated

DOI: 10.1007/s44163-025-00415-5

Keywords: Distributed optimization, formation control, networked robots, time-varying reference signals, directed graphs.

The challenge of teaching robots to manipulate flexible materials, such as cables and fabric sheets, is a complex task not easily resolved. Traditional aerial robots are designed primarily to handle rigid objects like boxes, as their mechanics are optimized for stable items. The dynamics of flexible materials present unique challenges, primarily due to the need for real-time adaptation. As Saldaña reflects on his own experiences, he articulates the nuances of this challenge through a simple yet profound analogy: gripping an apple from a tree branch while the branch provides resistance. As humans, we instinctively adjust our grip to counter these external forces. However, current robotic platforms are not yet capable of such nuanced adjustments.

The ambitious research project led by Saldaña aims to change this paradigm. His vision includes the application of reinforcement learning—a method wherein robots learn optimal behaviors through trial and error in dynamic environments. This learning approach will allow aerial robots to develop the skills necessary to manipulate flexible objects without extensive pre-programmed knowledge about the materials they’re interacting with. Saldaña’s novel methodology integrates control systems with reinforcement learning, establishing a framework that enables a type of real-time compensation for unexpected forces, thereby enhancing stability and learning speed simultaneously.

The implications of this research extend far beyond theoretical exploration; they delve into real-world applications that could revolutionize several industries, particularly construction and disaster response. For instance, in the construction sector, aerial robots could deliver and position items like cables or rods currently managed by human workers. This transformation not only promises to decrease operational costs but also significantly enhances worker safety by minimizing their exposure to hazardous environments.

Moreover, the scope of application broadens in the context of emergency scenarios such as natural disasters or catastrophic events. Imagine drones effectively unfurling water hoses in the wake of a fire or wrapping plastic sheeting around structures to prevent hurricane damage. The potential for aerial robots that can respond with dexterity to flexible materials embodies a crucial advancement in intelligent robotics, emphasizing the growing importance of adaptability in autonomous systems.

To begin this groundbreaking project, Saldaña and his team will develop an adaptive controller designed to maintain stability amidst variable external forces. This controller will serve as a foundational component for the reinforcement learning aspect. Through a structured exploration of various control strategies, aerial robots will learn to interact with flexible objects, optimizing their behaviors in a process reminiscent of how living organisms adapt to their surroundings.

Intrigued by the possibilities, Saldaña notes that integrating adaptive control with reinforcement learning has not been done before—a claim that underlines the ambitious essence of his research. By pioneering a new hybrid methodology, Saldaña is not only enhancing the functionality of aerial robots but also challenging existing paradigms within the field of robotics.

One of the prominent areas poised to benefit from these advancements is the construction industry, particularly concerning high-rise buildings. With the risks and costs associated with manual labor on skyscrapers, drones equipped with the capabilities to manipulate flexible materials could fundamentally alter how construction projects are managed. No longer confined to the ground, these robots would operate alongside human workers but from a safe distance, establishing a beneficial synergy between humans and technology.

Before any of these exciting applications can become a reality, there are formidable challenges that must be surmounted. Real-time adaptation in the face of constant external interactions underscores the complexity of the environments aerial robots are likely to operate in. Simulating the innate abilities of a squirrel, which adjusts its movements effortlessly in response to shifting branches, presents an intricate task. However, Saldaña remains optimistic and energized at the prospect of designing solutions that can replicate such behaviors in machines.

Saldaña’s commitment to tackling these challenges reflects a broader vision for the future of robotics—one characterized by resilience and adaptability. By granting aerial robots "squirrel-like" capabilities, it becomes feasible to envision intelligent systems that learn from experience and make decisions autonomously in real-time. This could ultimately lead to completely customizable robotic behavior based on unique situational demands.

In this age of rapid technological advancement, Saldaña’s research exemplifies the intersection of nature-inspired design and cutting-edge engineering. By drawing parallels between the complexities of animal movement and robotic capabilities, he not only redefines the boundaries of what robots can achieve but also plays a pivotal role in transforming our relationship with technology. As hurdles are crossed and innovative methodologies are employed, the future may reveal aerial robots capable of tasks that blend efficiency, safety, and adaptability—signifying a monumental leap toward the realization of intelligent robotics.

Through the support of the NSF CAREER award, Saldaña stands poised to contribute vital knowledge that will shape the next generation of aerial robotics. By addressing the intricacies of flexible manipulation and real-time adaptation, his work embodies a blending of educational mentorship and cutting-edge research, ultimately inspiring future generations of engineers and roboticists to build technologies that harmonize with the natural world.

As technological innovation unfolds, the possibility of drones that seamlessly interact with their environment brings excitement, offering glimpses of advancements that may redefine numerous industries. In solving the problems of aerial manipulation and robotic adaptability, David Saldaña paves the way for transformative changes that echo the intelligence and ingenuity we find in nature—an endeavor that could reshape our understanding of automation and its myriad applications.

Subject of Research: Aerial manipulation and adaptation of flexible materials in robotics
Article Title: Empowering Aerial Robots: Nature’s Inspiration Sparks Innovations in Robotics
News Publication Date: October 2023
Web References: David Saldaña Faculty Page, SwarmsLab Website, NSF Award Abstract (# 2442475)
References: N/A
Image Credits: Credit: Christa Neu/Lehigh University

Keywords

The recently published study in the journal Science Robotics explores the innovative designs behind this soft robot, which utilizes the principles of fluid dynamics to coordinate movements and grasping in a manner akin to that of an octopus. This design approach leverages the octopus’s unique anatomy, showcasing a system that does not rely on traditional computational frameworks, setting a new paradigm in robotic operation and manipulation.

At the heart of the robot’s design is a cutting-edge suction system, which not only allows for adhesion to various surfaces but also serves as a sensory mechanism. It enables the robot to gauge the environmental conditions surrounding it, including the identification of contact with different mediums such as air, water, and varying surface textures. This dual functionality of suction as both an attachment method and a sensory input paves the way for a new understanding of how robots can interact with their environment.

Tianqi Yue, the lead author of the research, articulated the significance of their findings, drawing parallels between their robotic innovations and the natural world. They previously established a concept for an artificial suction cup that emulates the stickiness of an octopus’s suckers. This evolution further develops the concept of ’embodied suction intelligence,’ a term that encapsulates the robot’s capacity to mimic the octopus’s intricate neuromuscular coordination through soft materials combined with fluidic systems.

The functionality exhibited by the soft robot operates on two distinct levels. At a low level, the robot achieves a baseline of intelligence through its fluidic circuitry that combines suction flow with responsive actions. This allows it to handle delicate items with care, adaptively curl around objects, and encapsulate items of indeterminate shapes. At a higher level, by analyzing the pressure changes from the suction mechanism, the robot can discern subtle environmental variations, classify surface roughness, detect contact points, and even predict the forces acting on it during interaction with objects.

This sophisticated level of function presents several practical implications for the future of robotics. The research team envisions applications in various sectors, including agriculture, where soft robots could gently harvest fruits without damaging them. Factories could utilize these advancements for processing fragile components, while medical settings might benefit from robots that can anchor tools inside the human body. Additionally, the potential for creating soft toys and interactive wearables that engage safely with users signifies an exciting frontier for consumer products.

The current research highlights the simplicity and cost-effectiveness of integrating suction intelligence into soft robotic designs. This ability to replicate nature’s solutions not only facilitates the creation of robots that are more intuitive and user-friendly but also emphasizes the potential for new developments that align closely with ecological principles. By harnessing the inherent efficiency present in natural systems, the development team embarks on a mission to simplify the complexity often associated with robotic designs.

In seeking to revolutionize real-world applicability, the research team is actively pursuing advancements to make their system smaller and more robust. By combining their current findings with smart materials and artificial intelligence, they anticipate an increase in adaptability and decision-making prowess in complex, unpredictable environments. The direction of this research signifies a movement toward intelligent soft robots that can navigate diverse tasks with ease.

The innovation of a suction cup, devoid of any electronic components, yet capable of sensory perception, cognitive processing, and actionable responses mirrors the functionalities inherent in octopus arms. Researchers believe this breakthrough opens the door to soft robots that can function more naturally, expanding their utility and interaction within human environments. The implications of such technology permeate various domains, setting the stage for a future enriched by intelligent, responsive soft robotic systems.

This synthesis of biology and engineering not only enriches robotic technology but also invites deeper inquiry into biomimicry as a tool for innovation. The integrated systems derived from the octopus serve as a springboard for enhancing soft robotic capabilities, with the potential to redefine how humans utilize robotics across multiple disciplines. By pioneering this unique approach, the researchers have not only elevated the field of soft robotics but also potentially laid the groundwork for further explorations into biologically inspired robotic systems that can adapt and evolve similarly to living organisms.

As the research unfolds, the scientific community remains eager to observe how these developments might catalyze transformative changes across industries that depend on both automation and delicate handling. Balancing functionality with safety and efficiency, these soft robots could potentially reshape the landscape of human-robot interaction and redefine standards across various applications.

Furthermore, the collaborative spirit behind this research underlines the importance of interdisciplinary efforts in advancing technological innovation. By uniting expertise in robotics, biology, and fluid dynamics, the team has managed to produce a critical advancement with far-reaching implications. Their work will not only contribute to the field of robotics but may also inspire future collaborations, emphasizing the importance of looking to nature for answers to contemporary technological challenges.

The journey toward integrating this soft robotic intelligence into everyday life is just beginning, with the researchers at the University of Bristol leading the charge. As they refine their technologies and explore new applications, the world awaits to discover the true potential of soft robotics that can move, think, and interact with the world just like an octopus does.

Subject of Research: Not applicable
Article Title: Embodying soft robots with octopus-inspired hierarchical suction intelligence
News Publication Date: Not specified
Web References: Not specified
References: Not specified
Image Credits: Tianqi Yue

Keywords

Traditional robotic hands often rely on rigid components controlled by discrete actuators, which limits their adaptability when encountering unexpected forces or complex objects. The inability of these systems to absorb shocks or conform to shapes dynamically has restricted their practical applications, especially in unstructured environments. Junge and Hughes’ research addresses this limitation by integrating compliance—essentially a form of controlled mechanical flexibility—that is spatially distributed throughout the robotic structure. This biomimetic compliance mimics the nonlinear, adaptive properties of biological tissues, enabling the robotic hand to maintain grasp stability while adapting to varying external conditions.

At the heart of their design philosophy lies the concept of distributed compliance, where flexibility and energy dissipation are embedded across the hand’s structure rather than localized to specific joints. This approach contrasts sharply with prior models that often concentrated compliance in single parts, leading to uneven responses and potential mechanical failure points. By distributing compliant elements strategically, the robotic hand absorbs and channels forces more effectively, resulting in smoother interactions with objects and surfaces of differing textures and shapes.

The technological leap is not merely mechanical but also deeply integrated with sophisticated control algorithms. The research team developed novel control schemes that harmonize with the mechanical compliance, ensuring that the robotic hand does not become excessively floppy or unresponsive. Instead, sensors embedded within the compliant materials feed real-time data into adaptive controllers, which modulate actuator responses on the fly. This synergy between hardware and software creates a dynamic feedback loop reminiscent of biological sensorimotor integration, dramatically enhancing manipulation robustness.

The implications of such a system extend well beyond conventional robotics. In prosthetic technology, for example, a biomimetic compliant hand could restore users’ ability to perform intricate, delicate tasks—such as picking up a fragile glass or typing on a keyboard—that have so far remained out of reach for artificial limbs. The compliance not only improves grip adaptability but also significantly reduces the risk of damaging objects or losing grip due to sudden perturbations, mimicking the subtle adjustments present in a biological hand.

From an engineering perspective, the materials selected to replicate this biomimicry play a crucial role. Junge and Hughes utilized advanced elastomers and shape-memory polymers engineered to replicate the nonlinear elastic properties of human tendons and skin. These materials exhibit hysteresis and energy-dissipating behaviors that classical robotics materials typically lack. Integrating these substances into a multi-layered architecture allowed the researchers to design a system with gradated stiffness, which varies smoothly from the fingertips to the palm, closely replicating human hand biomechanics.

Moreover, the layered compliance model was designed with modularity in mind. This modular approach suggests that robotic hands could be customized or scaled depending on specific task requirements—from delicate microsurgery instruments to heavy-duty industrial grippers. The scalability of this approach offers a universal framework for future robotic manipulator designs, bridging a wide spectrum of applications that demand both precision and power.

In their experiments, Junge and Hughes demonstrated the robotic hand’s ability to handle a myriad of objects ranging from soft fruits and fragile ceramics to irregularly shaped tools. Crucially, the robotic system showed remarkable resilience under unexpected disturbances—such as being bumped during manipulation—and swiftly recalibrated its grip to prevent dropping or damaging the object. These results bear witness to the successful real-world translation of spatially distributed biomimetic compliance and underscore the potential for deployment in dynamic, uncontrolled environments.

The research also delved into the sensor integration systems that underpin the robotic hand’s autonomous adaptability. By embedding arrays of tactile and force sensors within the biomimetic materials, the hand gains a rich sensory palette that informs its control system. This broadband sensory input mimics the human hand’s complex sensory network, providing critical information about texture, pressure distribution, and slippage without relying solely on external cameras or optical systems.

Their control algorithms employ advanced machine learning techniques, including reinforcement learning, to refine motor responses through continuous interaction feedback. This learning capacity allows the robotic hand not only to perform preprogrammed tasks but also to improve over time, adapting to new objects and conditions with increasing proficiency. Such adaptive control marks a significant departure from rigid, rule-based automation and ushers in a new era of robots capable of genuine dexterity and situational awareness.

Beyond individual performance, the integration of spatially distributed compliance addresses the longstanding durability challenge posed by repetitive, forceful interactions. By diffusing mechanical stress across multiple compliant interfaces, the robotic hand minimizes wear and tear, potentially extending the operational lifespan of robotic limbs and reducing maintenance costs—a critical factor for applications like space exploration, hazardous material handling, or continuous manufacturing.

In the broader landscape of robotics, this research represents a substantial paradigm shift, underpinning a move towards machines whose physical interaction capabilities closely mirror those of living organisms. The embracement of compliance as a feature rather than a drawback reflects a maturing understanding of soft robotics, biomechanics, and human-robot collaboration. Junge and Hughes’ work places them at the forefront of this transformative wave, providing a blueprint for robotics that integrates form, function, and feedback in unprecedented harmony.

The study also paves the way for more immersive and intuitive human-robot interfaces. By leveraging biomimetic compliance, robotic prostheses or exoskeletons could not only enhance strength or mobility but also provide wearers with nuanced sensory feedback, improving embodiment and user comfort. Such advances could revolutionize rehabilitation technologies and expand the horizons for individuals with motor impairments, blending biological and artificial systems in seamless synergy.

The implications for industrial automation are especially profound in light of the ongoing ambition to deploy robots alongside humans in shared workspaces safely and productively. The compliance embedded in these robotic hands inherently reduces the risk of harmful impacts, enabling safer physical human-robot interaction and collaborative task execution. This feature may accelerate robotics adoption in domains traditionally resistant due to safety concerns, including healthcare, manufacturing, and construction.

Perhaps one of the most fascinating aspects of the research lies in its convergence of multiple scientific disciplines: materials science, control theory, biomechanics, and artificial intelligence. This interdisciplinary approach underscores the complexity of replicating human-like manipulation and highlights the importance of cross-cutting innovation in overcoming the technical challenges involved. The researchers’ success exemplifies how integrated strategies can yield solutions reflecting the elegance and efficiency of natural systems.

In conclusion, Junge and Hughes have introduced a transformative concept and practical demonstration that brings anthropomorphic robotic hands closer to the extraordinary capabilities of their biological counterparts. Their spatially distributed biomimetic compliance framework not only enriches the mechanical and sensory performance of robotic manipulators but also enhances their resilience, adaptability, and longevity. As robotics continues to advance towards more human-centric applications, research efforts like these illuminate the pathway towards machines that can truly grasp the world as deftly and delicately as we do.

Subject of Research: Spatially distributed biomimetic compliance in anthropomorphic robotic manipulation

Article Title: Spatially distributed biomimetic compliance enables robust anthropomorphic robotic manipulation

Article References:
Junge, K., Hughes, J. Spatially distributed biomimetic compliance enables robust anthropomorphic robotic manipulation. Commun Eng 4, 76 (2025). https://doi.org/10.1038/s44172-025-00407-4

Image Credits: AI Generated

The foundational work centers on a collection of disk-shaped autonomous robots, designed to resemble small hockey pucks. These robots have been programmed to assemble into diverse shapes and configurations, effectively behaving like a new type of material with distinct properties. The researchers have ventured into the complex realm of material properties, where the robots demonstrate the ability to be both strong and rigid, yet fluid and adaptable, depending on the task or form required.

A remarkable aspect of this robotic system is how it draws inspiration from living systems, particularly from embryonic tissues. Researcher Otger Campàs, who previously worked at UCSB and is now at TU Dresden, delineates that living tissues possess extraordinary capabilities such as self-healing, self-shaping, and manipulating their material strength in response to different stimuli. This biological insight drives the design in which robots can flexibly switch between solid and fluid states, akin to how cells in an embryo reorganize as they develop.

The collaboration between mechanical engineering and biological studies yields insight into the mechanics behind these processes. Emergent properties of the robots are controlled not only by external forces but also by intricate internal mechanisms. This internal signaling allows the robots to coordinate their movements and adjust their shapes dynamically—an ability that could revolutionize fields like robotics and material science. During the developmental processes of embryos, cells interact through numerous active forces, leading to a profound reorganization from a formless collection into structured forms like limbs and organs. Similarly, by imitating these forces in their robotic systems, researchers can generate transformative capabilities for the robots.

Illustrating the theoretical underpinnings, scientists implemented mechanical components in the robots that facilitate this inter-unit force. This is achieved through eight motorized gears located on the circumference of each robot, enabling them to navigate around one another and constructively push each other even in confined spaces. Such mechanical manipulation mirrors the natural interactions observed in living organisms, where cells work harmoniously to reshape the material they compose.

Moreover, the researchers have utilized light sensors equipped with polarized filters that serve as a method for the robots to perceive their environment. When exposed to targeted light, these sensors direct the robots to rotate their gears in specific ways, altering their formation effortlessly. This feature allows for synchronized movements within the collective, as robots can respond collectively to changes in light, much like how cells respond to biochemical signals in embryonic development.

A particularly fascinating area of study within this research involves the concept of adhesion among the robots. This adhesion is achieved through strategically integrated magnets on the edges of the robotic units. Activation or deactivation of these magnets allows robots to attract or repel each other as necessary, cultivating a more cohesive or fluid collective when interacting with various environments or tasks.

Research findings demonstrate that fluctuations in signal strength are pivotal to enabling shape-shifting capabilities in this robotic ensemble. By simulating the natural unpredictability found in cell interactions, researchers noted improved fluidity and adaptability in the robotic collective. A key takeaway is that robots mimic biological systems in converting solid states to fluid states based on these fluctuations, enhancing their ability to respond to environmental changes.

By modulating both signal fluctuations and inter-unit forces, the researchers successfully transformed the collective from a rigid configuration to a dynamic, flowing entity. This approach not only reduces overall power consumption but unlocks potential for advanced robotic applications. Dynamic switching between these states means that robots can perform tasks efficiently by conserving energy when rigid and activating movement when flow is required.

These developments suggest broader applications beyond mere robotics—they pave the way for further investigations into the principles of active matter and phase transitions. The findings could offer fresh insights into biological research, opening pathways to understanding how living systems function at a material level. This powerful interplay between robotics and biology could potentially lead to unprecedented advances in intelligent materials and adaptive systems.

As scientific exploration continues, the potential of scaling down these robotic units leads towards the creation of larger, more versatile assemblies that can function much like traditional materials but with inherent active capabilities. Researchers are optimistic about leveraging machine learning strategies to unlock new behaviors and enhance the usability of such robotic systems. The work done by Devlin, Hawkes, and their team signifies an exciting step forward in redefining what materials can do, encouraging a wave of innovation that merges material science with robotics.

With this foundational knowledge and continuing exploration, the researchers envision expansive applications for their material-like robotic collectives. By integrating sophisticated controls and self-regulatory mechanisms, they aim to develop robotic systems that can evolve beyond their initial programming. The possibilities are endless as they explore how these robots can reshape themselves, handle substantial loads, or even demonstrate self-healing abilities, much like biological materials.

As the world increasingly turns towards automation and intelligent systems, this groundbreaking research embodies a critical intersection between technology and biology, epitomizing the future of materials science. The implications stretch beyond academic curiosity, as we contemplate a new era of smart materials that can adapt to the environment, self-arrange, and ultimately enhance human capabilities in multifaceted and unforeseen ways.

Subject of Research: Material-like robotic collectives with spatiotemporal control of strength and shape
Article Title: Material-like robotic collectives with spatiotemporal control of strength and shape
News Publication Date: 21-Feb-2025
Web References: Science
References: 10.1126/science.ads7942
Image Credits: N/A

Keywords

• Applied sciences and engineering
• Robotics
• Mechatronics
• Robot control
• Robotic designs
• Swarm robotics
• Robotic locomotion