Robotic Flying Fish: Bridging the Gap Between Water and Sky

The Robotic Flying Fish is a bio-inspired machine designed to emulate the remarkable locomotion of its biological counterpart, the flying fish. This creation stands as a testament to the power of biomimicry in robotics, where the principles of nature are harnessed to engineer robots capable of transcending the boundaries between water and sky. One of the most captivating features of the Robotic Flying Fish is its propulsion system. Powerful oscillations, akin to the movements of a real fish, are used to achieve high-speed swimming. The incorporation of these oscillations enables the robot to navigate the underwater domain with exceptional efficiency and agility. This design choice emphasizes the importance of imitating nature’s fluid dynamics to excel in aquatic environments. The swimming motion testing phase of this project is of paramount importance, as it serves as the foundation for the robot’s aquatic performance. The experiments conducted within a water tank, measuring 5 meters in length, 4 meters in width, and 1.5 meters in depth, have unveiled the impressive capabilities of the Robotic Flying Fish.

In a new study published in the peer-reviewed journal Bioinspiration & Biomimetics, scientists from Peking University, in collaboration with Taiyuan University of Technology, have reported an exciting development in the field of aquatic-aerial robotics. Led by Dr. Di Chen, Dr. Zhengxing Wu, Dr. Yan Meng, and Professor Junzhi Yu, along with Dr. Huijie Dong, their work focused on designing a robotic flying fish with the capability to seamlessly transition between water and air, a feat inspired by the remarkable locomotion of natural flying fish. To achieve high-speed swimming, the research team designed a propulsion system that mimics the undulating motion of a fish’s body. This design choice, combined with bilateral symmetry and slight positive buoyancy, imparts remarkable stability during high-speed swimming. Notably, the robot maintains partial buoyancy, allowing it to rely primarily on self-propulsion rather than buoyancy for acceleration. This innovation aligns with the overarching goal of achieving dynamic and agile underwater locomotion.

The results of the swimming motion tests are truly remarkable. As the actuation frequency increases, the swimming speed of the robot proportionally rises. The peak swimming speed reached an impressive 1.43 meters per second, equivalent to 5.42 body lengths per second. This achievement underscores the significance of designing a propulsion system that faithfully emulates nature’s mechanisms, as it is through this mimicry that such high-performance swimming is made attainable. Comparatively, it is important to note that the swimming speed, although impressive, has decreased slightly when compared to previous work due to the implementation of morphing pectoral fins. These fins were integrated to address the challenges posed by cross-domain locomotion. While this modification impacted the swimming speed, it also opened doors to exciting possibilities in other areas.

The pinnacle of this project’s success lies in its ability to perform cross-domain motion seamlessly. The transition from water to air, a feat rarely achieved in the realm of robotics, is the true litmus test for an aquatic-aerial robot like the Robotic Flying Fish. To validate the cross-domain capability of this robotic marvel, the authors conducted experiments with a specific focus on the intricate phases involved in transitioning from the underwater realm to the aerial domain. A robotic arm, equipped with two degrees of freedom, was ingeniously employed to assist in this challenging endeavor. The gripping mechanism of the arm, coupled with precise adjustments in pitch angle, enabled controlled and strategic movements. The experiment involved several phases, including underwater acceleration, water-air interface crossing, and the deployment of wing-like pectoral fins. The entire process was meticulously documented using high-speed cameras. What transpired was a mesmerizing display of technology and nature seamlessly merging into one.

The Robotic Flying Fish achieved a vertical leaping motion and wing spreading, precisely imitating the biological flying fish’s behavior. The robot’s ability to accomplish this cross-domain feat in a controlled and efficient manner heralds a significant advancement in aquatic-aerial robotics. It is the first of its kind to genuinely possess the capability of ‘fish leaping and wing spreading’ cross-domain locomotion. While the primary focus of this robotic creation is its aquatic prowess, the engineers behind the Robotic Flying Fish were also curious about its gliding capabilities. In nature, flying fish utilize a combination of swimming and gliding to cover significant distances above the water’s surface. Replicating this behavior in a robotic system poses considerable challenges.

During the gliding motion, the robot’s pitch angle increased gradually due to the wing’s pitch moment. This unique design choice aimed to mimic the characteristics of real flying fish. However, due to certain design limitations, such as the absence of an extra pair of pelvic fins for better longitudinal stability, the gliding distance was relatively short. It is essential to recognize that this exploration of gliding motion represents the first step in understanding the robot’s potential beyond aquatic environments.

Recognizing the need for further improvement in gliding performance, the engineers turned to simulation analysis. The goal was to optimize the gliding distance by dynamically adjusting the sweepback angle of the pectoral fins, a critical aspect of the robot’s design. The simulation conducted by the authors involved the use of a double Deep Q-Network (DQN) control strategy, with a reward function designed to reflect improvements in gliding distance. The training of the network was rigorous, comprising 30,000 episodes. The results of this simulation were promising. It was demonstrated that the gliding distance could be improved by dynamically adjusting the sweepback angle. The optimization control strategy effectively enhanced the gliding performance, offering a 7.2% increase in the maximum gliding distance. Their findings are of significant importance as they highlight the potential for further enhancements in gliding capabilities through meticulous control strategies and design optimizations.

The journey of creating the Robotic Flying Fish by Professor Junzhi Yu and colleagues is a testament to the fusion of engineering ingenuity and nature’s designs. It has yielded a robot that showcases remarkable aquatic speed, precise cross-domain motion, and initial gliding capabilities. However, this journey is far from over. One cannot overlook the considerable challenges that come with imitating the complex locomotion of flying fish. While the robotic flying fish achieved a notable swimming speed of 1.43 meters per second and executed cross-domain locomotion with finesse, the gliding performance still lags behind its biological counterpart. In conclusion, the Robotic Flying Fish is a remarkable engineering achievement that exemplifies the power of biomimicry in robotics. It serves as a bridge between the aquatic and aerial worlds, offering a glimpse into the future of autonomous cross-domain locomotion.