Understanding multi-fin swimming and maneuvering to develop highly capable swimming robots

Fish swim underwater with levels of agility and maneuverability that far exceed those of contemporary unmanned underwater vehicles (UUVs). While UUVs primarily rely on rectilinear motions, fish continuously move their fins and body--pitching, rolling, turning, and dynamically positioning and orienting themselves within the flow. The forces fish use to swim and maneuver are the result of highly coordinated time-varying forces produced by the motions of their body and fins. The modulation of the forces produced by fins and body occurs in multiple ways. Fish can manipulate the shape and local stiffness of their fins to alter the direction of forces. Among different fish species there is a high degree of variability in the relative location of the fins along the body. In addition, the kinematics and utilization of the different fins will vary depending on the gait or behavioral goal of the fish. In order to understand what contributes to this performance disparity and design swimming robots that are able to swim and turn effectively, it is necessary to understand which physical properties and biomechanical capabilities of fish's fins and body are crucial to their swimming and maneuvering capacity. The objectives of this research are to: 1. Understand how the coordination of fish fins and body produce and shape the propulsive forces necessary for effective forward swimming and turning. 2. Adapt and implement that understanding for the development of multi-fin swimming robots that effectively swim and turn. An iterative five-step process was adopted to accomplish these objectives. Firstly, previous biological and biorobotic research were used to identify characteristics of a fish-like body and fins that are suspected to substantially contribute to swimming performance. These characteristics were then modeled using purpose-built biorobotic platform(s), which were used to experimentally investigate how varying the said characteristics affect force production. The findings from these experiments were analyzed to assess the influence of individual characteristics on net force production and body motion. Insights from these analyses, as well as from previous biological and biorobotic research, were then synthesized to define a list of functionalities and design considerations that are crucial for the development of a multi-fin robot capable of effective swimming and turning. Throughout the course of this research, three experimental platforms were designed and built, each facilitating investigations to understand how different characteristics affect swimming and turning. The first platform is a free-swimming, paired-pectoral fin (PPF) robot, used to investigate forces produced by coordinated pectoral fins and serve as a testbed for waterproofing techniques and mechanism design (Chapter 2). The second is a robotic platform with a jointed peduncle along with dorsal, anal, and caudal fins (PDAC), to investigate how thrust and lateral forces are affected by changes in the phase and geometric relationships among the peduncle and fins. This platform also served as a testbed to further refine manufacturing techniques and develop mechatronic systems (Chapter 3). The third is a posable fish-shaped body and fins to investigate the forces and moments that are produced by body and fins in different body configurations associated with turning motions (Chapter 4). The results from research using these platforms led to the development of a highly-maneuverable, multi-finned, free-swimming robotic experimental platform, called SAMUNO. The detailed design and performance evaluation of the SAMUNO robot is discussed in Chapter 5. Experimentation conducted with both the PPF and PDAC robots has shown that changes in the relative phase of the fins significantly alter the magnitude and temporal profile of the net force vector. Though coordinated paired pectoral fin motions produced net forces that were remarkably similar to the linear sum of individual fin forces, the same was not found for net forces produced by coordinated median fin motions. The observed changes in propulsive forces while varying the phase and geometric relationships between the median fins are different from those that would be expected if net forces were the result of the linear superposition of individual fin forces. Experiments have shown that propulsive forces were highly dependent on the phase relationships among the median fins and peduncle and the spacing between the location of the fins along the body. The effect of phase is considerable; appropriate phase relationships between the dorsal/anal fins and the caudal fin can more than double mean thrust or reduce lateral forces to nearly zero. Mean thrust and root mean square (RMS) lateral forces varied cyclically as the phase of the dorsal and anal fin was varied relative to the phase of the caudal fin and peduncle. Furthermore, changes in the relative location of the fins affect the phase at which the changes in forces occur, with the thrust forces being more sensitive to changes in fin location relative to lateral forces. In addition to the changes in mean forces, the time-varying magnitude of the 2D forces produced by the fins varies significantly with relative fin phasing as well. Observations of the fluid flows during experiments and 2D simulated flows under similar conditions suggest that these changes in forces may be attributed, in large part, to the interaction of the downstream fins and body with the wake produced by the upstream fin(s). The fin phase and geometric relationships that produce large mean thrust correspond to flow conditions where the wake produced by upstream fin(s) smoothly transitions to, and is accelerated by, the downstream fin. When maximum mean thrust is produced, the downstream fin's motion is such that its leading edge maintains a preferential angle-of-attack relative to the wake produced by the upstream fin(s). This condition supports smooth flow over the surface of the downstream fin, enhancing the energy in that fin's wake in the direction of thrust. Body bending was found to enhance force production during steady swimming and play a critical role during the execution of turning maneuvers. The actuation of a single peduncle joint using the PDAC robot significantly increased the overall magnitude of thrust and lateral forces produced during a steady swimming gait without disrupting the fin-fin interactions that enhance thrust production with proper fin phasing. This suggests that additional body bending may further increase swimming performance. Experiments with the posable fish platform found that drag-based force from body bending may play a larger role in changing body orientation during steady swimming whereas zero-angular-momentum rotations are more useful when stationary or at low speeds. Analysis of the findings from investigations with each biorobotic system were synthesized into a list of functionalities and capabilities that are necessary for agile, multi-fin swimming. Each functionality was evaluated for its relative contribution to overall swimming and turn performance. This list of functionalities was then adapted into the design requirements for an agile, multi-finned swimming robotic experimental platform (Chapter 5). The built platform, dubbed SAMUNO, is a free-swimming, wirelessly controlled robot that includes a two-jointed body, dorsal, anal, caudal and a pair of pectoral fins. The SAMUNO robot's swimming and turning abilities were validated through comparison with other contemporary multi-finned swimming robots in the literature. Its swimming and maneuvering performance is comparable to, and in some cases exceeds, that of similar platforms. SAMUNO maintains steady swimming speeds exceeding 1.29 m/s (2.03 body lengths per second, BL/s) and is capable of executing rapid turns up to 90° (1.57 radians) in less than one second from a complete standstill. These achievements demonstrate that the set of crucial functionalities identified in this research are critical for agile, effective underwater multi-fin locomotion. More broadly, SAMUNO serves not only as a validation of key principles but also as a foundation for the next generation of agile underwater robots. The SAMUNO robot embodies a powerful experimental platform that could deepen our understanding of fin-body coordination, pioneer new swimming gaits, and advance control strategies to eventually enable autonomous robots to maneuver through complex, dynamic aquatic environments with fish-like agility.