Multiple Robotic Fish Research Articles

A Novel Obstacle Traversal Method for Multiple Robotic Fish Based on Cross-Modal Variational Autoencoders and Imitation Learning.

Precision control of multiple robotic fish visual navigation in complex underwater environments has long been a challenging issue in the field of underwater robotics. To address this problem, this paper proposes a multi-robot fish obstacle traversal technique based on the combination of cross-modal variational autoencoder (CM-VAE) and imitation learning. Firstly, the overall framework of the robotic fish control system is introduced, where the first-person view of the robotic fish is encoded into a low-dimensional latent space using CM-VAE, and then different latent features in the space are mapped to the velocity commands of the robotic fish through imitation learning. Finally, to validate the effectiveness of the proposed method, experiments are conducted on linear, S-shaped, and circular gate frame trajectories with both single and multiple robotic fish. Analysis reveals that the visual navigation method proposed in this paper can stably traverse various types of gate frame trajectories. Compared to end-to-end learning and purely unsupervised image reconstruction, the proposed control strategy demonstrates superior performance, offering a new solution for the intelligent navigation of robotic fish in complex environments.

A Circular Formation Method for Biomimetic Robotic Fish Inspired by Fish Milling.

Circular motion phenomena, akin to fish milling, are prevalent within the animal kingdom. This paper delineates two fundamental mechanisms underlying such occurrences: forward following and circular topological communication. Leveraging these pivotal concepts, we present a multi-agent formation circular model based on a second-order integrator. This model engenders the attainment of homogeneous intelligence convergence along the circumferential trajectory. The convergence characteristics are intricately linked to the number of agents and the model parameters. Consequently, we propose positive and negative solutions for ascertaining the convergent circle property and model parameters. Furthermore, by integrating our proposed formation control methodology with a robotic fish dynamics model, we have successfully implemented simulations and experiments, demonstrating the circular formation of multiple biomimetic robotic fish. This study provides a mathematical explication for the circular motion observed in animal groups and introduces a novel approach to achieving circular formation in multiple robots inspired by biological phenomena.

Towards end-to-end formation control for robotic fish via deep reinforcement learning with non-expert imitation

Collaboration of multiple robotic fish can accomplish various underwater tasks effectively. However, controlling robotic fish to maintain a specific formation remains a huge challenge, especially in complex and changing flow fields. This paper presents an end-to-end formation control approach in the leader–follower topology by combining deep reinforcement learning and imitation learning. First, we build a high-fidelity environment based on computational fluid dynamics (CFD) to generate samples for training the formation controller. In this environment, we maneuver the robotic fish by adjusting the maximum swing of its tail. Then, we model the formation control problem as a Markov decision process (MDP), where a compound reward function is tailored to guide the training. To improve the learning efficiency of the deep reinforcement learning (DRL) based controller, we propose a novel DRL algorithm on the top of deep Q-networks (DQN) and behavior cloning, which we call dueling double DQN (D3QN) with imitation. Combining with the designed imitation-based action selection strategy, this algorithm significantly reduce the blindness of agent exploration at the beginning of training. A series of experiments demonstrate the advantages of the proposed algorithm in terms of control accuracy, training efficiency, as well as generalization ability for different formation configurations.

Bottom-level motion control for robotic fish to swim in groups: modeling and experiments

Moving in groups is an amazing spectacle of collective behaviour in fish and has attracted considerable interest from many fields, including biology, physics and engineering. Although robotic fish have been well studied, including algorithms to simulate group swimming, experiments that demonstrate multiple robotic fish as a stable group are yet to be achieved. One of the challenges is the lack of a robust bottom-level motion control system for robotic fish platforms. Here we seek to overcome this challenge by focusing on the design and implementation of a motion controller for robotic fish that allows multiple individuals to swim in groups. As direction control is essential in motion control, we first propose a high-accuracy controller which can control a sub-carangiform robotic fish from one arbitrary position/pose (position and direction) to another. We then develop a hydrodynamic-model-based simulation platform to expedite the process of the parameter tuning of the controller. The accuracy of the simulation platform was assessed by comparing the results from experiments on a robotic fish using speeding and turning tests. Subsequently, extensive simulations and experiments with robotic fish were used to verify the accuracy and robustness of the bottom-level motion control. Finally, we demonstrate the efficacy of our controller by implementing group swimming using three robotic fish swimming freely in prescribed trajectories. Although the fluid environment can be complex during group swimming, our bottom-level motion control remained nominally accurate and robust. This motion control strategy lays a solid foundation for further studies of group swimming with multiple robotic fish.

Coordination of Multiple Robotic Fish With Applications to Underwater Robot Competition

This paper is concerned with the coordination control of multiple biomimetic robotic fish in highly dynamic aquatic environments by building a hybrid centralized system. With the aid of the results of biorobotics and control techniques, a radio-controlled multijoint robotic fish and its locomotion control are developed. To enable a closed control loop, a visual subsystem that is responsible for tracking of multiple moving objects is constructed and implemented in real time. Furthermore, a behavior-based hierarchical architecture in conjunction with fuzzy reinforcement learning is proposed to accomplish effective coordination among multiple swimming robots. Finally, experiments on 2vs2 water polo game are carried out to verify the proposed coordination control scheme. Over the past eight years, this multirobot platform has been successfully applied to international underwater robot competitions to promote innovative research and education in underwater robotics.

Leader–Follower Flocking of Multiple Robotic Fish

In this paper, we study cohesive flocking and formation flocking of multiple robotic fish swimming in the water surface under the guidance of only one leader with zero-value external input. Combining consensus algorithms and attraction/repulsion functions, a distributed flocking algorithm is proposed for the robotic fish system to execute the cohesive flocking task. According to the LaSalle–Krasovskii invariance principle, the proposed algorithm enables followers to asymptotically track the leader's velocity and approach the equilibrium distances with their neighbors, provided that the initial interaction network of the system is a leader–follower connected graph. Furthermore, by adding the information of a desired formation topology to the potential function term, the proposed algorithm can be extended to solve arbitrarily shaped formation flocking problem of multiple robotic fish. The experimental results demonstrate that the proposed approaches are effective for three robotic fish. Finally, several simulation examples are given to verify the functionality of the proposed approaches for a larger system with ten agents.

Influence of robotic shoal size, configuration, and activity on zebrafish behavior in a free-swimming environment

In animal studies, robots have been recently used as a valid tool for testing a wide spectrum of hypotheses. These robots often exploit visual or auditory cues to modulate animal behavior. The propensity of zebrafish, a model organism in biological studies, toward fish with similar color patterns and shape has been leveraged to design biologically inspired robots that successfully attract zebrafish in preference tests. With an aim of extending the application of such robots to field studies, here, we investigate the response of zebrafish to multiple robotic fish swimming at different speeds and in varying arrangements. A soft real-time multi-target tracking and control system remotely steers the robots in circular trajectories during the experimental trials. Our findings indicate a complex behavioral response of zebrafish to biologically inspired robots. More robots produce a significant change in salient measures of stress, with a fast robot swimming alone causing more freezing and erratic activity than two robots swimming slowly together. In addition, fish spend more time in the proximity of a robot when they swim far apart than when the robots swim close to each other. Increase in the number of robots also significantly alters the degree of alignment of fish motion with a robot. Results from this study are expected to advance our understanding of robot perception by live animals and aid in hypothesis-driven studies in unconstrained free-swimming environments.

Distributed adaptive flocking of robotic fish system with a leader of bounded unknown input

A distributed leader-follower flocking problem of multiple robotic fish governed by extended second-order unicycles is studied in this paper. The multi-agent system consists of only one leader with pre-appointed and bounded speeds. A distributed flocking algorithm on the basis of the combination of consensus and attractive/repulsive functions is investigated, in which adaptive strategy is adopted to compute the weight of the velocity coupling strengths. The proposed control algorithm enables followers to asymptotically track the leader’s varying velocities and approach the equilibrium distances with their neighbors. Furthermore, the arbitrarily-shaped formation flocking problem of the system can also be solved by adding the information of a desired formation topology to the potential function term. Finally, simulations are carried out to verify the effectiveness of the proposed theoretical results.

Experimental implementation of distributed flocking algorithm for multiple robotic fish

This paper presents the experimental implementation of a flocking algorithm for multiple robotic fish governed by extended second-order unicycles. Combing consensus protocols with attraction/repulsion functions, a flocking algorithm is proposed to make the agents asymptotically converge to swim with consistent velocities and approach the equilibrium distances to their neighbors. The LaSalle–Krasovskii invariance principle is applied to verify the stability of the system. Besides numerical simulations, platform simulations involving robotic fish kinematic constraint and control mechanism are shown. An experiment with three robotic fish is implemented to illustrate the effectiveness of the proposed flocking algorithm in the presence of external disturbance and boundary collision.

Decentralized Leader-Follower Flocking of Multiple Non-Holonomic Agents

In this paper, we study the cohesive flocking problem of multiple non-holonomic agents governed by extended second-order unicycle dynamics, such as robotic fish. Consider the system with only one leader. Combining the consensus algorithm and attraction/repulsion functions, a distributed flocking algorithm is presented for multiple robotic fish to complete the cohesive flocking task. According to LaSalle-Krasovskii invariance principle, the proposed algorithm enables followers to asymptotically track the leader's velocity and approach the equilibrium distances with their neighbors, provided that the initial interaction network among the followers is an undirected connected graph, and there exists at least one follower having a leader neighbor at the initial time. During the evolutionary process, the connectivity of the communication network can be preserved due to the effect of the potential function. Finally, a simulation example is given to verify the proposed theoretical analysis.

Cooperative box-pushing with multiple autonomous robotic fish in underwater environment

This study presents a cooperative underwater box-pushing scenario, in which three autonomous robotic fish that sense, plan and act on its own move an elongated box from some initial location to a goal location. With the onboard monocular camera, the robotic fish can estimate the pose of the object in the swimming tank. Considering the complexity of the underwater environment and the limited capability of a single robotic fish, the authors address the task by decomposing it into three subtasks and assigning them to capable robotic fish. With one robotic fish observing the box at the goal location and two robotic fish pushing the left and right ends of the box, the box can be moved gradually towards the goal location. The subtask consists of a series of behaviours, each designed to fulfil one step of the subtask. The robotic fish coordinate through explicit communications and distribute the subtasks with a market-based dynamic task allocation method. Task reallocation mechanism that permits robotic fish to auction its assigned task to capable ones is used to cope with unexpected changes in the environment and the limited sensing range of the robotic fish. Experiments are conducted to verify the feasibility of the proposed methods.

Coordinated Transport by Multiple Biomimetic Robotic Fish in Underwater Environment

This paper presents a coordination method for multiple biomimetic robotic fish in underwater transport task. Based on our robotic fish prototype, we further investigate the coordination problems of multiple robotic fish. Considering the kinematic constraints of the robotic fish, and the unstructured and dynamic characteristics of the water environment, we utilize the limit cycle approach to control the posture of the fish and realize collision avoidance. To control the transport orientation in the particular underwater environment, the fuzzy logic method is adopted. Experimental results of transportation performed by three robotic fish validate the effectiveness of the proposed method.

Development of multiple robotic fish cooperation platform

This article presents the development of a multiple robotic fish cooperation platform, which is established by employing a group of radio-controlled, multi-link fish-like robots. This work is inspired by the observation from nature that the capability of one single fish is limited, as in order to survive the atrocious circumstances in the sea, fish often swim in schools. The analogical situations occur in the robotic fish case. In engineering applications, most missions are so complex that they must be accomplished by effective cooperation of multiple fish robots. The platform presented in this article, as a novel test bed for multiple robotic fish cooperation, can be applied to different types of complex tasks. More importantly, it provides a good platform to test and verify all kinds of algorithms and strategies for cooperation of multiple underwater mobile robots. We use two cooperative tasks as examples to heuristically demonstrate the performance of this platform.

Control and Coordination of Multiple Biomimetic Robotic Fish

Investigation in biomimetic robotic fish is a multidisciplinary area entirely integrating hydrodynamics-based control and robotic technology. In this brief, grounded on an optimized kinematic and dynamic model of robotic fish which synthesizes both the carangiform and anguilliform swimming modes, a free-swimming, multilink robotic fish and its motion control are designed. Employing top-down design approach, we propose a hierarchical architecture for an artificial multifish system which consists of five levels: task level, role level, behavior level, action level, and controller level, to formalize the processes from task decomposition, role assignments, and control performance. A vision-based, closed-loop experimental system for multiple robotic fish is finally constructed to verify the proposed architecture. The running experiments of fish versus man primarily show that much greater capabilities are exhibited when the fish try to cooperate with each other