Swarm Of Agents Research Articles

Achieving task allocation in swarm intelligence with bi-objective embodied evolution

In this paper, we seek to achieve task allocation in swarm intelligence using an embodied evolutionary framework, which aims to generate divergent and specialized behaviors among a swarm of agents in an online and self-organized manner. In our considered scenario, specialization is encouraged through a bi-objective composite fitness function for the genomes, which is the weighted sum of a local and a global fitness function. The former depends only on the behavior of an agent itself, while the latter depends on the effectiveness of cooperation among all nearby agents. We have tested two existing variants of embodied evolution on this scenario and compared their performances against those of an individual random walk baseline algorithm. We have found out that those two embodied evolutionary algorithms have good performances at the extreme cases of weight configurations, but are not adequate when the two objective functions interact. We thus propose a novel bi-objective embodied evolutionary algorithm, which handles the aforementioned scenario by controlling the proportion of specialized behaviors via a dynamic reproductive isolation mechanism. Its performances are compared against those of other considered algorithms, as well as the theoretical Pareto frontier produced by NSGA-II.

A modified Marine Predator Algorithm based on opposition based learning for tracking the global MPP of shaded PV system

Under partial shading condition, the power-voltage curve of the photovoltaic (PV) system contains several maximum power points (MPPs). Among these points, there is only single global and some local points. Accordingly, modern optimization algorithms are highly required to tackle this problem. However, the methods are considered as time consuming. Therefore, finding a new algorithm that capable to solve the problem of tracking global maximum power point (GMPP) with minimum number of population is highly appreciated. Several new straightforward methods as well as meta-heuristic approaches are exist. Recently, the Marine Predator Algorithm (MPA) has been developed for engineering applications. In this study, an alternative method of MPA, integrating Opposition Based Learning (OBL) strategy with Grey Wolf Optimizer (GWO), named MPAOBL-GWO, is proposed to cope with the implied weaknesses of classical MPA. Firstly, Opposition Based Learning (OBL) strategy is adopted to prevent MPA method from searching deflation and to obtain faster convergence rate. Besides, the GWO is also implemented to further improve the swarm agents’ local search efficiency. Due to that, the MPA explores the search space well better than exploiting it; so, this combination improves the efficiency of the MPA and avoids it from falling in local points. To verify the effectiveness of the enhanced method, the well-known CEC’17 test suite and the maximum power point tracking (MPPT) of photovoltaic (PV) system problem are solved. The obtained results illustrate the ability of the proposed MPAOBL-GWO based method to achieve the optimum solution compared with the original MPA, GWO and Particle Swarm Optimization (PSO). The findings revealed that, the proposed method can be viewed as an efficient and effective strategy for more complex optimization scenarios and the MPPT as well.

Multiagent Planning and Control for Swarm Herding in 2-D Obstacle Environments Under Bounded Inputs

This article presents a method for herding a swarm of adversarial agents toward a safe area in a 2-D obstacle environment. The team of defending agents (defenders) aims to block the path of a swarm of risk-averse, adversarial agents (attackers) and guide it to a safe area while navigating in an obstacle-populated environment. To achieve this, a closed formation (StringNet) of defenders is formed around the adversarial swarm. A combination of open-loop, near time-optimal controllers (that result in forming the defenders’ formation), and state-feedback controllers with finite-time convergence guarantees under bounded inputs (that guide the formation around attackers and toward the safe area) synthesize the herding strategy. For demonstration purpose, we consider that the attacking swarm moves under a flocking model, which however is unknown to the defenders. Collision-free trajectory generation for the defenders, as well as their convergence to the desired formations, is proved formally, and simulations are provided to demonstrate the efficacy of the proposed approach. An implementation of the proposed approach on quadrotor vehicles simulated in the Gazebo simulator is also provided.

Aerial Swarm Defense by StringNet Herding: Theory and Experiments.

This paper studies a defense approach against one or more swarms of adversarial agents. In our earlier work, we employed a closed formation (“StringNet”) of defending agents (defenders) around a swarm of adversarial agents (attackers) to confine their motion within given bounds, and guide them to a safe area. The adversarial agents were assumed to remain close enough to each other, i.e., within a prescribed connectivity region. To handle situations when the attackers no longer stay within such a connectivity region, but rather split into smaller swarms (clusters) to maximize the chance or impact of attack, this paper proposes an approach to learn the attacking sub-swarms and reassign defenders toward the attackers. We use a “Density-based Spatial Clustering of Application with Noise (DBSCAN)” algorithm to identify the spatially distributed swarms of the attackers. Then, the defenders are assigned to each identified swarm of attackers by solving a constrained generalized assignment problem. We also provide conditions under which defenders can successfully herd all the attackers. The efficacy of the approach is demonstrated via computer simulations, as well as hardware experiments with a fleet of quadrotors.

Control of a Robotic Swarm Formation to Track a Dynamic Target with Communication Constraints: Analysis and Simulation

We describe and analyze the Local Charged Particle Swarm Optimization (LCPSO) algorithm, that we designed to solve the problem of tracking a moving target releasing scalar information in a constrained environment using a swarm of agents. This method is inspired by flocking algorithms and the Particle Swarm Optimization (PSO) algorithm for function optimization. Four parameters drive LCPSO—the number of agents; the inertia weight; the attraction/repulsion weight; and the inter-agent distance. Using APF (Artificial Potential Field), we provide a mathematical analysis of the LCPSO algorithm under some simplifying assumptions. First, the swarm will aggregate and attain a stable formation, whatever the initial conditions. Second, the swarm moves thanks to an attractor in the swarm, which serves as a guide for the other agents to head for the target. By focusing on a simple application of target tracking with communication constraints, we then remove those assumptions one by one. We show the algorithm is resilient to constraints on the communication range and the behavior of the target. Results on simulation confirm our theoretical analysis. This provides useful guidelines to understand and control the LCPSO algorithm as a function of swarm characteristics as well as the nature of the target.

Energy-Efficient Navigation of an Autonomous Swarm with Adaptive Consciousness

The focus of this work is to analyze the behavior of an autonomous swarm, in which only the leader or a dedicated set of agents can take intelligent decisions with other agents just reacting to the information that is received by those dedicated agents, when the swarm comes across stationary or dynamic obstacles. An energy-aware information management algorithm is proposed to avoid over-sensation in order to optimize the sensing energy based on the amount of information obtained from the environment. The information that is needed from each agent is determined by the swarm’s self-awareness in the space domain, i.e., its self-localization characteristics. A swarm of drones as a multi-agent system is considered to be a distributed wireless sensor network that is able to share information inside the swarm and make decisions accordingly. The proposed algorithm reduces the power that is consumed by individual agents due to the use of ranging sensors for observing the environment for safe navigation. This is because only the leader or a dedicated set of agents will turn on their sensors and observe the environment, whereas other agents in the swarm will only be listening to their leader’s translated coordinates and the whereabouts of any detected obstacles w.r.t. the leader. Instead of systematically turning on the sensors to avoid potential collisions with moving obstacles, the follower agents themselves decide on when to turn on their sensors, resulting in further reduction of overall power consumption of the whole swarm. The simulation results show that the swarm maintains the desired formation and efficiently avoids collisions with encountered obstacles, based on the cross-referencing feedback between the swarm agents.

The impact of catastrophic collisions and collision avoidance on a swarming behavior

Swarms of autonomous agents are useful in many applications due to their ability to accomplish tasks in a decentralized manner, making them more robust to failures. Due to the difficulty in running experiments with large numbers of hardware agents, researchers often make simplifying assumptions and remove constraints that might be present in a real swarm deployment. While simplifying away some constraints is tolerable, we feel that two in particular have been overlooked: one, that agents in a swarm take up physical space, and two, that agents might be damaged in collisions. Many existing works assume agents have negligible size or pass through each other with no added penalty. It seems possible to ignore these constraints using collision avoidance, but we show using an illustrative example that this is easier said than done. In particular, we show that collision avoidance can interfere with the intended swarming behavior and significant parameter tuning is necessary to ensure the behavior emerges as best as possible while collisions are avoided. We compare four different collision avoidance algorithms, two of which we consider to be the best decentralized collision avoidance algorithms available. Despite putting significant effort into tuning each algorithm to perform at its best, we believe our results show that further research is necessary to develop swarming behaviors that can achieve their goal while avoiding collisions with agents of non-negligible volume.

Hybrid slime mould algorithm with adaptive guided differential evolution algorithm for combinatorial and global optimization problems

The Slime Mould Algorithm (SMA) is a recent metaheuristic inspired by the oscillation of slime mould. Similar to other original metaheuristic algorithms (MAs), SMA may suffer from drawbacks, such as being trapped in minimum local regions and improper balance between exploitation and exploration phases. To overcome these weaknesses, this paper proposes a hybrid algorithm: SMA combined to Adaptive Guided Differential Evolution Algorithm (AGDE) (SMA-AGDE). The AGDE mutation method is employed to enhance the swarm agents’ local search, increase the population’s diversity, and help avoid premature convergence. The SMA-AGDE’s performance is evaluated on the CEC’17 test suite, three engineering design problems – tension/compression spring, pressure vessel, and rolling element bearing – and two combinatorial optimization problems – bin packing and quadratic assignment. The SMA-AGDE is compared with three categories of optimization methods: (1) The well-studied MAs, i.e., Biogeography-Based Optimizer (BBO), Gravitational Search Algorithm (GSA), and Teaching Learning-Based Optimization (TLBO), (2) Recently developed MAs, i.e., Harris Hawks Optimization (HHO), Manta Ray Foraging optimization (MRFO), and the original SMA, and (3) High-performance MAs, i.e., Evolution Strategy with Covariance Matrix Adaptation (CMA-ES), and AGDE. The overall simulation results reveal that the SMA-AGDE ranked first among the compared algorithms, and so, over different function landscapes. Thus, the proposed SMA-AGDE is a promising optimization tool for global and combinatorial optimization problems and engineering design problems.

Decentralized Control of Large Collaborative Swarms Using Random Finite Set Theory

Controlling large swarms of robotic agents presents many challenges including but not limited to computational complexity due to a large number of agents, uncertainty in the functionality of each agent in the swarm, and uncertainty in the swarm's configuration. The contribution of this work is to decentralize random finite set (RFS)-based control of large collaborative swarms for controlling individual agents. The RFS-based control formulation assumes a Gaussian mixture probability hypothesis density approximation and a complete topology for centralized swarm control. To generalize the control topology in a localized or decentralized manner, a sparse linear quadratic regulator (LQR) is used to sparsify the RFS-based control gain matrix obtained using iterative LQR. This allows agents to use information of agents near each other (localized topology) or only the agent's own information (decentralized topology) to make a control decision. Sparsity and performance for decentralized RFS-based control are compared for different degrees of localization in feedback control gains which show that the stability and performance compared to centralized control do not degrade significantly in providing RFS-based control for large collaborative swarms.

Flocking motion in swarms with limited sensing radius and heterogeneous input constraints

This paper addresses the flocking motion problem for swarms of agents with two restrictions: limited communication/detection ranges and different input constraints. In this problem, the distance between pairs of agents determines if a communication/detection link exists among them, while each agent has a different control action bound. We use the notion of proximity graph to model communication/detection between agents and provide distributed controllers designed for leaderless and leader-followers flocking motion scenarios. Our proposed designs preserve the connectivity of the proximity graph while the control effort satisfies the bound of each particular agent. Unlike previous results, our protocols take advantage of the group’s input heterogeneity to use agents with a greater capacity to compensate for neighbors that are less capable of meeting the group’s requirements. Additionally, our designs are based only on local state errors and are robust to non-modeled edge failures. To illustrate the effectiveness of our proposal, we use numerical simulations of different flocking scenarios.

Tracking and Relative Localization of Drone Swarms With a Vision-Based Headset

We address the detection, tracking, and relative localization of the agents of a drone swarm from a human perspective using a headset equipped with a single camera and an Inertial Measurement Unit (IMU). We train and deploy a deep neural network detector on image data to detect the drones. A joint probabilistic data association filter resolves the detection problems and couples this information with the headset IMU data to track the agents. In order to estimate the drones' relative poses in 3D space with respect to the human, we use an additional deep neural network that processes image regions of the drones provided by the tracker. Finally, to speed up the deep neural networks' training, we introduce an automated labeling process relying on a motion capture system. Several experimental results validate the effectiveness of the proposed approach. The approach is real-time, does not rely on any communication between the human and the drones, and can scale to a large number of agents, often called swarms. It can be used to spatially task a swarm of drones and also employed without a headset for formation control and coordination of terrestrial vehicles.

An Improved Tunicate Swarm Algorithm for Global Optimization and Image Segmentation

This study integrates a tunicate swarm algorithm (TSA) with a local escaping operator (LEO) for overcoming the weaknesses of the original TSA. The LEO strategy in TSA–LEO prevents searching deflation in TSA and improves the convergence rate and local search efficiency of swarm agents. The efficiency of the proposed TSA–LEO was verified on the CEC’2017 test suite, and its performance was compared with seven metaheuristic algorithms (MAs). The comparisons revealed that LEO significantly helps TSA by improving the quality of its solutions and accelerating the convergence rate. TSA–LEO was further tested on a real-world problem, namely, segmentation based on the objective functions of Otsu and Kapur. A set of well-known evaluation metrics was used to validate the performance and segmentation quality of the proposed TSA–LEO. The proposed TSA-LEO outperforms other MA algorithms in terms of fitness, peak signal-to-noise ratio, structural similarity, feature similarity, and segmentation findings.

Human Factors Issues and Challenges In Research on ‘Many- Agent’ Control Applications

Human control of multiple agents in swarm or multi-agent deployments are still novel and understudied—especially when real-world research involving many agents (e.g., 50+) are considered, though many organizations are rushing to deploy them for diverse purposes from package delivery to military engagements. Challenges include the scale of situation awareness, workload and attention splitting associated with division of labor across many functions, human perceptions of emergent swarm behaviors, etc. The challenges in either fielding many automated agents in real world settings using current technologies, or of accurately simulating those agents and behaviors sufficiently are posing difficulties—and will become ever more difficult as scale and complexity increases. This panel will focus on recent efforts to create and evaluate methods to support human control of large numbers of swarm or multi-agent vehicles, with an emphasis on deployment and lessons from real world “in the wild” evaluations using multiple representative machine agents.

Random Finite Set Theory and Centralized Control of Large Collaborative Swarms

Controlling large swarms of robotic agents presents many challenges, including, but not limited to, computational complexity due to a large number of agents, uncertainty in the functionality of each agent in the swarm, and uncertainty in the swarm’s configuration. This work generalizes the swarm state using random finite set (RFS) theory and solves a centralized control problem with a quasi-Newton optimization through the use of model predictive control (MPC) to overcome the aforementioned challenges. This work uses the RFS formulation to control the distribution of agents assuming an unknown or unspecified number of agents. Computationally efficient solutions are also obtained via the MPC version of the iterative linear quadratic regulator (ILQR), a variant of differential dynamic programming. Information divergence is used to define the distance between the swarm RFS and the desired swarm configuration through the use of the modified distance. Simulation results using MPC and ILQR show that the swarm intensity converges to the desired intensity. Additionally, the RFS control formulation is shown to be very flexible in terms of the number of agents in the swarm and configuration of the desired Gaussian mixtures. Lastly, the ILQR and the Gaussian mixture probability hypothesis density filter are used in conjunction to solve a spacecraft relative motion problem with imperfect information to show the viability of centralized RFS control for this real-world scenario.

Bioinspired Visuomotor Feedback in a Multiagent Group/Swarm Context

Practical aerial swarm applications are limited by the need to provide fast, robust feedback. To consider how biological agents incorporate visual feedback in multiagent environments, this study extends a well-developed model of bioinspired visual feedback for individual agent to feedback in multiagent dynamic environments. A multiagent visual model is developed and analyzed for an individual agent operating in a group of other agents. The theoretical model indicates a stable equilibrium trajectory within the neighboring agents. When a minority of agents conducting bioinspired visual feedback are introduced into a planar swarm simulation, results show robustness of agents to perturbations from reference condition. The bioinspired feedback rule is implemented experimentally using ground microbots for two cases identified by a theoretical observability condition. The experimental performance of the single agent in an established swarm is consistent with the theoretical stability and observability predictions. The results indicate insect visuomotor response can help participate in established swarms, and indicate the importance of optic flow input in maintaining group motion.

Controllability and Stabilization for Herding a Robotic Swarm Using a Leader: A Mean-Field Approach

In this article, we introduce a model and a control approach for herding a swarm of “follower” agents to a target distribution among a set of states using a single “leader” agent. The follower agents evolve on a finite state space that is represented by a graph and transition between states according to a continuous-time Markov chain (CTMC), whose transition rates are determined by the location of the leader agent. The control problem is to define a sequence of states for the leader agent that steers the probability density of the forward equation of the Markov chain. For the case, when the followers are possibly interacting, we prove local approximate controllability of the system about equilibrium probability distributions. For the case, when the followers are noninteracting, we design two switching control laws for the leader that drive the swarm of follower agents asymptotically to a target probability distribution that is positive for all states. The first strategy is open-loop in nature, and the switching times of the leader are independent of the follower distribution. The second strategy is of feedback type, and the switching times of the leader are functions of the follower density in the leader's current state. We validate our control approach through numerical simulations with varied numbers of follower agents that evolve on graphs of different sizes, through a 3-D multirobot simulation in which a quadrotor is used to control the spatial distribution of eight ground robots over four regions, and through a physical experiment in which a swarm of ten robots is herded by a virtual leader over four regions.

AutoRWN: automatic construction and training of random weight networks using competitive swarm of agents

Random Weight Networks have been extensively used in many applications in the last decade because it has many strong features such as fast learning and good generalization performance. Most of the traditional training techniques for Random Weight Networks randomly select the connection weights and hidden biases and thus suffer from local optima stagnation and degraded convergence. The literature shows that stochastic population-based optimization techniques are well regarded and reliable alternative for Random Weight Networks optimization because of high local optima avoidance and flexibility. In addition, many practitioners and non-expert users find it difficult to set the other parameters of the network like the number of hidden neurons, the activation function, and the regularization factor. In this paper, an approach for training Random Weight Networks is proposed based on a recent variant of particle swarm optimization called competitive swarm optimization. Unlike most of Random Weight Networks training techniques, which are used to optimize only the input weights and hidden biases, the proposed approach will automatically tune the weights, biases, the number of hidden neurons, and regularization factor as well as the embedded activation function in the network, simultaneously. The goal is to help users to effectively identify a proper structure and hyperparameter values to their applications while obtaining reasonable prediction results. Twenty benchmark classification datasets are used to compare the proposed approach with different types of basic and hybrid Random Weight Network-based models. The experimental results on the benchmark datasets show that the reasonable classification results can be obtained by automatically tuning the hyperparameters using the proposed approach.

A Novel Paradigm for Underwater Monitoring Using Mobile Sensor Networks.

This paper presents a novel autonomous environmental monitoring methodology based on collaboration and collective decision-making among robotic agents in a heterogeneous swarm developed within the project subCULTron, tested in a realistic marine environment. The swarm serves as an underwater mobile sensor network for exploration and monitoring of large areas. Different robotic units enable outlier and fault detection, verification of measurements and recognition of environmental anomalies, and relocation of the swarm throughout the environment. The motion capabilities of the robots and the reconfigurability of the swarm are exploited to collect data and verify suspected anomalies, or detect potential sensor faults among the swarm agents. The proposed methodology was tested in an experimental setup in the field in two marine testbeds: the Lagoon of Venice, Italy, and Biograd an Moru, Croatia. Achieved experimental results described in this paper validate and show the potential of the proposed approach.

Optimization test of a rule-based swarm intelligence simulation for the conceptual design process

AbstractToday, in the field of architecture, bio-inspired algorithms can be used to design and seek solutions to design problems. Two of the most popular algorithms are the genetic algorithm (GA) and swarm intelligence algorithm. However, no study has examined the simultaneous use of these two bio-inspired algorithms in the field of architecture. Therefore, this study aims to test whether these two bio-inspired algorithms can work together. To this end, GA is used in this study to optimize the rule-based swarm algorithm for the conceptual design process. In this optimization test, the objective was to increase the surface area, and the constraints are parcel boundary and building height. Further, the forms associated with swarm agents were determined as variables. Following the case studies, the study concludes that the two bio-inspired algorithms can effectively work together.

Collective olfactory search in a turbulent environment.

Finding the source of an odor dispersed by a turbulent flow is a vital task for many organisms. When many individuals concurrently perform the same olfactory search task, sharing information about other members' decisions can potentially boost the performance. But how much of this information is actually exploitable for the collective task? Here we show, in a model of a swarm of agents inspired by moth behavior, that there is an optimal way to blend the private information about odor and wind detections with the public information about other agents' heading direction. Our results suggest an efficient multiagent olfactory search algorithm that could prove useful in robotics, e.g., in the identification of sources of harmful volatile compounds.