Flocking Behavior Research Articles

Edge intelligence for poultry welfare: Utilizing tiny machine learning neural network processors for vocalization analysis.

The health of poultry flock is crucial in sustainable farming. Recent advances in machine learning and speech analysis have opened up opportunities for real-time monitoring of the behavior and health of flock. However, there has been little research on using Tiny Machine Learning (Tiny ML) for continuous vocalization monitoring in poultry. This study addresses this gap by developing and deploying Tiny ML models on low-power edge devices to monitor chicken vocalizations. The focus is on overcoming challenges such as memory limitations, processing power, and battery life to ensure practical implementation in agricultural settings. In collaboration with avian researchers, a diverse dataset of poultry vocalizations representing a range of health and environmental conditions was created to train and validate the algorithms. Digital Signal Processing (DSP) blocks of the Edge Impulse platform were used to generate spectral features for studying fowl vocalization. A one-dimensional Convolutional Neural Network (CNN) model was employed for classification. The study emphasizes accurately identifying and categorizing different chicken noises associated with emotional states such as discomfort, hunger, and satisfaction. To improve accuracy and reduce background noise, noise-robust Tiny ML algorithms were developed. Before the removal of background noise, our average accuracy and F1 scores were 91.6% and 0.92, respectively. After the removal, they improved to 96.6% and 0.95.

Perancangan Mesin Klasifikasi Menggunakan Particle Swarm Optimization

Designing an effective classification engine is very important in various pattern recognition and machine learning applications. In this research, the Particle Swarm Optimization (PSO) algorithm is applied for the development of classification engines on various datasets. PSO is a population-based optimization method inspired by the behavior of flocks of birds or fish, which is effectively used to find optimal solutions in large search spaces. This research aims to develop a classification model by using Particle Swarm Optimization (PSO) as a training element to determine weights and biases. To test the performance on several different datasets, namely on a dummy multi-class dataset, Sasak Aksara image dataset, and the well-known Iris dataset. In the Sasak Aksara data, Discrete Cosine Trasnform (DCT) is used as feature extraction with the aim of reducing computation time. The results show that PSO can be used in the implementation of several datasets used, in the classification of dummy data, iris data, and Sasak Aksara image data. The model achieved 100% accuracy, precision, recall, and F1-Score on dummy data and iris data. However, on the Sasak Aksara image dataset, the performance of the model decreased with accuracy only reaching 65%, precision 50%, recall 32%, and F1-Score 39%. This research contributes in demonstrating the effectiveness of PSO in optimizing Perceptron models on simpler datasets and highlights the need for further development to handle more complex datasets.

Route Optimization for UVC Disinfection Robot Using Bio-Inspired Metaheuristic Techniques.

The COVID-19 pandemic highlighted the urgent need for effective surface disinfection solutions, which has led to the use of mobile robots equipped with ultraviolet (UVC) lamps as a promising technology. This study aims to optimize the navigation of differential mobile robots equipped with UVC lamps to ensure maximum efficiency in disinfecting complex environments. Bio-inspired metaheuristic algorithms such as the gazelle optimization algorithm, whale optimization algorithm, bat optimization algorithm, and particle swarm optimization are applied. These algorithms mimic behaviors of biological beings such as the evasive maneuvers of gazelles, the spiral hunting patterns of whales, the echolocation of bats, and the collective behavior of flocks of birds or schools of fish to optimize the robot's trajectory. The optimization process adjusts the robot's coordinates and the time it takes to stops at key points to ensure complete disinfection coverage and minimize the risk of excessive UVC exposure. Experimental results show that the proposed algorithms effectively adapt the robot's trajectory to various environments, avoiding obstacles and providing sufficient UVC radiation exposure to deactivate target microorganisms. This approach demonstrates the flexibility and robustness of these solutions, with potential applications extending beyond COVID-19 to other pathogens such as influenza or bacterial contaminants, by tuning the algorithm parameters. The results highlight the potential of bio-inspired metaheuristic algorithms to improve automatic disinfection and achieve safer and healthier environments.

Emergent behaviors of relativistic flocks with adaptive coupling laws

In this paper, we investigate the emergent behaviors of the relativistic Cucker–Smale (RCS) model equipped with adaptive couplings. To do this, we first divide adaptive couplings into two types, Hebbian or anti-Hebbian. For the Hebbian case, we demonstrate the asymptotic flocking of the RCS model in two ways based on the Lyapunov functional approach and continuous argument. Meanwhile, for the anti-Hebbian case, depending on the regularity of the adaptive law at the origin, we prove the various emergent behaviors such as the slow velocity alignment or the group formation.

Tree structures may be the leadership structures employed by group motions exhibiting hierarchy

Collective behaviors leading to various fascinating movement patterns are believed to be the product of complex interplay among individuals. Previous studies have identified two types of leadership structures in pigeon flocks, i.e., hierarchical networks and reciprocal relationships. However, both of these leadership structures are predicated based on data analysis and lack substantial empirical evidence. Additionally, it is difficult to delineate a direct correspondence between leadership structures and trajectory data for pigeon flocks because birds cannot report their leadership structure. Herein, based on experiments involving volunteers, we found that tree structures may serve as the leadership structures employed by group motions exhibiting hierarchy. In the tree structure, each follower follows its only leader during collective motions, and the single top leader determines the direction of collective motions. By employing the tree structure, we introduced a model for describing collective motions. A method for obtaining the leadership structure through experimental trajectory data was proposed. This strategy was used to analyze flight trajectory data from a pigeon flock and elucidate the pigeons' leadership relationships. Our model can simulate the collective behavior of pigeon flocks, thus accurately replicating the findings of previous experimental studies. The results of this study provide insights regarding the leadership structure in pigeon flocks and have implications for artificial collective systems, e.g., autonomous formation control of multiple unmanned aerial vehicles or unmanned surface vehicles. Published by the American Physical Society 2024

Spraying drones: efficacy of integrating an avian repellent with drone hazing to elicit blackbird flock dispersal and abandonment of sunflower fields

Multiple management strategies exist to combat bird damage to agriculture. We explored combining two tools, drones as frightening devices and an avian repellent, to assess effectiveness of an integrated method to deter large flocks on complex landscapes. We evaluated the ability of a spraying drone (DJI Agras MG‐1P) deploying Avian Control (i.e. active ingredient: methyl anthranilate; hereafter MA) or water to elicit abandonment, flock reductions, latency to return, and behavioral changes of blackbirds (Icteridae) foraging in sunflower Helianthus annuus. Following hazing and spraying (MA = 32; water = 32 trials), the percent of flocks abandoning, partially abandoning, or remaining was 56%, 31%, and 13% for MA and 50%, 25%, and 25% for water, respectively. Following full abandonment, 14% more flocks returned following MA (83%) than water (69%), averaging 3.96 min ± 3.51 (SD) to return (MA = 4.12 min ± 4.03; water = 3.73 min ± 2.68). When reduction occurred, average decline was 47% ± 35 (SD) with MA and 44% ± 39 (SD) with water. Site conditions impacted the ability to maneuver the drone and observe flock behaviors, potentially resulting in variables other than treatment explaining the probability of abandonment and flock reduction. When controlling for flock size, number of lift‐offs following water treatments (0.23 min−1 ± 0.17 SD) was statistically less than the pre‐hazing period (0.43 min−1 ± 0.28 SD), however no relationship existed for MA treatments (post: 0.29 min−1 ± 0.32 SD; pre: 0.31 min−1 ± 0.20 SD). This difference may be due to a longer latency to return, decreasing the post‐hazing time period, or flightier birds after MA exposure. We found eight mins of hazing, and a 9‐l tank of repellent, was insufficient to elicit differences between water and repellent applications. We suggest extended hazing or additional negative stimuli (e.g. multiple drones, increased repellent) to increase efficacy.

DCWPSO: particle swarm optimization with dynamic inertia weight updating and enhanced learning strategies.

Particle swarm optimization (PSO) stands as a prominent and robust meta-heuristic algorithm within swarm intelligence (SI). It originated in 1995 by simulating the foraging behavior of bird flocks. In recent years, numerous PSO variants have been proposed to address various optimization applications. However, the overall performance of these variants has not been deemed satisfactory. This article introduces a novel PSO variant, presenting three key contributions: First, a novel dynamic oscillation inertia weight is introduced to strike a balance between exploration and exploitation; Second, the utilization of cosine similarity and dynamic neighborhood strategy enhances both the quality of solution and the diversity of particle populations; Third, a unique worst-best example learning strategy is proposed to enhance the quality of the least favorable solution and consequently improving the overall population. The algorithm's validation is conducted using a test suite comprised of benchmarks from the CEC2014 and CEC2022 test suites on real-parameter single-objective optimization. The experimental results demonstrate the competitiveness of our algorithm against recently proposed state-of-the-art PSO variants and well-known algorithms.

Self-organized vortex phases and hydrodynamic interactions in Bos taurus sperm cells.

Flocking behavior is observed in biological systems from the cellular to superorganismal length scales, and the mechanisms and purposes of this behavior are objects of intense interest. In this paper, we study the collective dynamics of bovine sperm cells in a viscoelastic fluid. These cells appear not to spontaneously flock, but transition into a long-lived flocking phase after being exposed to a transient ordering pulse of fluid flow. Surprisingly, this induced flocking phase has many qualitative similarities with the spontaneous polar flocking phases predicted by Toner-Tu theory, such as anisotropic giant number fluctuations and nontrivial transverse density correlations, despite the induced nature of the phase and the clearly important role of momentum conservation between the swimmers and the surrounding fluid in these experiments. We also find a self-organized global vortex state of the sperm cells, and map out an experimental phase diagram of states of collective motion as a function of cell density and motility statistics. We compare our experiments with a parameter-matched computational model of persistently turning active particles and find that the experimental order-disorder phase boundary as a function of cell density and persistence time can be approximately predicted from measures of single-cell properties. Our results may have implications for the evaluation of sample fertility by studying the collective phase behavior of dense groups of swimming sperm.

Anomalous flocking in nonpolar granular Brownian vibrators

Using Brownian vibrators, we investigated the structures and dynamics of quasi-2d granular materials, with packing fractions (ϕ) ranging from 0.111 to 0.832. Our observations revealed a remarkable large-scale flocking behavior in hard granular disk systems, encompassing four distinct phases: granular fluid, flocking fluid, poly-crystal, and crystal. Anomalous flocking emerges at ϕ = 0.317, coinciding with a peak in local density fluctuations, and ceased at ϕ = 0.713 as the system transitioned into a poly-crystal state. The poly-crystal and crystal phases resembled equilibrium hard disks, while the granular and flocking fluids differed significantly from equilibrium systems and previous experiments involving uniformly driven spheres. This disparity suggests that collective motion arises from a competition controlled by volume fraction, involving an active force and an effective attractive interaction resulting from inelastic particle collisions. Remarkably, these findings align with recent theoretical research on the flocking motion of spherical active particles without alignment mechanisms.

A comprehensive survey on the chicken swarm optimization algorithm and its applications: state-of-the-art and research challenges

The application of optimization theory and the algorithms that are generated from it has increased along with science and technology's continued advancement. Numerous issues in daily life can be categorized as combinatorial optimization issues. Swarm intelligence optimization algorithms have been successful in machine learning, process control, and engineering prediction throughout the years and have been shown to be efficient in handling combinatorial optimization issues. An intelligent optimization system called the chicken swarm optimization algorithm (CSO) mimics the organic behavior of flocks of chickens. In the benchmark problem's optimization process as the objective function, it outperforms several popular intelligent optimization methods like PSO. The concept and advancement of the flock optimization algorithm, the comparison with other meta-heuristic algorithms, and the development trend are reviewed in order to further enhance the search performance of the algorithm and quicken the research and application process of the algorithm. The fundamental algorithm model is first described, and the enhanced chicken swarm optimization algorithm based on algorithm parameters, chaos and quantum optimization, learning strategy, and population diversity is then categorized and summarized using both domestic and international literature. The use of group optimization algorithms in the areas of feature extraction, image processing, robotic engineering, wireless sensor networks, and power. Second, it is evaluated in terms of benefits, drawbacks, and application in comparison to other meta-heuristic algorithms. Finally, the direction of flock optimization algorithm research and development is anticipated.

Interaction Rules Supporting Effective Flocking Behavior.

Several simulation models have demonstrated how flocking behavior emerges from the interaction among individuals that react to the relative orientation of their neighbors based on simple rules. However, the precise nature of these rules and the relationship between the characteristics of the rules and the efficacy of the resulting collective behavior are unknown. In this article, we analyze the effect of the strength with which individuals react to the orientation of neighbors located in different sectors of their visual fields and the benefit that could be obtained by using control rules that are more elaborate than those normally used. Our results demonstrate that considering only neighbors located on the frontal side of the visual field permits an increase in the aggregation level of the swarm. Using more complex rules and/or additional sensory information does not lead to better performance.

Optimization of the concentration of ozone generated by DBD using PSO algorithm for water treatment process

The water treatment process with ozone is influenced by various operating parameters and environmental factors that can impact its efficiency. In this study, experiments were conducted using a Venturi pumping frame to investigate the effects of three controllable variables: oxygen flow height, applied voltage level, and water flow rate. The tests aimed to develop a mathematical model that accurately represents the relationship between these input variables and the resulting ozone concentration in the treated water. The experimental data was analyzed using the MODDE 5.0 software, a specialized application for statistical modeling and design of experiments. By fitting the data to appropriate model equations, a mathematical model was obtained that quantifies the influence of each variable and their interactions on the ozone concentration response. To optimize the process performance, a particle swarm optimization (PSO) algorithm was employed to extract the best-fit parameters for the mathematical model. PSO is a computational technique inspired by the social behavior of bird flocks or fish schools, utilizing a population of candidate solutions that evolve iteratively to converge on the global optimum solution. In this case, PSO searched for the model parameter values that minimized the error between predicted and experimentally measured ozone concentrations, rapidly converging to an accurate solution. The optimized mathematical model enables predicting the ozone concentration under any combination of oxygen flow height, voltage, and water flow rate within the experimental range. This predictive capability facilitates identifying the optimum operating conditions to maximize ozone concentration, thereby enhancing the efficiency of the water treatment process. The model serves as a valuable tool for process control, monitoring, and optimization, ensuring consistent treatment quality while minimizing resource consumption and operational costs.

UAV formation control using enhanced behavior mechanism and artificial potential field

Inspired by formation flight of pigeon flock, this paper proposes a enhanced method of autonomous formation control of multiple Unmanned Aerial Vehicles (UAVs) that can maintain high symmetry based on pigeon flock behavior mechanism. Addressing the instability of formation in the original method, the follow improvements have been made. Firstly, improve leadership of top three UAVs, Secondly, modify artificial potential field strategies for top two followers. Finally, through a series of simulation experiments, it is verified that the UAVs can form the expected formation under the autonomous formation control, and can maintain the formation under the complex motion of leader UAV.

Plantation forestry of Alder-leaf Birch (Betula alnoides) affects composition but not interactions of mixed-species bird flocks in southwestern China

Human disturbance impacts mixed-species bird flocks (“flocks”). Unfortunately, the impact on flocks by one large-scale disturbance, plantation forestry, has remained little explored. We examined how plantation forestry of a widespread yet understudied timber species, Alder-leaf Birch (Betula alnoides, “birch”), affects the composition and interactions of flocks in the Indo-Burma biodiversity hotspot in southwestern China. We conducted transect surveys to sample flocks in birch plantations and natural forests of two age classes (mature and young). While flock size and species richness per flock were similar across land-use types, rarefied species richness accounting for unequal sampling effort was noticeably higher in mature forests. Furthermore, flock composition differed across land-use types, with differences related to species' morphological characteristics, dietary preferences, and foraging substrates. Specifically, mature forests supported flock participants with large bodies, poor dispersal ability, and a fruit-eating diet. Birch plantations offered equal support to invertivores as both natural forests, and additional support to bark foragers. Lastly, interactions among flocking species quantified by social network metrics were similar across land-use types, suggesting that birch plantations perserved the flocking behavior itself. Our study reveals the conservation potential of birch plantations in supporting invertivorous birds and preserving interactions in flocks. More importantly, it highlights the irreplaceability of mature forests because of their unique species composition. We recommend promoting birch planting without compromising local economies and protecting remnant mature forests through education programs and continued research.

Flocking Behavior of Boids Driven by Hyperchaotic MACM System

In the present work a detailed study is presented, on the design, programming, and investigation of the behavior of flocking (Movement type flock), through the model of BOIDS, for its acronym in English "Bird Oid Object" (Object type bird), which was devised by Craig Reynolds in 1986. This complex flocking behavior that occurs arises from the interaction of simple local rules, in which complexity and sensitivity to initial conditions are present. A measure of chaotic compound will be introduced to the algorithm by means of a new four-dimensional autonomous hyperchaotic system based on the 3D Méndez-Arellano-Cruz-Martínez (MACM) system. The measures proposed herein, therefore, may have the potential to predict, control, and exemplify the behavior of group intelligence study systems that occur in nature, allowing the implementation of these systems in groups of robots through the implementation of hyperchaotic trajectories in the future, to obtain greater speed and efficiency, obstacle and collisions avoidance in their flights.

Behavioral Discrimination in Chicken Flocks: A Camera-Based Movement Analysis for Distinguishing Normal and Scared Behaviors

Assessing chicken behavior in coops is crucial for high-quality meat production and stress prevention. This paper introduces an innovative system for evaluating behavior in chicken flocks by integrating collective area and its rate of change, surpassing traditional metrics like individual chicken distance walked, velocity, and acceleration. The novelty lies in simultaneously using multiple methods, providing a nuanced understanding of chicken responses and enhancing precision in behavior discrimination. The proposed system utilizes camera-based movement analysis, employing procedural image processing and data association for early detection and intervention in closed cage systems. This cost-effective solution is accessible to local farmers, contributing to improved poultry farming practices. The hardware setup includes a camera with a 470 nm bandpass lens, yielding computable results suitable for procedural system processes, especially in streamlining thresholding image processing.

Automating poultry farm management with artificial intelligence: Real-time detection and tracking of broiler chickens for enhanced and efficient health monitoring.

The broiler industry plays a vital role in meeting the growing global demand for poultry meat. However, maintaining the health and well-being of broiler birds is crucial to ensure both optimal productivity and animal welfare. The increasing mortality rate of broiler chickens developed into an unavoidable issue that required attention. The major goal of this research is to monitor individual chickens for early disease identification, which will then allow for prompt isolation and treatment of sick birds, stopping the spread of pathogens and preserving the health of the flock as a whole. With an accuracy of 96%, the chosen model, YOLOv5s (You Only Look Once), performed the best. Based on their age, the algorithm was able to categorise broiler chickens. The model is converted to ONNX (Open Neural Network Exchange) format after custom training, and the centroid tracker is used for real-time tracking. After that, the output data is kept in a MySQL (My Structured Query Language) database for later use. The OpenCV (Open-Source Computer Vision Library) library is used to deploy this model on a local machine. This model seeks to identify the broiler chicken in the video frame, classify them, and maintain track of them using the tracker. Based on their age, the birds are divided into categories. Since most monitoring is required between 1 and 4weeks of age, they are divided into four age groups. The potential application of this model is in the detection of temperature, weight, flock behaviour, etc.

Dynamic Leadership Mechanism in Homing Pigeon Flocks.

In recent years, an increasing number of studies have focused on exploring the principles and mechanisms underlying the emergence of collective intelligence in biological populations, aiming to provide insights for human society and the engineering field. Pigeon flock behavior garners significant attention as a subject of study. Collective homing flight is a commonly observed behavioral pattern in pigeon flocks. The study analyzes GPS data during the homing process and utilizes acceleration information, which better reflects the flock's movement tendencies during turns, to describe the leadership relationships within the group. By examining the evolution of acceleration during turning, the study unveils a dynamic leadership mechanism before and after turns, employing a more intricate dynamic model to depict the flock's motion. Specifically, during stable flight, pigeon flocks tend to rely on fixed leaders to guide homing flight, whereas during turns, individuals positioned in the direction of the flock's turn experience a notable increase in their leadership status. These findings suggest the existence of a dynamic leadership mechanism within pigeon flocks, enabling adaptability and stability under diverse flight conditions. From an engineering perspective, this leadership mechanism may offer novel insights for coordinating industrial multi-robot systems and controlling drone formations.

Study on organic goat production as a tool for sustainability of nomadic pastoralism in north-western Himalayan region, India

Migratory Gaddi sheep and goat constitutes approximately 70% of the total population of small ruminants in Himachal Pradesh, which is in north-western Himalayan region of India. The present investigation was carried out on 32 Gaddi migratory flocks consisting of around 5000 animals (sheep and goat) through structured pre-validated questionnaire to find out scope of nomadic pastoralism for organic goat production. Through cross-sectional surveys, it was found that 100% of breeds reared by the nomads are indigenous in origin, which is well adapted for grazing and browsing in different hilly terrains and topographies. Tethering was not at all practiced among the nomadic flocks of goats. The flocks were herd in corrals during night to prevent predatory loss. The main source of nutrition for goats was rangeland, natural pastures and common property resources. The nomadic goat production system does not substantially compromise the welfare of the animals. Natural suckling and freedom of social behaviour in nomadic flocks are the ethological advantages same as that are required for standards of organic production. Natural breeding is commonly practiced in this nomadic goat production system whereas artificial insemination, embryo transfer and estrus synchronization is unpopular. Ethno-veterinary practices of Gaddi farmers give them an edge over other countries in organic goat production; it can be effective substitute for allopathic medicines but needs to be validated on efficacy and dosage. Exploring and supporting nomadic pastoralism for organic production will help in improving its traceability, meeting consumer satisfaction and sustaining the dwindling nomadic pastoralism.

Distributed Adaptive Flocking Control for Large-Scale Multiagent Systems.

This article presents a novel distributed flocking control method for large-scale multiagent systems (LS-MASs) operating in uncertain environments. When dealing with a massive number of flocking agents in uncertain environments, existing flocking methods encounter the problem of communication complexity and "Curse of dimensionality" caused by the exponential growth of agent interactions while solving PDE-based optimal flocking control for large-scale systems. The mean field game (MFG) method addresses this issue by transforming interactions among all agents into the interaction of each individual agent with average effects represented by a probability density function (pdf) of other agents. However, relying solely on a pdf term to consider other agents' states can result in inefficient flocking performance due to the absence of a proficient coordination mechanism encompassing all agents involved in flocking. To overcome these difficulties and achieve the desired flocking performance for LS-MASs, the agents are decomposed into a finite number of subgroups. Each subgroup comprises a leader and followers, and a hybrid game theory is developed to manage both inter-and intragroup interactions. The method incorporates a cooperative game that links leaders from different groups to formulate distributed flocking control, a Stackelberg game that teams up leaders and followers within the same group to extend collective flocking behavior, and an MFG for followers to address the challenges of LS-MASs. Furthermore, to achieve distributed adaptive flocking using the hybrid game structure, we propose a hierarchical actor-critic-mass-based reinforcement learning technique. This approach incorporates a multiactor-critic method for leaders and an actor-critic-mass algorithm for followers, enabling adaptive flocking control in a distributed manner for large-scale agents. Finally, numerical simulation including comparison study and Lyapunov analysis demonstrates the effectiveness of the developed method.