A5037488149- Distributed Control Multi-Agent Systems
- Modular Robots and Swarm Intelligence
- Anomaly Detection Techniques and Applications
- Fault Detection and Control Systems
- Robotic Path Planning Algorithms
- Robotic Locomotion and Control
- Evacuation and Crowd Dynamics
- Micro and Nano Robotics
- Reinforcement Learning in Robotics
- Advanced Statistical Process Monitoring
- Mineral Processing and Grinding
- Neural Networks and Reservoir Computing
- Machine Fault Diagnosis Techniques
- Gaussian Processes and Bayesian Inference
- Target Tracking and Data Fusion in Sensor Networks
- UAV Applications and Optimization
- Optimization and Search Problems
- Fluid Dynamics Simulations and Interactions
- Data Stream Mining Techniques
- Insect and Arachnid Ecology and Behavior
- Power System Reliability and Maintenance
- Distributed Sensor Networks and Detection Algorithms
- Neural dynamics and brain function
- Model Reduction and Neural Networks
- Internet Traffic Analysis and Secure E-voting
École Nationale Supérieure d'Informatique
2021-2023
Laboratoire de Recherche en Informatique
2022
Nottingham Trent University
2022
Universitat Politècnica de València
2022
Brunel University of London
2022
University of Biskra
2015-2020
As an emergent research area by which swarm intelligence is applied to multi-robot systems; robotics (a very particular and peculiar sub-area of collective robotics) studies how coordinate large groups relatively simple robots through the use local rules.It focuses on studying design amount robots, their physical bodies controlling behaviors.Since its introduction in 2000, several successful experimentations had been realized, till now more projects are under investigations.This paper seeks...
Machine Learning (ML) for swarm motion prediction is a relatively unexplored area that could help sustain and monitor daily robotics collective tasks. This paper focuses on specific application of which pattern formation, to demonstrate the ability Ensemble (EL) approaches predict speed robots. Specifically, boosted trees (BST) bagged (BT) algorithms are introduced miniature two-wheels differential driver mobile robots performing circle-formation via viscoelastic control model. choice's...
Aggregation is a vital behavior when performing complex tasks in most of the swarm systems, such as robotics systems. In this paper, three new aggregation methods, namely distance-angular, distance-cosine, and distance-Minkowski k-nearest neighbor (k-NN) have been introduced. These methods are mainly built on well-known metrics: cosine, angular, Minkowski distance functions, which used here to compute distances among robots' neighbors. Relying these each robot identifies its neighborhood set...
We report a swarm robots circle formation control model that is based on intra virtual viscoelastic connectivity between the neighbors. The can dynamically form uniform using only distance estimation among neighbors, fully decentralized and scalable by which be formed whatever number of being implicated. Based equilibrium forces exerted robots, could positions at equal angular ranges boundary with or without positioning robot center circle. A developed, implemented evaluated ARGoS simulator.
In this work, we study a simple collective behaviour, called aggregation, performed by swarm of mobile robots system. We mainly proposed the Distance-Minkowski k-Nearest Neighbours (DM-KNN) as new approach to aggregation behaviour The method introduced Minkowski distance function in computing distances between robots' neighbours. approach, set k-nn members with which each robot will interact is identified. Then an artificial viscoelastic mesh among built perform aggregation. When Analyzing...
Abstract Collective decision-making by a swarm of robots is paramount importance. In particular, the problem collective perception wherein aims to achieve consensus on prevalent feature in environment. Recently, this has been formulated as discrete estimation scenario estimate their proportion rather than deciding about one. Nevertheless, performance existing strategies resolve either poor or depends higher communication bandwidth. work, we propose novel strategy based maximum likelihood...
An innovative and flexible approach is introduced to address the challenge of self-organizing a group mobile robots into cubic-spline-based patterns without any requirement control points. Besides self-organization robots, incorporates potential field-based for obstacle/collision avoidance. This will offer more flexibility swarm efficiently deal with many practical situations, including smoothly avoiding obstacles during movement or exploring covering areas complex curved patterns....
Fault detection in robotic swarm systems is imperative to guarantee their reliability, safety, and maximize operating efficiency avoid expensive maintenance. However, data from these are generally contaminated with noise, which masks important features the degrades fault capability. This paper introduces an effective approach against noise uncertainties data, integrates multiresolution representation of using wavelets sensitivity small changes exponentially weighted moving average scheme....
Detecting anomalies in a robot swarm play core role keeping the desired performance, and meeting requirements specifications. This paper deals with problem of detecting swarm. In this regards, an unsupervised monitoring approach based on principal component analysis k-nearest neighbor is proposed. The model employed to generate residuals for anomaly detection. Then, are examined by computing proposed exponentially smoothed statistic purpose Here, instead using parametric thresholds derived...
In this paper, a Distance-Weighted K Nearest Neighboring (DW-KNN) topology is proposed to study self-organized aggregation as an emergent swarming behavior within robot swarms. A virtual physics approach applied among the neighborhood keep robots together. distance-weighted function based on Smoothed Particle Hydrodynamic (SPH) interpolation used key factor identify K-Nearest neighbors taken into account when aggregating robots. The intra physical connectivity these achieved using...
Swarm robotics requires continuous monitoring to detect abnormal events and sustain normal operations. Indeed, swarm with one or more faulty robots leads degradation of performances complying the target requirements. This paper present an innovative data-driven fault detection method for swarm. The combines flexibility principal component analysis (PCA) models greater sensitivity exponentially-weighted moving average control chart incipient changes. We illustrate through simulated data...
Extracting local interaction rules that govern the dynamics of a swarm is central challenge in many robotics application domains. Reverse engineer such might be highly beneficial preventing serious design handcrafting errors engineers may implicitly make. Advances data-driven based systems identification techniques, called SINDy, are currently enabling tractable equations governing systems. However, they have not yet to applied In this work, we aim combine sparsity-promoting techniques with...
Fault detection plays an important role in supervising the operation of robotic swarm systems. If faults are not detected, they can considerably affect performance robot swarm. In this paper, we present a robust fault mechanism against noise and uncertainties data, by merging multiresolution representation data using wavelets with sensitivity to small changes exponentially weighted moving average scheme. Specifically, monitor robotics systems performing virtual viscoelastic control model for...
Human-robot interaction has been a major focus of research over the past decade. However, appearance swarm robotics brought different topic for study. Human (HSI), blend biology, robotics, computer science, and psychology, is field that investigates how humans interact with robots. Furthermore, there noticeable need relevant studies using bibliometric analyses in HIS to assess trends growth research. The purpose this paper explore have better grasp HSI issue by analysing 402 works selected...
One of the most significant security concerns confronting network technology is detection distributed denial service (DDOS). This paper introduces a semi-supervised data-driven approach to DDOS attacks. The proposed method employs normal events data without labeling train model. Specifically, this an improved autoencoder (AE) model by incorporating Gated Recurrent Unit (GRU) based on attention mechanism (AM) at encoder and decoder sides AE GRU enhances AE's ability learn temporal...
This work introduces a short term forecasting gated recurrent unites framework for swarm motion speed forecasting. is motivated by the growing need of addressing such challenges in order to keep robotic systems executing daily collective operations and accomplishing tasks more successfully groups. The based on BiGRU model its performances compared base GRU model. built upon sensor measurements collected using simulated e-puck robots performing simple pattern formation task free/no-free...