ICR 2024

Autonomous Landing Algorithm for UAV on a Mobile Robotic Platform with a Fractal Marker

2024-01-22T11:20:46+00:00

This article describes experiments related to the simulation of automatic landing of UAV on a mobile robotic platform using computer vision and a control system based on PID and polynomial regulators within the Gazebo environment. Algorithm has been developed to generate control inputs for maintaining the velocities of the UAV based on the data from the computer vision system and feedback from onboard sensor devices. A series of experiments were conducted at altitudes ranging from 5 to 20 meters, which allowed for identifying limitations that affect the successful landing of the UAV. Measurements were taken of the landing time and landing error, calculated as the distance between the platform center and the UAV's center of mass upon completion of the landing. The average landing time and error ranged from 19.64 seconds and 0.16 meters at an initial altitude of 5 meters to 121.01 seconds and 0.27 meters at an initial altitude of 20 meters. Analysis of the obtained results revealed that both the average error and landing time increase with the initial altitude, which is associated with the accuracy of marker recognition at altitudes above 15 meters. The obtained results can be valuable for further improvement of systems for automatic landing of UAVs on mobile platforms.

GBMILs: Gradient Boosting Models for Multiple Instance Learning

2024-01-22T11:20:47+00:00

An approach based on using the gradient boosting machine for solving the Multiple Instance Learning (MIL) problem under condition of small tabular data is proposed. The MIL deals with labeled objects called bags which consist of several parts of the objects called instances with unknown labels and each bag label depends on the instance labels. Three modifications of the approach are developed and studied. They are determined by different aggregation functions which combine the intermediate predictions of the instance classes in each bag and allow gradient-based optimization through them. The modifications are based on the following aggregation functions: the Hard Max Aggregation, the Simple Attention Aggregation, and the Ensemble of gradient boosting machines in fusion with the Attention Neural Networks. The former two modifications can use an arbitrary decision tree gradient boosting model, which allows iterative training on loss gradients. The later modification simultaneously optimizes an ensemble of parallel gradient boosting models and the parameters of neural network. Numerical experiments with tabular datasets illustrate the proposed modifications of the approach and their superiority comparing to available MIL approaches accompanied by accuracy improvement up to 8%.

Ground mobile robot localization algorithm based on semantic information from the urban environment

2024-01-22T11:20:48+00:00

This paper presents the SLAM algorithm, which use the semantic information extracted from the urban environment to increase the accuracy of ego-vehicle localization in ORB-SLAM2 system. For this purpose, a semantic segmentation module is added to the standard algorithm to assign an object on each frame to one of a given set of classes. The CARLA Simulator was used as a simulation environment, which generates a photorealistic urban environment with the ability to run an arbitrary number of active elements in it, which usually make localization difficult, causing interference with the system. Based on the environment, a training dataset for semantic segmentation was collected. The training dataset consists of 3,696 pairs of city images and corresponding segmentation masks in which each pixel corresponds to one of 23 semantic labels. Using this dataset, the DeepLabV3+ segmentation model was trained with mean per-class IoU metric equals to 81.48%. By using semantic information to filter potentially dynamic objects and matching key points, we were able to increase the localization accuracy relative to the base algorithm by an average of 23% and build a semantic map of the environment.

Construction of a three-dimensional UAV movement planner when the latter moves in conditions of difficult terrain

2024-01-22T11:20:49+00:00

Abstract. The construction of high-speed algorithms for planning the routes of robotic platforms in a three-dimensional mapped environment is an urgent task. Known methods of planning based on cellular decomposition of the field of motion in a three-dimensional formulation are severely limited in speed.

This article proposes a method for planning the movement of robotic platforms in this environment, combining the use of the well-known Dijkstra algorithm for constructing a two-dimensional projection curve with subsequent projection reconstruction and multi-stage correction of the target spatial piecewise polyline curve. The restoration of the original spatial curve by its two-dimensional projection onto the horizontal plane is performed on the basis of a given discrete elevation map of the motion area, and the specified adjustment is made taking into account the requirements, firstly, the minimality of the total length of the final piecewise polyline, and, secondly, taking into account the specified known kinematic limitations of the apparatus.

The algorithm for the synthesis of a spatial curve is detailed for the common case when obstacles are represented in the form of rectangular cylinders with polygonal generators.

The effectiveness of the developed global scheduler algorithm is confirmed by the results of numerical modeling.

Optimization of the placement of measurement points and control of the power of moving sources in rod heating

2024-01-22T11:20:51+00:00

The problem of synthesis of power control of the sources moving according to the given rules and trajectories when the rod is heated is considered. The current values of the controls are determined depending on the values of the temperature of the bar at the points of measurement. Formulas for the components of the gradient of the objective functional are obtained with respect to the feedback parameters and the coordinates of the measurement points, which are used to numerically solve the test problem using first-order numerical optimization methods. The results of computer experiments are presented.

Approach to Numerical Solution of Nonlinear Optimal Feedback Control Problems

2024-01-22T11:20:52+00:00

In the paper, we consider optimal feedback control problems for dynamic, in the general case, nonlinear systems with lumped parameters based on continuous and discrete feedback on the object’s state. To calculate the values of the parameters of the feedback control, we propose to use the measured values of the observable components of the phase vector or the output of the object at the current and some previous (past) moments of time in order to compensate for the inability to measure all the components of the object’s phase state. As a result of this kind of formation of the dependence of the parameters of the synthesized control on a part of the object’s state, the process under consideration will be described by ordinary differential equations with time-constant delay arguments in the phase state. The feedback control problem is solved numerically by reducing it to a finite-dimensional optimization problem. To this end, we derive formulas for the gradient components of the objective functional of the reduced problem with respect to the optimizable parameters - the zonal values of the feedback parameters. These formulas made it possible to formulate necessary first-order optimality conditions, as well as to use them for numerical solution to model problems using first-order optimization methods.

Remote Control Robotic System For The Perimeter Security

2024-01-22T11:20:54+00:00

The the results of the development, preparation and preliminary laboratory and field tests of the robotic security system along the perimeter of any object remotely controlled by radio waves have been presented in paper. Radio signals are coded using a cryptographic method. The robotic system consists of a remote control and controlled units (blocks) that can perform various functions to protect the perimeter of the selected zone. The principle of operation and the structure of the robotic security system are presented. To ensure safety, the blocks are controlled by coded signals. The working principle, management, operation rules, management and telemetry of the security system were explained. Cryptographic coding based on the Neyman algorithm was used to encrypt the telemetry of the security system. The robotic security system can be used for the protection of territory, borders or objects, including for military purposes. The working principle and management of the protection system are given in the paper.

Model-based Policy Optimization with Neural Ordinary Differential Equations

2024-01-22T11:20:54+00:00

Applying learning-based control methods to real robots presents hard challenges, including the low sample efficiency of model-free reinforcement learning algorithms. The widely adopted approach to tackling this problem uses an environment dynamics model. We propose to use the Neural Ordinary Differential Equations to approximate transition dynamics as this allows for finer control of a trajectory generation process. We check our approach on several tasks from simulation environment including learning 6-DoF robotic arm to open the door, which represents particular challenges for policy search. The NODE model is trained to predict movement of the arm and the door, and is used to generate trajectories for the model-based policy optimization. Our method shows better sample efficiency on this task comparing to the model-free and model-based baseline. It also shows comparable results on several other tasks.

Development of a mechanism for recognizing the emotional state based on the unconscious movements of the subject

2024-01-22T11:20:55+00:00

The paper deals with the creation of a mechanism capable of detecting the emotional state of a person based on his interaction with a computer. To solve the problem, we collected a dataset using a web application developed in Python. The application is focused on accumulating data about the subject's cursor movement, such as the distance travelled, the maximum deviation from the line which connects the start and end points of movement, the time of interaction with the mouse during the session and the maximum speed of the cursor movement. Trained classifiers for emotion analysis based on human control of the computer mouse on the created dataset and analyzed the results.

Attention guided In-Hand Mechanical Tools Recognition in Human-Robot Collaborative Process

2024-01-22T11:20:55+00:00

The task of recognition of human behavior in a collaborative robotic system is crucial for the organization of seamless and productive collaboration. We design a vision system for the industrial scenario for riveting a metal plate, and concentrate on the task of recognizing in-hand mechanical tools. However, there is a severe occlusion problem during hand-object interaction process. Incorporating attention modules into the backbone part are often utilized to handle occlusion and enhance the ability of extract features with contextual information. In view of that, three modified occlusion-aware models based on YOLOv5 for in-hand mechanical tools recognition are proposed: by adding SimAM into each of bottleneck network in the backbone part, inserting a Criss-Cross attention layer between the last C3 block and the SPPF block of the back-bone network, and replacing the last C3 block of the backbone network with Criss-Cross attention layer. We create a dataset specifically for our task of in-hand mechanical tools recognition and validate four modified models after training separately, which proves the effectiveness of SimAM module and ineffectiveness of Criss-Cross attention module. The real-time detection is still imperfect under the occlusion of various directions of the hands.

Movement Along the Trajectory of a Home Quadruped Robot

2024-01-22T11:20:57+00:00

Home walking robots imitating pets have a high appeal due to increased maneuverability in a cramped home environment. Unlike robots on a wheelbase, walking home robots can climb stairs, get up when tipping over and change their height when overcoming obstacles. Planning the movements of a home robot is an important component of the control system of a quadruped walking robot. The article deals with the problem of following the trajectory of a walking robot, which relates to motion planning. The article presents a model of a robot and a mathematical model of its legs. Two methods of approximation of the trajectory of motion are proposed in the article – piecewise linear approximation and approximation by arcs of a circle. The use of piecewise linear approximation makes it possible to solve the problem using simple robot movements. The use of approximation by arcs of a circle allows you to build a universal gait for a walking robot. Such a gait can replace several different types of gaits of a quadruped walking robot. The proposed methods of trajectory approximation for a home robot have a small numerical complexity. The simulation of robot movements using two types of approximations is carried out. The article presents experimental modeling data. It is shown that the average speed of a walking robot with piecewise linear approximation is significantly lower than the speed of the robot moving in a straight line. The article presents the conclusions drawn from the results of experiments.

Experimental Validation of an Interface for a Human-Robot Interaction within a Collaborative Task

2024-01-22T11:20:58+00:00

This paper presents a non-contact UR robots – Virtual Control (UR-VC) system for collaborative robots of the UR family, which is based on computer vision techniques and a virtual interaction interface. The control method involved specific hand movements within the robot's virtual control zones. We present the system and the results of an experimental validation. To evaluate the system, we designed a test case that simulates a joint product assembly in a collaborative workspace. 24 participants were involved in the experiments. All participants successfully controlled the robot using the proposed interface and their experience of operating the robot was analyzed via surveys.

An Ontological Approach to the Organization of Computing in Distributed Monitoring Systems with Mobile Components Based on a Distributed Ledger

2024-01-22T11:21:00+00:00

This paper considers issues related to the efficiency of distributed monitoring systems with mobile components based on the concept of fog computing and distributed ledger technologies. The problem of organizing the computational process in terms of placement and workload reallocation is highlighted. It is proposed to use the ontological approach in order to reduce the search space of candidate nodes for the placement of computational load. The core of this approach is the ontological model built with the subject domain in mind. Experimental studies have shown that the application of the ontological approach can reduce the search space by 80%. The obtained data are consistent with the results of the experiment conducted for another subject area.

Moving Person Detection based on modified YOLOv5

2024-01-22T11:21:01+00:00

Visual Dynamic SLAM (Simultaneous Localization and Mapping) is a fundamental technology for intelligent mobile systems, enabling applications in robotics, augmented reality, and self-driving cars. This paper presents a novel approach to improve the performance of Visual Dynamic SLAM by integrating the YOLOv5 object detection framework with attention mechanisms (CBAM) and a BiFPN (Bidirectional Feature Pyramid Network) structure. The dynamic feature points, which are located in the bounding box of the dynamic object, are removed in the tracking thread, and only the static feature points are used to estimate the position of the camera. The primary focus is on improving the detection performance of dynamic objects, particularly persons, and addressing challenges such as occlusion and small object detection. Overall, the integration of YOLOv5 with attention mechanisms and a BiFPN structure presents a significant advancement in visual dynamic SLAM. The proposed approach enhances the detection of persons, addresses challenges related to small objects and occlusion, and improves the overall performance of the system in dynamic environments. These findings demonstrate the effectiveness of the proposed methodology and its potential for real-world applications in various domains, including robotics, augmented reality, and self-driving cars.

Sliding-mode Control of Phase Shift for Two-Rotor Vibration Setup

2024-01-22T11:21:02+00:00

This paper focuses on the development and study of a phase shift control system for a two-rotor vibration mechatronic setup, with the objective of maintaining the desired revolving speed of the rotors. The sliding mode motion is achieved through the use of a relay controller in the phase loop, while PI controllers are employed in the velocity control loops. Numerical study and simulations are conducted using the parameters of the Mechatronic Vibration Setup SV-2M from IPME RAS. The velocity and phase shift control laws are proposed, and the possibility of sliding mode occurrence in the phase shift loop is examined through both analytical and numerical analysis. The simulation results confirm the appearance of sliding mode motion after a finite transient time and demonstrate the dynamical properties of the closed-loop system.

DHC-R: Evaluating “Distributed Heuristic Communication” and Improving Robustness for Learnable Decentralized PO-MAPF

2024-01-22T11:21:03+00:00

Multi-agent pathfinding (MAPF) is a problem of coordinating the movements of multiple agents operating a shared environment that has numerous industrial and research applications. In many practical cases the agents (robots) have limited visibility of the environment and must rely on local observations to make decisions. This scenario, known as partially observable MAPF (PO-MAPF), can be solved through decentralized approaches. In recent years, several learnable algorithms have been proposed for solving PO-MAPF, that leverage the increased availability of data and computational capabilities. However, their performance is oftentimes not validated out-of-distribution (OOD) or in adversarial scenarios, and the code is often not properly open-sourced. In this study, we conduct a comprehensive empirical evaluation of one of the state-of-the-art decentralized PO-MAPF algorithms, Distributed Heuristic Communication (DHC) [1], which incorporates communication between agents. Our experiments reveal that the performance of DHC deteriorates when agents encounter complete packet loss during communication. To address this issue, we propose a novel algorithm called DHC-R (DHC-robust). DHC-R employs a similar architecture to the original DHC but introduces randomness into the Graph Neural Network-based communication block, preventing the passage of some data packets during training. Empirical evaluation confirms that DHC-R outperforms DHC in scenarios with message loss. Open-sourced model weights and the codebase are provided: https://github.com/acforvs/dhc-robust-mapf.

3D-CNNs-based Touchless Human Machine Interface

2024-01-22T11:21:04+00:00

Interacting with machines via hand gestures is a common way for people to communicate with robots. Human utilize gestures in a regular talk to convey meaning and emotions to one another. Gesture-based interactions are utilized in a wide range of applied to a wide range of fields, as telephones, TVs, monitors, video games, and other electronic devices. By technological improvements, gesture recognition is now a more realistic and appealing approach in the context of human interaction. In this research, the relevant experiments are conducted using numerous types of convolutional neural networks, including the proposed customized model, to see which ones performs the best. Because of the introduction of such Microsoft Kinect sensor, increased depth and vision sensing has been widely important for several purposes. Given its ability to measure ranges to objects at a fast frame rate, these types of sensors are widely being employed for 3D acquisitions, as well as for other purposes in robotics and machine learning. This research made use of the Kinect sensor and the use of an RGB-D camera and a 3D convolution neural network, which offer a novel approach for fingertips identification and hand gesture classification in real time that is both accurate and fast (3DCNN).

Development of a Firmware for Multirotor UAV Flight Controller Implemented on MCU MDR 32

2024-01-22T11:21:05+00:00

The paper is devoted to study unmanned aerial vehicle (UAV) flight operational features and UAV’s automatic control system operating in manual mode (it processes pilot’s control commands). The analysis of operation in special cases and conditions is carried out for the aircraft equipped with a radio-operating human-controlled system and automatic flight control system. The design of UAV Euler's angles automatic stabilization system is described. The results of research and development of the flight controller based on the domestic made microcontroller unit (MCU) MDR 32 are presented.

Autonomous robot navigation system as part of a human-machine team based on self-organization of distributed neurocognitive architectures

2024-01-22T11:21:06+00:00

The relevance of this study is due to the solution of the problem of developing the basic principles and algorithms for providing adaptive settings for autonomous robots intelligent control systems as part of a human-machine team based on the general method of machine learning. To do this, the paper proposes to use a formalism based on multi-agent neurocognitive architectures. Implementation of the possibility of adaptation is considered on the example of performing the task of orientation and navigation of an autonomous robot in an unfamiliar environment.

An autonomous robot navigation system based on self-organization of distributed neurocognitive architectures has been developed.

A multi-agent neurocognitive architecture is presented, which forms an active map containing all the locative information necessary to ensure the orientation and navigation of an autonomous robot between loci.

The use of a multi-agent architecture to provide the representation of locative information in the task of implementing an interface in human-machine team will make it possible to build an ontology responsible for representing the location of objects in the external environment, as well as provide interaction with the user in natural language, taking into account its semantics. Further development of the approach is aimed at the implementation of the natural language interface of the human-machine team, which is not inferior in efficiency to the interaction between people.

Development of a robot for agricultural field scouting

2024-01-22T11:21:08+00:00

Since agricultural environments are mostly in unstructured feature information, in order to facilitate agricultural robots to better adapt to environmental change problems in agricultural environments, this paper proposes a robot system architecture for agricultural fields scouting. Firstly, the hardware architecture of the robot and its required sensors are considered, then the AMCL route planning algorithm is applied. The depth camera+lidar+RTK approach is used for the robot's map construction, the simulation model of robot route planning by ROS, the robot design by SolidWorks, and finally the required sensors and hardware structure parts are analyzed and summarized. Based on the considered project, it is planned to create a field agricultural scouting robot, which will become part of the implementation of a digital twin in crop production.

Relay Algorithm for Controlling Phase of Unbalanced Rotors of Electromechanical Plant

2024-01-22T11:21:09+00:00

The paper considers an complex mechatronic plant consisting of two induction motors connected to unbalanced rotors and installed on a common base. A relay control algorithm stabilizing the given phase between the rotors is proposed. The results of the influence of the relay controller coefficients on the performance quality and the achievement of the goal of control by means of modeling are shown.

Comparison of ROS Local Planners for a Holonomic Robot in Gazebo Simulator

2024-01-22T11:21:10+00:00

Safe robot navigation in a dynamic environment is an essential part of an autonomous exploration path planning. The path planning part of navigation involves global and local planners. While a global planner finds an optimal path with a prior knowledge of an environment and static obstacles, a local planner recalculates the path to avoid dynamic obstacles. The main goal of the local planning is adjusting an initial plan produced by the global planner in an online fashion. It is a crucial step to ensure a robot operation in dynamic environments because in real world scenarios an environment usually contains people and thus a dynamic obstacles avoidance must respond quickly and recalculate an actual route. Holonomic robotic platforms are robotic vehicles that use omni-wheels to move in any direction, at any angle, without an additional rotation. Because of a construction and a wheel design these robotic platforms are ideal for working zones with a limited space access. This paper provides a comparison of ROS local planners that support omni-wheel mobile robots: Trajectory Rollout, DWA, EBand, and TEB. The algorithms were compared using a path length, a travelling time and a number of obstacle collisions. Gazebo simulator was used for modeling virtual scenes with dynamic obstacles.

Evaluation of representativeness of EEG data for zonal affiliation of brain waves by leads

2024-01-22T11:21:12+00:00

The article highlights the results of scientific experimental research aimed at the evaluation of the representativeness of bioelectrical signals obtained by electroencephalography (EEG). The basic hypothesis is formulated and tested with the help of artificial neural network technology. The authors consider an experiment on the formation of steady-state visually evoked potentials in a group of people with the subsequent creation of an applied database. They describe an original approach for extracting representative features from the EEG signal. With the help of deep machine learning technology the representativeness of the data under study is evaluated. The main conclusions are formulated and the hypothesis that each brain lead reproduces unique waves which are characteristic of each brain zone is confirmed. The proposed model of a symmetric multilayer multi-adaptive direct propagation neuron, which provides functions of encoder and decoder, can find its application in solving problems related to the processing of EEG signals. Based on the results of this study, the authors suggest that the data on bioelectrical activity of the brain recorded by EEG can be represented as a multidimensional random variable, where the center position is also the mathematical expectation of its projections on the axes of the principal components.

DEVELOPING OF A DEVICE FOR POST-TRAUMATIC ANKLE REHABILITATION

2024-01-22T11:21:13+00:00

This article discusses a device for active-passive mechanotherapy of the ankle joint. The device is based on a controllable mobile platform equipped with force-moment sensors, on which the patient's foot is mounted by means of cuffs, and the platform rotation angles are controlled by linear motion sensors. The platform of the device is designed in such a way that the rotation axis of the platform always coincides with the centre of the ankle joint. For this purpose, a parallel kinematics mechanism is used, which is based on three linear electric drives. The control system of the device provides both active and passive movement of the platform. For realization of the control algorithm of the mobile platform movement, a mathematical model is developed, which allows establishing connections between angular motions of the mobile platform and linear drives of the parallel mechanism. Models of reaction forces of the platform support on the patient's foot during operation of the device are also described. A functional control diagram of the device is presented, and the modes of operation of the device are described.

Identification of the Quadcopter Rotational Dynamics for the Tilt Angle

2024-01-22T11:21:14+00:00

The pitch and roll angles of the quadcopter attitude are controlled by torque arising from the difference in thrust of different rotors. Rotational dynamics of the quadcopter as a rigid body is considered. The pitch and roll angles that form the quadcopter tilt angle have similar dynamics for the case of symmetric quadcopter frame, so the pitch angle is studied separately in the paper. The finite frequency identification approach is used to find the transfer function from the experimental flight data. The approach needs data, when sine wave test signals are fed to the system input. In the case of the unstable loop of the quadcopter tilt angle, the closed loop identification procedure is used where the test signal is added to the setpoint of the operating controller. As a result, a more complex model is identified than is commonly used. Moreover, the nonlinear model of the quadcopter pitch angle is considered, taking into account the translation velocity. Parameters of this model are estimated from the same experimental data using the nonlinear grey-box model parameters identification procedure. The transfer function founded from this model confirms the structure of the transfer function obtained via the finite frequency identification procedure.

Resource-Saving Multiobjective Task Distribution in the Fog- and Edge-Robotics

2024-01-22T11:21:15+00:00

In the current paper the question of the resource-saving tasks distribution is under consideration. As a wide range of computational tasks in robotics performs in a distributed manner in fog- and edge network layers, a lot of tasks are performed by devices with a relatively low computational complexity. At the same time, data preprocessing, machine learning, SLAM problems are computationally hard, and so the participants of the computational process can be overloaded, the latter causes the average residual life deterioration. In this paper a problem of resource-saving distribution task is formulated as structural-parametric multiobjective one, with paying attention to the workload of nodes, which transmit data. The general solution technique is proposed based on time constraints estimations and on the simulated annealing. Some selected experimental results are presented, and some comparisons with the previously conducted results are made.

The Design of Hybrid Control System for Nonaffine Plants

2024-01-22T11:21:16+00:00

In the theory of automatic control, an urgent problem is the development of methods to design nonaffine control systems. In such systems, the control affects the input of the object nonlinearly, so it affects the state variables in a non-additive way. The purpose of this article is to develop a design method that ensures the stability of the zero-equilibrium position of a closed nonaffine control system in a certain area. The objects described by a nonlinear system of differential equations with one control and one output are considered. A constraint is introduced, which consists in the differentiability of the right-hand sides of differential equations with respect to all state variables. The task of designing control in the form of a function of the setting action, a vector of state variables and control values at previous points in time is set. This problem is solved using a quasilinear model of the control object. Such model allows you to preserve all the features of the nonlinear equations of objects without simplifying them. In the quasilinear model, matrices and vectors are functions of the variables of the state of the control object. The control is found using an algebraic polynomial matrix method. This method allows you to find control under the condition of controllability of the object in the form of inequality. This article presents the calculation ratios for calculating the control in accordance with the polynomial-matrix method. Based on the given coefficients of the desired polynomial, as a result of solving an algebraic system of equations, coefficients are found that are a function of control and state variables. At the same time, the fulfillment of the controllability condition guarantees the existence of a solution of the specified algebraic system. An expression has been found that allows calculating the control by the coefficients found. The article also presents a condition for the possibility of providing a non-zero value of the output-controlled quantity of a nonlinear Hurwitz system in a steady-state mode. Under this condition, a zero value of the static error for the setting effect can also be provided. Further, the transformation of the obtained continuous control into a discrete one is proposed, which is implemented in a digital computer. The article also provides a numerical example of designing a control system for a nonaffine object of the second order, and the results of modeling a closed nonaffine system. The given example confirms the theoretical results obtained. Thus, the proposed approach makes it possible to design stable Hurwitz control systems for nonaffine plants using the algebraic polynomial matrix method with sufficiently small sampling periods of variables of the control plant and small modules of the roots of the characteristic polynomial of the matrix of a closed system in its quasilinear model.

Curl-free vector field for collision avoidance in a swarm of autonomous drones

2024-01-22T11:21:17+00:00

FIRAS function-based potential field method is the standard for collision avoidance as implemented in isolated drones. However, its use in an autonomous swarm can be problematic. Its complex interconnected structure causes one of the known issues when there are multiple simultaneously active control objectives. They are intra-swarm collision avoidance and reaching the target relative distance (normally referred to as formation control). This paper shows that with collision avoidance active, the standard potential field method will cause a local minima-like effect in an autonomous swarm. To prevent it, this paper proposes a modified curl-free vector field-based algorithm. This modification enables extended lateral circular motion to prevent swarm members from getting stuck in a local minimum. Stability theory methods are invoked to show that the formation remains stable when running the proposed algorithm. Comparative numerical experiments were run on a drone swarm model in MATLAB/Simulink to illustrate the functioning of this algorithm. To prove the proposed method effective, the paper presents simulation results for standard vs modified potential field.

Reliability Of Robot’s Controller Soft

2024-01-22T11:21:18+00:00

The reliability of robots’ digital control systems, based on Von Neumann type computer platform is investigated. It is shown that in systems of suchlike class the importance of software in ensuring the reliability of robot operation as a whole increases significantly. It is determined that software failure potential is being laid to the control program at its designing stage due to neglect of such digital controllers properties, as time delays when data processing. The model of control program failures emerging, caused by time factor, is worked out. For delays between transactions estimation the semi-Markov model of poling procedure is used, that permit to estimate probabilities of exceeding data skew, pure lag and sampling period the threshold, beyond which the control system failure takes place. With use the stochastic matrix, describing poling procedure, probability of failure, caused by transactions order disturbances is estimated. Probabilities and sampling period obtained are used for simulation of failure flow generator, describing reliability of control program.

An Improved Model of Greedy Tasks Assignment in Distributed Robotic Systems

2024-01-22T11:21:19+00:00

In the current paper the improved model for computational tasks distribution efficiency estimation is presented and discussed. Taking into account the tight connection between distributed robotic systems and IoT concepts, including fog and edge, the problem of computational resource spending is considered. As the failure rate of the node depends on the workload, we consider the strategy, when each node chooses to transmit or to process data. The model developed allows to implement the greedy strategy of tasks distribution, in which every node choses the best individual state. Also some selected experimental results are presented, pros and cons of such greedy approach are discussed.

STUDY OF PATH PLANNING METHODS IN TWO-DIMENSIONAL MAPPED ENVIRONMENTS

2024-01-22T11:21:19+00:00

The article studies the problem of motion planning in two-dimensional mapped environments. The review and analysis of known planning algorithms based on Voronoi diagrams, probabilistic road maps, rapidly growing random trees, Dijkstra algorithms, A*, D* and their modifications, artificial potential fields, and intelligent heuristics are carried out. Based on the analysis, it is concluded that classical methods in dynamic environments require significant costs in terms of calculation time and the amount of memory used. The conclusion is made about the relevance of the development of algorithms that increase the efficiency of known planning methods. In this regard, this article is devoted to the development of a modified algorithm for rapidly growing random trees and the study of its effectiveness in comparison with known methods. The article presents a modified algorithm for rapidly growing random trees, characterized in that when checking for a path to a new potential node of the tree, the path to some area near the specified node is checked. This reduces the number of nodes in the tree under construction. The developed algorithm is first compared with the traditional algorithm of fast-growing random trees. The comparison criteria are the trajectory calculation time, the amount of memory required, the path length, and the percentage of situations in which the trajectory to the target point was successfully found. Next, the developed algorithm is compared with the planning algorithms of other classes. The study uses representative samples of numerical experiments and various environments that differ in the density of obstacles and the presence of mazes. A study of planning algorithms using the results of experiments on a ground-based wheeled robot is also being conducted. Based on the results of numerical and real experiments, conclusions are drawn about the advantages and disadvantages of the developed algorithm of motion planning and the feasibility of its application in various environments.

Design and Implementation of a Multimodal Combination Framework for Robotic Grasping

2024-01-22T11:21:22+00:00

Robotic grasping plays a crucial role in manipulation tasks. However, due to the complexity of human-robot interaction, service robots still face significant challenges in handling task-oriented operations in real-world environments. To address this issue and better meet practical interaction needs, we propose a multimodal combination framework for robotic grasping. It leverages language texts to facilitate communication and detects and grasps target objects based on point clouds and feedback. The framework comprises several multimodal components, including ChatGPT, stereo cameras, and wearable devices, to complete instruction processing, grasp detection, and motion execution. To enable effective interaction, ChatGPT facilitates basic communication and responds to instructions between humans and robots. Additionally, the robot can detect the 6-DoF grasp of objects based on point clouds obtained by stereo cameras. These grasps are combined with the feedback provided by ChatGPT to further meet the requirement from human. Finally, we utilize wearable devices to teach robots generalized motor skills. This enables the robot to learn corresponding movements and perform them effectively in various scenarios, further improving its manipulation abilities. The experimental results from simulated conversations and real-scene tasks highlight that our proposed framework provides logical communication, stable grasping, and effective motion.

Monitoring the state of robotic systems based on time series analysis

2024-01-22T11:21:22+00:00

In view of the close integration of robotic systems into industrial and technological systems, critical infrastructure objects, as well as a significant number of possible entry points, the task of monitoring operational safety for robotic systems is more complex than ensuring information security in classical information systems. The paper presents a method for monitoring the state of robotic systems based on time series analysis. The developed method differs from the existing ones by the combined approach of using an ensemble of parallel classifiers and Fishburn weight coefficients in the security event management system. The time series is composed of a set of informative features, characterizing the functioning of a robotic system. Values for previous discrete time points are ranked using significance weights. The method was approved on a data set of a real industrial system. Due to parallel computing, it was possible to significantly increase the speed of determining the state of robotic systems. The identification precision due to the combined approach increased by 1.45 % compared to the best results presented in scientific papers, the recall increased by 4.45 % and amounted to 99.85 % for both indicators. The results of the study can be applied in monitoring the safety of robotic systems.