Academia.eduAcademia.edu

Figure 12 – uploaded by IAES Journals

See full PDFdownloadDownload figure
Figure 13. GAZEBO Scene Simulator _ Connectivity, rayload, and Autonomy conditions are satisfactory for ali units within the Cluster. The simulation's outcomes are presented in the subsequent Figures 11 to 15 and Tables | to 8. The plan involves finding the time to reach each victim, pinpointing the longest time to reach two victims, and averaging the maximum values. We focus on the maximum and minimum average durations. The regions in Figure 15 show designated areas in RGB colors within the All Area. These areas come from segmenting the total ROI using the greedy algorithm. The process is repeated three times. The first iteration involves a single robot over ten attempts, and the results are in Table 1. Where time distribution by number of attempts is given (with trend function) in the next chart as shown in Figure 16.

Figure 13 GAZEBO Scene Simulator _ Connectivity, rayload, and Autonomy conditions are satisfactory for ali units within the Cluster. The simulation's outcomes are presented in the subsequent Figures 11 to 15 and Tables | to 8. The plan involves finding the time to reach each victim, pinpointing the longest time to reach two victims, and averaging the maximum values. We focus on the maximum and minimum average durations. The regions in Figure 15 show designated areas in RGB colors within the All Area. These areas come from segmenting the total ROI using the greedy algorithm. The process is repeated three times. The first iteration involves a single robot over ten attempts, and the results are in Table 1. Where time distribution by number of attempts is given (with trend function) in the next chart as shown in Figure 16.

Related Figures (17)

Figure 1. Nominal current discharge characteristic (0.65A) it to make decisions for both collaborative and individual purposes. v) Every agent unit within the cluster i outfitted with essential equipment, including a Camera, LiDAR (light detection and ranging), a motor drivet actuators, and sensors as shown in Figure 2. vi) The micro-robot group operates in two modes: Supervise: mode and collaboration mode.
Figure 1. Nominal current discharge characteristic (0.65A) it to make decisions for both collaborative and individual purposes. v) Every agent unit within the cluster i outfitted with essential equipment, including a Camera, LiDAR (light detection and ranging), a motor drivet actuators, and sensors as shown in Figure 2. vi) The micro-robot group operates in two modes: Supervise: mode and collaboration mode.
Figure 3. ROS basics and concepts We aim to effectively locate and aid victims in challenging situations through Collaborative MRTA. Our approach involves the implementation of 3D digitization utilizing LiDAR or a depth camera, ensuring optimal outcomes for Search and simultaneous localization and mapping (SLAM) purposes. The ROS environment serves to manage the context, and, in addition to interoperability, ROS can engage with various platforms dedicated to simulating MRTA problems [26]-[30]. The ROS.MSG embedded module facilitates message exchange among micro-robots, with inertial measurement unit (IMU) data conversion to ODOM (odometry) enabling the validation and voting of agent moves in areas that are still unoccupied as shown in Figure 3.
Figure 3. ROS basics and concepts We aim to effectively locate and aid victims in challenging situations through Collaborative MRTA. Our approach involves the implementation of 3D digitization utilizing LiDAR or a depth camera, ensuring optimal outcomes for Search and simultaneous localization and mapping (SLAM) purposes. The ROS environment serves to manage the context, and, in addition to interoperability, ROS can engage with various platforms dedicated to simulating MRTA problems [26]-[30]. The ROS.MSG embedded module facilitates message exchange among micro-robots, with inertial measurement unit (IMU) data conversion to ODOM (odometry) enabling the validation and voting of agent moves in areas that are still unoccupied as shown in Figure 3.
Figure 2. AL Mustaksheef3D, wheeled robot developed
Figure 2. AL Mustaksheef3D, wheeled robot developed
Figure 4. Cluster architecture
Figure 4. Cluster architecture
Figure 5. N-gon of a scanned area
Figure 5. N-gon of a scanned area
Figure 6. Potential scenarios for the robot's posture (a) simple case with 3 robots, (b) complex case with 3 robots, (c) singular case, and (d) insoluble case
Figure 6. Potential scenarios for the robot's posture (a) simple case with 3 robots, (b) complex case with 3 robots, (c) singular case, and (d) insoluble case
Figure 7. Flowchart of our approach The weights k; for i€{1,4] symbolize the contributions of each specific part to the agent's functional cost, as defined earlier. Our goal is to discover the optimal outcome (10) for the function fos; for every agent in the cluster, representing the best solution. Subsequently,
Figure 7. Flowchart of our approach The weights k; for i€{1,4] symbolize the contributions of each specific part to the agent's functional cost, as defined earlier. Our goal is to discover the optimal outcome (10) for the function fos; for every agent in the cluster, representing the best solution. Subsequently,
Figure 10. Scene with target location Figure 9. Model area of the simulation The simulation scenario is illustrated in Figure 9. It is characterized by various constraints that evaluate the efficiency and reliability of our approach in navigating a cluster of micro-robots, both with and without collaboration. The simulation scene consists of rooms (isolated spaces) containing furniture (obstacles). Micro-robots work together to rescue multiple victims as shown in Figure 10, to minimize time and enhance reliability. The simulation illustrates three severity levels involving 1-3 micro-robots as shown in Figure 11.
Figure 10. Scene with target location Figure 9. Model area of the simulation The simulation scenario is illustrated in Figure 9. It is characterized by various constraints that evaluate the efficiency and reliability of our approach in navigating a cluster of micro-robots, both with and without collaboration. The simulation scene consists of rooms (isolated spaces) containing furniture (obstacles). Micro-robots work together to rescue multiple victims as shown in Figure 10, to minimize time and enhance reliability. The simulation illustrates three severity levels involving 1-3 micro-robots as shown in Figure 11.
Figure 11. Space segmentation
Figure 11. Space segmentation
Figure 8. JetBot during domain exploration
Figure 8. JetBot during domain exploration
Figure 14. Robots while performing a mission Figure 12. The used robot model
Figure 14. Robots while performing a mission Figure 12. The used robot model
Figure 16. Duration taken by a single robot to reach the target on the initial attempt Figure 15. Agents ROI updates
Figure 16. Duration taken by a single robot to reach the target on the initial attempt Figure 15. Agents ROI updates
Table 4. Two robots with collaboration algorithm averaoce time (cec) Table 6. Three robots with collaboration algorithm average time (sec)
Table 4. Two robots with collaboration algorithm averaoce time (cec) Table 6. Three robots with collaboration algorithm average time (sec)
Table 3. Two robots without collaboration algorithm @iveawRawer Tamia Cae ¥ Table 5. Three robots without collaboration algorithm average time (sec)
Table 3. Two robots without collaboration algorithm @iveawRawer Tamia Cae ¥ Table 5. Three robots without collaboration algorithm average time (sec)
Figure 18. Max and min time duration by experiment until the victim is found Figure 17. Progression of time duration through experimentation The ROI was evaluated in two ways: by pinpointing the targets and by 3D scanning the area for potential reconnaissance. This was done using an enhanced version of a horizontal LiDAR (RPLiDAR A1MB8) as part of the new features on our AL Moustaksheef3D platform, a robotic payload for 3D SLAM currently in development. The enhanced LiDAR as shown in Figures 19 and 20 was tested on a wheeled machine and a Drone. The collected data as shown in Figure 21 is provided in point cloud (PCL) of “ped” format. This model integrates two LiDARs, with the first used for geolocation and the second for 3D vision. Scene reproduction yields relative data for estimating facts in the study area. Point-cloud data is used for scene reconstruction. Software like CloudCompare, MeshLab, Blender, and Gimp enable background processing on point cloud data. The micro-cobot agents integrate APIs into their embedded logic, employing robust algorithms for identifying specific targets in hostile environments. Using an MRTA approach with the enhanced Greedy algorithm, our strategy optimizes target identification through cluster age nt cooperation, significantly reducing task time. This collaborative approach requires substantial processor capabilities, with basic calculations centralized for collaboration and decentralized for independent agent decision-making post-task assignment.
Figure 18. Max and min time duration by experiment until the victim is found Figure 17. Progression of time duration through experimentation The ROI was evaluated in two ways: by pinpointing the targets and by 3D scanning the area for potential reconnaissance. This was done using an enhanced version of a horizontal LiDAR (RPLiDAR A1MB8) as part of the new features on our AL Moustaksheef3D platform, a robotic payload for 3D SLAM currently in development. The enhanced LiDAR as shown in Figures 19 and 20 was tested on a wheeled machine and a Drone. The collected data as shown in Figure 21 is provided in point cloud (PCL) of “ped” format. This model integrates two LiDARs, with the first used for geolocation and the second for 3D vision. Scene reproduction yields relative data for estimating facts in the study area. Point-cloud data is used for scene reconstruction. Software like CloudCompare, MeshLab, Blender, and Gimp enable background processing on point cloud data. The micro-cobot agents integrate APIs into their embedded logic, employing robust algorithms for identifying specific targets in hostile environments. Using an MRTA approach with the enhanced Greedy algorithm, our strategy optimizes target identification through cluster age nt cooperation, significantly reducing task time. This collaborative approach requires substantial processor capabilities, with basic calculations centralized for collaboration and decentralized for independent agent decision-making post-task assignment.
Related topics:
Multirobot systemsRobot operating system (ROS)3D DigitizationSimultaneous Localization and Mapping (SLAM)Light Detection and Ranging SystemsMulti-Robot Task Allocation
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved