Loading...

Path Planning and Control of Aerial-Terrestrial Swarm Robots Using Image Processing and Artificial Intelligence

Mirfakhar, Amin | 2023

71 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56299 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Alasti, Aria
  7. Abstract:
  8. Nowadays, there are various applications for swarm robots in different fields such as agriculture, healthcare, and entertainment. Among these, the field of agriculture is of greater importance due to the increasing population and the growing need for food resources. To address this need, industrialization of agriculture and the use of swarm robots have been proposed as a solution in most developed countries. Despite extensive research in this area, there are still challenges in implementing swarm robots. In many of the mentioned applications, determining the position and orientation of the robots relative to each other at long distances and accurately perceiving their surroundings are vital for executing control algorithms and enhancing safety. Additionally, seeking economic efficiency of the robots discourages the use of sensors such as RTK, GPS, and LIDAR. Therefore, the present study was conducted to address these needs. In this study, the combination of aerial and ground robots and the utilization of aerial imaging and image processing were attempted to identify the robots and ground obstacles. The ground robots were guided towards the goal through path planning, while their relative positions and orientations, along with obstacles, were determined using image processing and artificial intelligence. The outputs were then utilized for implementing control algorithms. Considering this, after providing an introduction to the subject, basic concepts in the field of swarm robots and image processing are presented. Subsequently, a literature review in this area is conducted, followed by the presentation of the kinematic and dynamic models of ground robots. Swarming and flocking algorithms, robotic chain formation, leader following, obstacle identification, and avoidance algorithms are examined through MATLAB and Python software simulations. In the next step, to execute these collective behaviors, they are developed for the existing swarm of robots in the laboratory, and the mentioned algorithms are implemented on the experimental swarm. The practical results obtained demonstrate the satisfactory performance of the algorithms, particularly in the case of stationary camera imaging
  9. Keywords:
  10. Image Processing ; Artificial Intelligence ; Swarm Robotics ; Collective Behavior ; Object Recognition ; Robot Pathplanning

 Digital Object List

 Bookmark

...see more