Loading...

Model Predictive Based Cooperative Search and Target Tracking With Dynamic Constraints

Moradi Pari, Hossein | 2013

456 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 45979 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Haeri, Mohammad
  7. Abstract:
  8. In this Thesis, a new algorithm is presented to cooperatively guide a set of robots, modeled as dynamic vehicles, toward a moving object while their heading angles variation are limited and collision avoidance is also required. The trajectory of object has unknown behavior, such that, at each moment in time, during the mission period, the available information about the object's trajectory is merely its position and instantaneous velocity. Due to this situation, vehicles are supposed to determine their movements, based on the predicted object's trajectory. To meet the goals of the algorithm, a time-decreasing reward is associated with the object, such that vehicles can collect the reward just if they visit the object. Model predictive control approach is employed in which the vehicles predict the objects position at a suitable horizon time and choose their directions via solving e sequential optimization problem to maximize their collected reward, which is the main objective of algorithm. In order to avoid the collision between the vehicles, we partitioned them into swarms. Each swarm contains the vehicles which are closed to each other and a leader that guide whole the swarm toward the object. Also, a new algorithm is presented to guide a set of robots, toward a moving object and dynamically decide whether to track the object or search in entire of the mission region for future object. The trajectory of object is unknown and vehicles are supposed to determine their movements, based on the predicted object trajectory. To meet the goals of the algorithm, a new cost function designed to satisfy the goals of the algorithm. Model predictive control has been used to provide this feature. Each robot's path planning would be based on solving a sequential optimization problem which determines that which robot follows the searcher and which robots take place at most probable points
  9. Keywords:
  10. Dynamic Target Tracking ; Cooperative System ; Cooperative Control ; Predictive Horizon Control ; Dynamic Decision Making

 Digital Object List

 Bookmark

No TOC