COVERAGE CONTROL TO ESTIMATE DENSITY FUNCTION MAP BY MULTI HEXACOPTER WITH COLLISION AVOIDANCE ABILITY

• Main Author: Aris Prabowo, Yaqub
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/35760
• Institution: Institut Teknologi Bandung

In this thesis, research is conducted whereas multi hexacopters can do coverage control so that they can estimate a density function in environment with known obstacles. Gaussian Process Regression (GPR) algorithm is used to estimate the time-invariant density function. Hexacopters are doing exploration or exploitation in the environment where the density function is not known based on the centroid of Voronoi partitions constructed. Exploration or exploitation are decided based on the maximum posteriori variance calculated using GPR. All of hexacopters use Hybrid Reciprocal Velocity Obstacle (HRVO) as the local planning algorithm combined with Theta* as the global planning algorithm so that the hexacopters can do the mission without having a collision. The existence of obstacles may create non-convex Voronoi polygon so that the centroid may be located inside the obstacle. Consequently, the location point must be moved optimally to the outside obstacles. This research is done with Software In the Loop Simulation (SILS) scheme in Robot Operating System (ROS) with autopilot system provided by Pixhawk PX4 and using mathematic model provided by Gazebo. The coverage strategy is also modified in this research to increase the time performance. The result of this research is the comparison of different strategies, obstacle locations, and number of hexacopters. At the end, with the simulation based on ROS, hexacopters successfully map the density function with the metrix are small maximum variance that is 0,2 with time spent 656.7 s.