مهندسی مکانیک شریف (May 2018)
A BEHAVIOR-BASED APPROACH TO SIMULTANEOUS REALIZATION OF LEADER-FOLLOWING AND OBSTACLE-AVOIDANCE BEHAVIOURS FOR A FLYING ROBOT
Abstract
With the purpose of performing successful realization of multi-robot missions, an accurate guidance and control system is needed to enable the robots to simultaneously follow a leader and avoid obstacles. In this research, a modular guidance and control system is proposed, which considers the dynamics of quad-copters to perform flight behaviors formation. Hovering in the air and the higher speed of aerial robots in comparison to ground robots makes obstacle avoidance a crucial function for their flight. Furthermore, these robots are employed frequently in urban environments where there are numerous obstacles. A multi-agent mission is a virtuous notion that improves the effectiveness and range of aerial robots in their applications. As a result, robots have to get far from their reference to be able to avoid the obstacle. A guidance system is ideal for these robots, which can maintain their path close enough to the original reference and at the same time avoiding obstacles safely. In this paper, the behavioral algorithm is utilized to simultaneously realize the leader following and obstacle avoidance behaviors in order to complete the formation flight successfully. Control system has to be consistent with the produced reference of the guidance system. Moreover, the main challenge in designing the control system is the underactuated-ness of quadrotors; this dynamical concept means that quadrotor has limited inputs and is not able to straightforwardly track three-dimensional trajectories. In this work, the control system is designed based on dynamic inversion approach and is consistent with the guidance system. Performance of guidance and control systems is examined in a detailed simulation environment, which enjoys the incorporation of the noise of sensors and lag of rotors. For this purpose, the flight dynamics of quadrotors is carefully modeled to make simulation outcomes more realistic. The simulation results specify that the guidance and control systems are capable to follow the leader while attaining 2 m accuracy in position and 3.5 m in altitude.
Keywords
