Waypoint-Based Generation of Guided and Optimal Trajectories for Autonomous Tracking Using a Quadrotor UAV

Unmanned aerial vehicles (UAVs) have extensively been investigated by the researchers over the past and recent decades. Specifically, a huge body of research has been directed towards quadrotors with huge contributions towards its design and control [4, 9, 13]. Currently, major research focus is on intelligent swarm control and aggressive maneuvering with quadrotor UAVs [7, 12, 15, 16]. To perform such tasks, generation of smooth and dynamically feasible trajectories is inevitable [6, 8, 17]. Smoothness can be defined as a function being continuous over time along with its continuous first-order derivative. Sometimes, a second derivative having continuity over time is also desirable for a specific dynamical system. Rough trajectories cause an increased burden on the robot dynamics and its control system by exciting undesirable vibrations. Any smooth function of time can be used to characterize a path and the shape of the path depends on the type of the chosen function. Additional constraints can also be applied to these functions for further smoothness and to attain the desired characteristics [3, 11]. Characteristically, trajectory is a time history that holds information regarding position, velocity, acceleration, and orientation of the robot. The user does not need to specify the complicated functions of time and space for describing the trajectories. This work is usually done by the robot online and trajectories are specified by the users in simple descriptions i.e., initial and final positions of the robot or waypoints. Furthermore, other spatial constraints like the elapsed time, velocity, and acceleration profiles are also considered equally by the robot in order to bring smoothness and desired attributes in the trajectory to be generated. Differential flatness is an important theory which helps in generating the dynamically feasible spatial trajectories subject to the proof that output states of the dynamical system are also differentially flat or simply flat [11, 12]. Flatness offers a way of decomposing the desired UAV trajectory into a chain of flat outputs and their derivatives. This further enables the calculation of controller commands to follow the computed trajectory [14]. Studies in Informatics and Control, 27(2) 225-236, June 2018