Abstract:
Path planning and trajectory generation for nonholonomic mobile robot in dynamic
environment is a difficult problem. It has to consider the nonholonomic constraints of robot
as well as the detection and avoidance of the possible collision with dynamic obstacles. In
this research an adaptive path planning technique has been proposed for mobile robot motion
in an environment with dynamic obstacles. A trajectory is formulated between the starting
and target positions for the three wheeled mobile robot. For this purpose kinematics and
dynamic models of the mobile robot have been used. Inverse kinematics model has been
developed by the wheel space technique for calculating the required wheels displacements to
reach the target location.
A collision-free motion of robot between its starting and goal positions is required to
be executed therefore, a safe global path which is a sequence of control points is planned,
incorporating the prior knowledge of environment. Artificial potential function has been
implemented for calculating intermediate target positions between the initial and final
position of mobile robot, to avoid obstacles in its path. A feasible global path is calculated for
vthe known environment and a time based trajectory is generated. This global path is set as a
reference path for the wheeled mobile robot. When dynamic obstacles moving with certain
velocity come in the path of mobile robot, it has to modify the path to avoid possible
collision.
An algorithm is proposed in this dissertation which efficiently detects and avoids the
possible collision with dynamic obstacles. This novel algorithm detects possible collision
with dynamic obstacle even if its velocity is not uniform. Expected collision of obstacle
coming in the path of mobile robot is detected by measuring the relative velocity of obstacle
and mobile robot at each time step. The reactive scheme of algorithm is activated for
changing the velocity of mobile robot to avoid the expected possible collision and the local
path planner plans a modified path to be executed by the mobile robot.
