Online smooth trajectory planning for mobile robots by means of nonlinear filters

The paper presents a nonlinear filtering technique that can be adopted to generate smooth trajectories for mobile robotic applications. The proposed trajectory planner can be fully executed online by the robot control system, thanks to its inherently discrete-time behavior and to its limited computational requirement. The outputs of the nonlinear filter used as a trajectory planner include the derivatives of the desired position in the cartesian plane up to the third order. This allows the implementation of feedback linearization control schemes that can transform the dynamics of a mobile robot in a double chain of three integrators, exploiting the highest derivative of the filter's output as a feedforward action. Finally, the paper reports experimental results obtained by the full implementation of the proposed trajectory planning and control scheme on a real unicycle-like robot.