4.4: Path Smoothing

Last updated
Save as PDF

Page ID: 14791

\( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \) \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)\(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\) \(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\)\(\newcommand{\AA}{\unicode[.8,0]{x212B}}\)

As paths are randomly sampled, they will be most likely shaky and not optimal. For exampe, a grid-map will generate a series of sharp turns and a sampling-based approach will return zig-zag random paths. Results can be drastically improved by running an additional algorithm that smoothes the path. One way of doing this is to connect points of the path using splines, curves or even trajectory snippets that are known to be feasible for a specific platform. Alternatively, one can also use a model of the actual platform and use a feedback controller such as described in Section 3.5.1 for mobile robots and Section 3.5.2 for arms, sample a series of points in front of the robot, and generate a trajectory that the robot can actually drive. When combined with dynamics, this approach is known as model-predictive control. Care needs to be taken, however, that the resulting paths are indeed collision free.