Materialien zur Vorlesung Mobile Roboter 

Eine Sammlung der in der Vorlesung gezeigten Videos und Animationen:

Localization

Monte Carlo Sampling with 1000 samples

The video shows the estimated distribution of a differential drive robot. The robot drives on the green trajectory and uses Monte Carlo Sampling with 1000 samples to estimate its pose.



Particle Filter with 100 samples

Again the robot follows the green trajectory. The robot is additionally equipped with a sensor which measures the distance and the angle to the green landmark. The odometry and the measurement is merged by a Particle Filter with 100 particles.



Particle Filter 100 samples with probabilities

The same example as above. The background visualizes the probability function of the measurement. It correlates to the weight a particle gets.

 


Navigation

Convex Hull - QuickHull

click on the image to advance the animation, or select a frame below

Javascript needed



Convex Hull - Sklansky

click on the image to advance the animation, or select a frame below

Javascript needed



Distance Transformation Algorithm - Dynamic Programming

  • Search Space dimensionality: 2
  • Motion Model: Omnidirectional


Distance Transformation Algorithm - Dynamic Programming - WORSTCASE

  • Search Space dimensionality: 2
  • Motion Model: Omnidirectional


Breadth-first Search

  • Search Space dimensionality: 2
  • Motion Model: Omnidirectional


Dijkstra

  • Search Space dimensionality: 2
  • Motion Model: Omnidirectional


Dijkstra

  • Search Space dimensionality: 4
  • Motion Model: Nonholonomic (car-like)


A* (using euclidean norm as heuristic)

  • Search Space dimensionality: 2
  • Motion Model: Omnidirectional
  • Heuristic: Euclidean norm


A*

  • Search Space dimensionality: 2
  • Motion Model: Omnidirectional
  • Heuristic: Chessboard distance


A*

  • Search Space dimensionality: 4
  • Motion Model: Nonholonomic (car-like)
  • Heuristic: Euclidean distance


A*

  • Search Space dimensionality: 4
  • Motion Model: Nonholonomic (car-like)
  • Heuristic: Optimal 2d heuristic, computed by dynamic programming in two dimensions


A* - Ignoring end orientation

  • Search Space dimensionality: 4
  • Motion Model: Nonholonomic (car-like)
  • Heuristic: Euclidean distance