Biosonar based mobile robot navigation

Echolocating bats can make nocturnal flights in acoustically cluttered environments with the use of echolocation. Their ability to evaluate targets in complete darkness provides mobile robots an opportunity to learn target detection, classification and identification with similar biomimetic platforms. In this work, natural landmark classification with a binaural system on a mobile robot, the purpose is to overcome some inherent shortcomings of airborne sonar and take advantages of bats’ perceived properties for mobile robots’ navigation in natural environments. Research now focuses on the topics of natural landmark classification with biosonar on mobile robot:

  • Stochastic feature based classification algorithms
  • 2D acoustic features from different sequential echolocation strategies
  • Clutter reflection model from acoustic volume reflection
The method we use consists of three basic steps:

First, the mobile robot  echolocates with certain sensing strategies with a biosonar head.



Biosonar head on mobile robot Robin

Then, the acquired echoes (random/sequential) are processed with various DSP and spatial correction algorithms to extract acoustic features.

At last the target is classified as a landmark for mobile robot navigation by matching algorithms.


Contact

Maosen Wang, Tel.: (07071) 29-77175, wangm@informatik.uni-tuebingen.de