Search

This paper considers active localization of a stationary target using time delay or together with Doppler shift measurements obtained by moving sensors. Non-negligible sensor motion effect occurs in the localization system when the dynamic monostatic sensors send and receive signals during its motion. The proposed new accurate models for time delay and Doppler shift are able to make up the sensor motion effect. Closed-form solutions when ignoring the sensor motion of the proposed models are derived to give an approximate position estimate. The solutions are further refined by the proposed models and Taylor series linearization to remove the effect caused by sensor motion. The simulations show that the refined solutions can increase the localization performance efficiently and the accuracy can reach the Cramér-Rao lower bound (CRLB). [ABSTRACT FROM AUTHOR]

Robust localization system has proven to be a cornerstone for mobile robot autonomy. Although passive robot localization is a mature field, it still could fail in challenging environments containing symmetries or open spaces. Active localization can fix this issue by allowing the robot to improve pose estimation by choosing specific actions. We propose an active localization strategy for the indoor position tracking problem in challenging environments. The proposed active localization is performed in three steps: (i) cluster the particle cloud with Spectral Clustering (or Kmeans + + ) algorithm, (ii) search and select the most informative point in a reduced search space, and (iii) execute rotational actions in order to sense the selected point. Hence, a significant number of wrong hypotheses are pruned. We also introduce a novel study that considers evaluates points in spatial neighborhoods all at once, instead of evaluating each cell independently. Simulated experiments show an improvement in robot pose estimation using the proposed strategy. Real-world validation in symmetric and open office-like environment is also presented. [ABSTRACT FROM AUTHOR]

In static scenarios, binaural sound localization is fundamentally limited by front-back ambiguity and distance non-observability. Over the past few years, “active” schemes have been shown to overcome these shortcomings, by combining spatial binaural cues with the motor commands of the sensor. In this context, given a Gaussian prior on the relative position to a source, this paper determines an admissible motion of a binaural head which leads, on average, to the one-step-ahead most informative audio-motor localization. To this aim, a constrained optimization problem is set up, which consists in maximizing the entropy of the next predicted measurement probability density function over a cylindric admissible set. The method is appraised through geometrical arguments, and validated in simulations and on real-life robotic experiments. [ABSTRACT FROM AUTHOR]