Method and system for identification and mitigation of errors in non-line-of-sight distance estimation
Ultra-wide bandwidth (UWB) transmission is a promising technology for indoor localization due to its fine delay resolution and obstacle-penetration capabilities. However, the presence of walls and other obstacles present a significant challenge in terms of localization, as they result in positively biased distance estimates. Measurement campaigns with FCC-compliant UWB radios can quantify effects of non-line-of-sight (NLOS) propagation. Features of waveforms measured during a campaign can be extracted for use in distinguishing between NLOS and line-of-sight situations in embodiments of the present invention. Embodiments further include classification and regression methods based on machine learning that improve the localization performance while relying solely on the received signal. Applications for systems employing an example embodiment of the invention include indoor or outdoor search and recovery with high accuracy and low cost.
1. A method of characterizing a waveform path, by an electronic device, the method comprising:
accessing training data in a training database assembled from waveforms transmitted between at least one pair of training wireless devices, the training database including line-of-sight and non-line-of-sight labels and feature vectors representing aspects of the waveforms transmitted between the training wireless devices;
learning a function using the training data, the function being nonparametric with respect to features representing aspects of a waveform transmitted via the waveform path between a first wireless device and a second wireless device;
by the electronic device, characterizing the waveform path by evaluating the function learned with a feature vector extracted from the waveform received at the first wireless device from the second wireless device to make a decision regarding a presence of an obstruction between the first and second wireless devices or to determine a likelihood of an obstruction between the first and second wireless devices; and
reporting the determined decision or likelihood of an obstruction between the first and second wireless devices to a device accessible to a user, another wireless device, or a different physical or logical layer in a wireless network including multiple wireless devices.
2. The method as claimed in claim 1 wherein learning the function includes using machine learning techniques by performing support vector machine (SVM) or Gaussian processes.
3. The method as claimed in claim 1 wherein the function is chosen from a class of functions including: polynomials and linear combinations of kernel functions.
4. The method as claimed in claim 1 wherein the first and second wireless devices are matched pairs of wireless devices employing compatible transceivers.
5. The method as claimed in claim 1 wherein the features representing aspects of a waveform include: estimated distance; energy of a received signal; maximum amplitudes; rise time; mean excess delay; root-mean-square delay spread; and kurtosis.
6. An apparatus for characterizing a waveform path, the apparatus comprising:
a training database with training data, the training database being assembled from waveforms transmitted between at least one pair of wireless devices, the training database being configured to provide line-of-sight and non-line-of-sight labels and feature vectors representing aspects of waveforms transmitted between the training wireless devices to a learning unit;
the learning unit configured to learn a function using the training data, the function being nonparametric with respect to features representing aspects of a waveform transmitted via the waveform path between a first wireless device and a second wireless device;
an evaluating unit configured to characterize the waveform path by evaluating the function learned with a feature vector extracted from the waveform received at the first wireless device from the second wireless device to make a decision regarding a presence of an obstruction between the first and second wireless devices or to determine a likelihood of an obstruction between the first and second wireless devices; and
a reporting unit configured to report the decision or likelihood of an obstruction between the first and second wireless devices to a device accessible to a user, another wireless device, or a different physical or logical layer in a wireless network including multiple wireless devices.
7. The apparatus as claimed in claim 6 wherein the learning unit learns the function uses machine learning techniques including: support vector machine techniques and Gaussian processes techniques.
8. The apparatus as claimed in claim 6 wherein the function is chosen from a class of functions including: polynomials and linear combinations of kernel functions.
9. The apparatus as claimed in claim 6 wherein the first and second wireless devices are matched pairs of wireless devices employing compatible transceivers.
10. The apparatus as claimed in claim 6 wherein the features representing aspects of a waveform include: estimated distance; energy of a received signal; maximum amplitudes; rise time; mean excess delay; root-mean-square delay spread; and kurtosis.
11. The method as claimed in claim 1 further comprising positioning the training wireless devices in a model environment representative of a real-world environment to obtain the training data.
12. The apparatus as claimed in claim 6 wherein the training wireless devices are positioned in a model environment representative of a real-world environment to obtain the training data.
13. The method of claim 1 wherein characterizing the waveform path yields information about the waveform.
14. The apparatus of claim 6 wherein characterizing the waveform path yields information about the waveform.
15. The method of claim 1 wherein learning the function further includes choosing the function from a class of functions including polynomials and linear combinations of kernel functions.
16. The apparatus of claim 6 wherein learning the function further includes choosing the function from a class of functions including polynomials and linear combinations of kernel functions.