Acoustic Helicopter Classification
conference paper
This presentation describes the results of a research project which has been funded by the Dutch Ministry of Defence, School of Military lntelligence. During the project "acoustic helicopter classification" a number of algorithms to classify helicopters were developed. A number of techniques have been refined, i.e. neural network, harmonic series and template matching. The algorithrns are trained and tested on a database of 8 different helicopters (hovering and moving) recorded at distanccs ranging from 90m up to 8km (Measurement campaign AMI 1 and 2). To investigate the sensitivity to noise; jet, tank and artillery noise has been used as input. For target distances up to 2 km all algorithms perform well. At longer distances the performance decreases. Overall the neural network has the best performance. With a combination of the evaluated techniques the development of an operational system seems possible. Preferably however is the development of a demonstrator, which can be used to
optimize the performance for different operational applications. Future work will be carried out on the deterioration of the classification results under the influence of propagation effects and wind and environment noise. With an automatic mearsurement station data for a range of meteo parameters will be gathered. At a later stage helicopter data will be distorted by propagation effects and by measured noise, and subsequently fed to the classification algorithms. The results will give insight in the possible detection and classification ranges for an operational system.
optimize the performance for different operational applications. Future work will be carried out on the deterioration of the classification results under the influence of propagation effects and wind and environment noise. With an automatic mearsurement station data for a range of meteo parameters will be gathered. At a later stage helicopter data will be distorted by propagation effects and by measured noise, and subsequently fed to the classification algorithms. The results will give insight in the possible detection and classification ranges for an operational system.
TNO Identifier
94738
Publisher
National Research Council
Source title
Sixth International Symposium on Long-Range Sound Propagation - Proceedings of a Symposium held at the Chateau Laurier Hotel Ottawa, Canada, 12-14 June 1994
Editor(s)
Havelock, D.I.
Stinson, M.R.
Stinson, M.R.
Place of publication
Ottawa
Pages
108-114
Files
To receive the publication files, please send an e-mail request to TNO Repository.