Synthetic data generation for classification via uni-modal cluster interpolation
The observations used to classify data from real systems often vary as a result of changing operating conditions (e.g. velocity, load, temperature, etc.). Hence, to create accurate classification algorithms for these systems, observations from a large number of operating conditions must be used in algorithm training. This can be an arduous, expensive, and even dangerous task. Treating an operating condition as an inherently metric continuous variable (e.g. velocity, load or temperature) and recognizing that observations at a single operating condition can be viewed as a data cluster enables formulation of interpolation techniques. This paper presents a method that uses data clusters at operating conditions where data has been collected to estimate data clusters at other operating conditions, enabling classification. The mathematical tools that are key to the proposed data cluster interpolation method are Catmull–Rom splines, the Schur decomposition, singular value decomposition, and a special matrix interpolation function. The ability of this method to accurately estimate distribution, orientation and location in the feature space is then shown through three benchmark problems involving 2D feature vectors. The proposed method is applied to empirical data involving vibration-based terrain classification for an autonomous robot using a feature vector of dimension 300, to show that these estimated data clusters are more effective for classification purposes than known data clusters that correspond to different operating conditions. Ultimately, it is concluded that although collecting real data is ideal, these estimated data clusters can improve classification accuracy when it is inconvenient or difficult to collect additional data.
KeywordsInterpolation Singular value decomposition Terrain classification Data clusters Pattern classification
This work was prepared through collaborative participation in the Robotics Consortium which is sponsored by the U. S. Army Research Laboratory under the Collaborative Technology Alliance Program, Cooperative Agreement DAAD 19-01-2- 0012. The U. S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation thereon. Funding for this research also provided by the National Science Foundation, Project EEC-0540865.
- Bartels, R. H., Beatty, J. C., & Barsky B. A. (1998). Hermite and cubic spline interpolation. In An introduction to splines for use in computer graphics and geometric modeling (pp. 9–17). San Francisco, CA: Morgan Kaufmann.Google Scholar
- Collins, E. G. Jr., & Coyle, E. J. (2008). Vibration-based terrain classification using surface profile input frequency responses. In Proceedings of IEEE International Conference on Robotics and Automation. Pasadena, California.Google Scholar
- Coyle, E. J., Collins, E. G. Jr., & Roberts, R. G. (2011). Speed independent terrain classification using singular value decomposition interpolation. In Proceedings of IEEE International Conference on Robotics and Automation. Shanghai, China. (submitted for publication).Google Scholar
- DuPont, E. M., Collins, E. G, Jr., Coyle, E. J., & Roberts, R. G. (2008a). Terrain classification using vibration sensors: Theory and methods. In E. V. Gaines & L. W. Peskov (Eds.), New Research on Mobile Robotics. Hauppauge, NY: Nova.Google Scholar
- DuPont, E. M., Moore, C. A., Collins, E. G, Jr, & Coyle, E. J. (2008b). Frequency response method for online terrain identification in unmanned ground vehicles. Autonomous Robots, 24(4), 337–347.Google Scholar
- DuPont, E. M., Moore, C. A., & Roberts, R. G. (2008c). Terrain classification for mobile robots traveling at various speeds an eigenspace manifold approach. In Proceedings of IEEE International Conference on Robotics and Automation. Pasadena, California.Google Scholar
- Feiveson, A. H. (1966). The generation of a random sample-covariance matrix. Technical, Report NASA-TN-D-3207, NASA.Google Scholar
- Rasmussen, C. E., & Williams, C. K. I. (2005). Gaussian processes for machine learning (Adaptive computation and machine learning). Cambridge, MA: The MIT Press.Google Scholar
- Ward, C. C., & Iagnemma, K. (2008). Speed-independent vibration-based terrain classification for passenger vehicles. Vehicle System Dynamics, 00, 1–19.Google Scholar
- Weiss, C., Fröhlich, H., & Zell, A. (2006). Vibration-based terrain classification using support vector machines. In Proceedings of the International Conference on Intelligent Robots and Systems. Beijing, China.Google Scholar
- Yuan, Q., Thangali A., Ablavsky V., & Sclaroff S. (2007). Parameter sensitive detectors. In Computer Vision and Pattern Recognition (CVPR) pp. 1–6.Google Scholar