Time Series Augmentation for Sensory Systems

Time Series Augmentation for Sensory Systems

Bachelor Thesis, Master Thesis

Deep Learning performs remarkably well on large datasets that covers wide range of cases. However, the labeled data of many world time series applications may be limited. As an effective way to enhance the size and quality of the training data, we rely on data augmentation which showed a huge success in computer vision problems. However, less attention has been paid to find better data augmentation methods for time series.

With that, Automotive sensor data will be used as a case study. Hence, the objective of this work is to review and implement methods to realistically augment multivariate timeseries data collected from CAN bus of Tesla Model 3. The augmented data along with the original data should be able to increase the performance of the Neural Network models used in regression tasks.

Requirements

  • Knowledge in Deep Learning models
  • Experience in python programming language.
  • Familiar Tensorflow and Keras.
  • (Bonus) Familiar in frequency and signal analysis

Tasks

  • Explore approaches in Time/Frequency Domain to augment data.
  • Implement the promising methods.
  • Evaluate the results by running the augmented data along with the original data to train NN models.

Please send your application to the following people:

  • Stéphane Foulard ()
  • Ousama Esbel ()
  • Alejandro Sanchez Guinea ()