Welcome to MIT Intelligent Anomaly Synthesiser! This website is for running test data remotely

Click here to enter your test parameters in the queue

The status of the GPASS system is publicly available via sensor and video online here


About Us

We are the MIT Mechatronics Research Laboratory (MRL). We are currently conducting research in health monitoring/anomaly detection on physical machines. We previously proposed a general purpose anomalous scenario synthesizer (GPASS) to generate various forms of reproducible, controlled dynamic anomalies to a rotary plant. For more technical details, refer to this paper. [ICRA paper link to be done]


        title = {The MIT MRL General Purpose Anomalous Scenario Data}, 

        author = {Yeung, Yip Fun and Alshehri, Ali and Wampler, Lois and Furokawa, Mikio and Hirano, Takayuki and Kim, Nicole and Paul-Ajuwape, Alex and Youcef-Toumi, Kamal}, 

        year = {2021}, 

        month = {January}, 

        note = {\url{ https://mit-anomalousdata.mit.edu/}} 


About the testbed

In the latest set up, the GPASS collects around 28 real-time sensor signals in multiple domains, e.g., angular displacement, velocity, acceleration, magnetic field, strain and so on. Most sensors run at a minimum of 80Hz. So, that’s a lot of data points.  


This is the stage of our setup. 



About the datasets 

Here, we published the data as we ran experiments locally for ourselves. The data are preliminarily pre-processed, such as fusing data streams from multiple sensors, synchronizing the time-frame, etc. However, the contents are intact. 

Good news is that we want to do more than just share the data. The GPASS works as a ‘cloud hardware platform’ that accepts ‘quotes’ for anomalous scenarios. If you want a characterized test run for your experiment, fill out the survey. The setup currently supports:

  1. Large lateral force (~20N)
  2. High frequency vibration (contact method: 50Hz, contactless method: regulated at 100Hz max)
  3. Grinding (force customizable)
  4. Scratching
  5. Rotational damping 

There is still more to come in the future! Feel free to contact us to develop a new anomalous mode together.