Introduction to Anomaly Detection for Engineers

Anomaly detection is the process of identifying events or patterns that differ from expected behavior. This is important for applications like predictive maintenance but can be hard to achieve by inspection alone. Machine learning and deep learning (AI) techniques for anomaly detection can uncover anomalies in time series or image data that would be otherwise hard to spot. Learn how and why to apply anomaly detection algorithms to identify anomalies in hardware sensor data.
Check out these other links:
- What Is Anomaly Detection?: bit.ly/3Re46SO
- What Is Automated Visual Inspection?: bit.ly/3fn3LQj
- Time Series Anomaly Detection Using Deep Learning (Example): bit.ly/3BFY6MS
- Want to see all the references in a nice, organized list? Check out this journey on Resourcium: bit.ly/3SrCI4Y
00:00 What is Anomaly Detection?
01:17 What is Anomaly Detection Used For?
03:10 How Anomaly Detection Works
03:47 Machine Learning Techniques for Time Series Data
05:00 Applying Autoencoders to Hardware for Anomaly Detection
08:55 Training and Testing Algorithms on Hardware
--------------------------------------------------------------------------------------------------------
Get a free product trial: goo.gl/ZHFb5u
Learn more about MATLAB: goo.gl/8QV7ZZ
Learn more about Simulink: goo.gl/nqnbLe
See what's new in MATLAB and Simulink: goo.gl/pgGtod
© 2022 The MathWorks, Inc. MATLAB and Simulink are registered trademarks of The MathWorks, Inc.
See www.mathworks.com/trademarks for a list of additional trademarks. Other product or brand names may be trademarks or registered trademarks of their respective holders.

[@rakshithb5806]

Another awesome Brian Douglas video and great demonstration! I really liked how we may not actually need failure data to train models for RUL estimation in predictive maintenance and how anamoly detection can do a pretty good job. Thinking of so many applications that can use this

[@trollgarten]

Wow, one of the best introduction into anomaly detections, very impressive video constructed with real content!

[@srodrmmabet6381]

Awesome explanation and with the example to illustrate the concepts 👏👏

[@posthocprior]

That was a great explanation.

[@zerddrez235]

Great video. Lots of thanks

[@Via.Dolorosa]

who simply explained, and very well demonstared

[@mukhtarsani9871]

Excellent work

[@standardio8270]

I really love the example it was really coot to see an example like this one.

[@thomasgamsjager7045]

Excellent!

[@sarette509]

Thank you so much sir for the great videos and the accurate informations you're providing. I'm a big fan! I have a suggestion for next videos: Can you talk about ML applications and approach in control theory? What are the limitations of control that favores an ML model? Thank u so much...

[@ruoxixi20]

super cool. Does it work for systems with high nonlinearity? Larger amount of data might be needed to capture nonlinear systems and there should be considerations to make sure the detector don't freak out under its acutal dynamics and acceptable amount of disturbance. Really cool topic!

[@dragolov]

Bravo!

[@dr.alikhudhair9414]

Wonderful

[@kwinvdv]

Have more standard estimators/observers from control theory, like a Luenberger observer, also been used for this? Because I can imagine that |y-yhat| (or maybe a low pass filter applied to that signal) might also be a good indicator of faults.

[@BrianBDouglas]

That’s a good point and I don’t know the answer. It seems like if the observer does a good job representing the nominal system then it could be used to flag anomalies.

[@Mrc93bpf]

Genius