We had fantastic entries in the Speed Bag Detection Algorithm contest. The solutions to the mystery datasets were 172 and 155. Out of all the entries, we had seven people whose solutions came really close to figuring out the mystery datasets and solving the problem. My next step was taking their code into the gym.
In the end, barryjh’s solution nailed it. His solution got great results in the field (nearly 100 percent) and he had great documentation to boot! He used a front end processor (FEP), detection processor (DP) and an FIR filter designed with TFilter. You can read Barry Hannigan's amazing tutorial on his solution.
I highly encourage you to have a look at the code in barry's repo. It's a complex solution to a complex problem with excellent results.
I want to give a special shout out to all the contestants!
- Patrickk (I'm working my way toward 4,294,967,295)
- gacrmonar (Your code was fully integrated - thanks!)
- marciopp (award for the simplest implementation)
- 821180
- UofI_1996 (Compiled first try! Well done!)
- Toad
- 284237 (award for best use of a Butterworth filter)
- waltemus
- 399747
- floppystick6
- Ben A
I learned a ton from your solutions, and I know other readers will for many years to come.
That's some amazing work. I wish my data collection/analysis was that complex in my projects, might be time to learn something new!