the first attempt
Well, let's say there is plenty of room for improvement.
Well, let's say there is plenty of room for improvement.
I’ve been a bit quiet over the past month. I took a break from punch recognition while I wait for a longer USB cable, so I can move around properly and train the system to recognize what different punches should look like.
In the meantime, I’ve been working on the firmware update process over Bluetooth. All sensors will come with a basic DFU bootloader with minimal BLE functionality. The app will then check the installed firmware version and update it automatically whenever necessary.
I spent some time developing the punch recognition algorithm on the sensors. The system can now reliably detect valid punches with approximately 90–95% accuracy. I integrated it into the existing BLE communication logic, but unfortunately the BLE connection is still unstable, with RSSI values remaining around -80 dBm or worse. Time to investigate the BLE side further.
I recently ran into another issue while testing the sensors. The RSSI is awful, averaging around -90 dBm. A week ago, I thought I had put this problem to bed, but unfortunately, that doesn’t seem to be the case. My developer partner—currently sleeping in the picture—believes we should go back to the drawing board, redesign the main loop, and process the data directly on the sensors rather than on the phone. It’s all part of the learning curve. I’m hoping to crack this one within a week.
| # | Username | Wins |
|---|---|---|
| 1 | KyleReese2045 | 49 |
| 2 | MartinRiggs1010 | 48 |
| 3 | PrisRunner5853 | 47 |
| 4 | JohnnyMnemonic5062 | 44 |
| 5 | MotokoKusanagi2759 | 43 |
| 6 | LeelooVector5616 | 39 |
| 7 | AeonPulse8008 | 39 |
| 8 | MaxRockatansky1334 | 37 |
| 9 | ViperZero6613 | 36 |
| 10 | AeonPulse1439 | 32 |