ๆญฃๅœจๅŠ ่ฝฝ่ง†้ข‘...

่ง†้ข‘ๅŠ ่ฝฝๅคฑ่ดฅ

โ€œ๐—œ ๐˜„๐—ฎ๐˜€ ๐—ฏ๐—ผ๐—ฟ๐—ป ๐—ฑ๐—ถ๐—ณ๐—ณ๐—ฒ๐—ฟ๐—ฒ๐—ป๐˜. ๐—Ÿ๐—ผ๐—ผ๐—ธ๐—ฒ๐—ฑ ๐—ฑ๐—ถ๐—ณ๐—ณ๐—ฒ๐—ฟ๐—ฒ๐—ป๐˜. ๐—•๐˜‚๐—ถ๐—น๐˜ ๐—ฑ๐—ถ๐—ณ๐—ณ๐—ฒ๐—ฟ๐—ฒ๐—ป๐˜.โ€ Kicking off #BlackHistoryMonth with an icon of our sport This is Maggie Alphonsi MBEโ€™s incredible story

77,318 ๆฌก่ง‚็œ‹ โ€ข 2 ๅนดๅ‰ โ€ขvia X (Twitter)

0 ๆก่ฏ„่ฎบ

ๆš‚ๆ— ่ฏ„่ฎบ

ๅŽŸๅง‹ๅธ–ๅญ็š„่ฏ„่ฎบๅฐ†ๆ˜พ็คบๅœจ่ฟ™้‡Œ

็›ธๅ…ณ่ง†้ข‘

๐—›๐—ฒ๐—ฟ๐—ฒโ€™๐˜€ ๐—ฎ ๐—บ๐—ถ๐˜€๐˜๐—ฎ๐—ธ๐—ฒ ๐—œ ๐˜€๐—ฒ๐—ฒ ๐—ฎ๐—น๐—น ๐˜๐—ต๐—ฒ ๐˜๐—ถ๐—บ๐—ฒ. Teams collect robot data at 30Hz because โ€œthatโ€™s what our robot runs atโ€ and then wonder why their models underperform. The truth is that ๐—ฑ๐—ถ๐—ณ๐—ณ๐—ฒ๐—ฟ๐—ฒ๐—ป๐˜ ๐˜๐—ฎ๐˜€๐—ธ๐˜€ ๐—ป๐—ฒ๐—ฒ๐—ฑ ๐—ฑ๐—ถ๐—ณ๐—ณ๐—ฒ๐—ฟ๐—ฒ๐—ป๐˜ ๐˜๐—ฒ๐—บ๐—ฝ๐—ผ๐—ฟ๐—ฎ๐—น ๐—ฟ๐—ฒ๐˜€๐—ผ๐—น๐˜‚๐˜๐—ถ๐—ผ๐—ป๐˜€. Pick-and-place often works best around 10Hz for smooth motions. Dynamic catching requires 30Hz or more. Assembly tasks move slower, so 5Hz can suffice. Precision insertion demands 50Hz or higher for micro-adjustments. Hereโ€™s the catch. The ๐—ผ๐—ฝ๐˜๐—ถ๐—บ๐—ฎ๐—น ๐—ณ๐—ฟ๐—ฒ๐—พ๐˜‚๐—ฒ๐—ป๐—ฐ๐˜† ๐—ถ๐˜€๐—ปโ€™๐˜ ๐˜€๐—ผ๐—บ๐—ฒ๐˜๐—ต๐—ถ๐—ป๐—ด ๐˜†๐—ผ๐˜‚ ๐—ฐ๐—ฎ๐—ป ๐—ด๐˜‚๐—ฒ๐˜€๐˜€. The traditional approach wastes time and demos. You pick a frequency, collect thousands of demos, train a model, get mediocre results, and realize you have to start over. A better way is to ๐—ฐ๐—ผ๐—น๐—น๐—ฒ๐—ฐ๐˜ ๐—ฑ๐—ฎ๐˜๐—ฎ ๐—ฎ๐˜ ๐—ต๐—ถ๐—ด๐—ต ๐—ณ๐—ฟ๐—ฒ๐—พ๐˜‚๐—ฒ๐—ป๐—ฐ๐˜†, experiment with sync rates, and find what actually works for your task. Weโ€™ve seen grasping tasks perform best at 12Hz - not 10Hz, not 15Hz - discovered only through systematic testing. ๐—ฆ๐˜†๐—ป๐—ฐ๐—ต๐—ฟ๐—ผ๐—ป๐—ถ๐˜‡๐—ฎ๐˜๐—ถ๐—ผ๐—ป ๐—ณ๐—ฟ๐—ฒ๐—พ๐˜‚๐—ฒ๐—ป๐—ฐ๐˜† ๐—ถ๐˜€ ๐—ฎ ๐—ต๐˜†๐—ฝ๐—ฒ๐—ฟ๐—ฝ๐—ฎ๐—ฟ๐—ฎ๐—บ๐—ฒ๐˜๐—ฒ๐—ฟ. ๐—ง๐—ฟ๐—ฒ๐—ฎ๐˜ ๐—ถ๐˜ ๐—น๐—ถ๐—ธ๐—ฒ ๐—น๐—ฒ๐—ฎ๐—ฟ๐—ป๐—ถ๐—ป๐—ด ๐—ฟ๐—ฎ๐˜๐—ฒ ๐—ผ๐—ฟ ๐—ฏ๐—ฎ๐˜๐—ฐ๐—ต ๐˜€๐—ถ๐˜‡๐—ฒ. ๐——๐—ผ๐—ปโ€™๐˜ ๐—ฑ๐—ฒ๐—ฐ๐—ถ๐—ฑ๐—ฒ ๐—ถ๐˜ ๐—ผ๐—ป๐—ฐ๐—ฒ ๐—ฎ๐—ป๐—ฑ ๐—ฟ๐—ฒ๐—ด๐—ฟ๐—ฒ๐˜ ๐—ถ๐˜ ๐—ณ๐—ผ๐—ฟ๐—ฒ๐˜ƒ๐—ฒ๐—ฟ. How do you choose your data collection frequency, fixed upfront or experimental?

Stephen James

22,574 ๆฌก่ง‚็œ‹ โ€ข 9 ไธชๆœˆๅ‰