Hang Yan
Hang Yan
  • Видео 26
  • Просмотров 21 372
RIDI: Robust IMU Double Integration
The demo video for our work on IMU double integration. Please refer to
yanhangpublic.github.io/ridi/index.html
for the paper, code and data.
Просмотров: 17 916

Видео

Turning an Urban Scene Video into a Cinemagraph - Demo Video
Просмотров 2,5 тыс.8 лет назад
My latest research. Paper link: arxiv.org/abs/1612.01235
Modeling static scenes with dynamic appearance for time-depend rendering
Просмотров 488 лет назад
Demo video made for CVPR 16 submission. Interesting part starts from 1:17
sample rescale7
Просмотров 108 лет назад
sample rescale7
sample rescale7 mp4 render 0 20
Просмотров 288 лет назад
sample rescale7 mp4 render 0 20
result shaky
Просмотров 488 лет назад
result shaky
result SANY0025
Просмотров 638 лет назад
result SANY0025
result panorama
Просмотров 358 лет назад
result panorama
result campus1
Просмотров 868 лет назад
result campus1
result lab
Просмотров 348 лет назад
result lab
lab
Просмотров 268 лет назад
lab
0073YC
Просмотров 288 лет назад
0073YC
panorama
Просмотров 168 лет назад
panorama
shaky
Просмотров 418 лет назад
shaky
18AF
Просмотров 478 лет назад
18AF
result 00006 1
Просмотров 368 лет назад
result 00006 1
00006 3dfail
Просмотров 348 лет назад
00006 3dfail
result 18AF
Просмотров 488 лет назад
result 18AF
0073YC 3dfail
Просмотров 298 лет назад
0073YC 3dfail
result 0073YC
Просмотров 288 лет назад
result 0073YC
campus1
Просмотров 1018 лет назад
campus1
00006 1
Просмотров 298 лет назад
00006 1
result campus2
Просмотров 268 лет назад
result campus2
campus2
Просмотров 278 лет назад
campus2
New 18AF ours wnd50
Просмотров 398 лет назад
New 18AF ours wnd50
SANY0025
Просмотров 468 лет назад
SANY0025

Комментарии

  • @basilavad
    @basilavad 2 года назад

    Hi, What would be the most appropriate method for calculating the position of an object using linear acceleration data from a BNO055 sensor, given the potential presence of noise and errors in the data? Additionally, what techniques or methods can be employed to mitigate these issues and improve the accuracy of the position calculation?

    • @JackOHaraEngineering
      @JackOHaraEngineering 3 месяца назад

      Lookup kalman filtering, sensor fusion, a complimentary filter perhaps. Bno055s calibrate themselves which is convenient but terrible. The drift has to be corrected, and even when corrected at standstill the bias changes over time. I’ve been trying to work this same thing, and my conclusion is that getting more than a few seconds of accurate data without other sensors is tough to impossible. Adding a gps and barometer in tandem seems to be the way. Edit: also make sure your polling speed is pretty high and personally I use doubles as my data type because floats don’t have as much precision, but this may be fruitless for the Euler angles anyways. Sorry for the rant lol, you’re not alone

  • @jackhanson1274
    @jackhanson1274 2 года назад

    Fantastic stuff!

  • @lrwerewolf
    @lrwerewolf 2 года назад

    What do you think might happen if you used multiple IMUs arranged in such a way that no IMU had parallel/co-planar planes to the others? Would the extra ability to isolate noise by calculating the virtual IMU help clean up the signal even more?

    • @jackhanson1852
      @jackhanson1852 2 года назад

      How would you remove noise in your example? Do you have a paper that details this technique that you could recommend? I'm using IMUs for pedestrian tracking and haven't come across two (or more) being used in this way. Interested to hear what you have to say!

  • @hamzasadik7521
    @hamzasadik7521 2 года назад

    Hi ,i wish you are good please would uu send me the data because there aren't anymore on the site

    • @jackhanson1852
      @jackhanson1852 2 года назад

      There is improved data for their work under a project called RoNIN!

  • @patmw
    @patmw 3 года назад

    Awesome work!!

  • @dariuszmaton5375
    @dariuszmaton5375 3 года назад

    Interesting, could directly integrating the regressed velocities (orange line) also work? (Edit: just read it in the paper, "Direct integration of the predicted velocities would produce positions but performs worse.")

  • @bigto925
    @bigto925 4 года назад

    great work

  • @steven-bt7ud
    @steven-bt7ud 4 года назад

    Can't wait for the next improvement 👍, hope its small enough to apply this in an arduino for vehicle tracking

  • @Biru_to
    @Biru_to 4 года назад

    Why blur the face of the person (I'm guessing it's the author's face) at 3:39, when at 0:38 there's 4 shots showing the (your) face? Really impressive research, nevertheless!

  • @muhammadsalmangalileo945
    @muhammadsalmangalileo945 4 года назад

    Very nice work

  • @HelloMynameisKPJ
    @HelloMynameisKPJ 6 лет назад

    Nice work! :)

  • @ivanrasierer3257
    @ivanrasierer3257 6 лет назад

    Good work!

  • @mariojuarez2951
    @mariojuarez2951 7 лет назад

    Why does the original error occur during the double integration?

    • @VladQuake
      @VladQuake 6 лет назад

      ruclips.net/video/_q_8d0E3tDk/видео.html

    • @jacksonkr_
      @jacksonkr_ 6 лет назад

      ruclips.net/video/_q_8d0E3tDk/видео.html

    • @snakehaihai
      @snakehaihai 5 лет назад

      various white noise, random walk noise and bias. once integrate them together, you amplify the noise. See github.com/ethz-asl/kalibr/wiki/IMU-Noise-Model for detail

  • @dafanghe1807
    @dafanghe1807 8 лет назад

    diao