-
Notifications
You must be signed in to change notification settings - Fork 1
/
robosub21.html
744 lines (677 loc) · 46.2 KB
/
robosub21.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
<!doctype html>
<html lang="en">
<head>
<!--<h2>hello</h2> -->
<title>Team Tiburon — The AUV Team Of NITR</title>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<link rel="shortcut icon" href="https://github.com/auvnitrkl/webAssets/blob/main/images/favicon.ico?raw=true">
<link rel="stylesheet" href="css/custom-bs.css">
<link rel="stylesheet" href="css/jquery.fancybox.min.css">
<link rel="stylesheet" href="fonts/icomoon/style.css">
<link rel="stylesheet" href="fonts/line-icons/style.css">
<!-- MAIN CSS -->
<link rel="stylesheet" href="css/style.css">
<link rel="stylesheet" href="css/robosub.css">
<!-- FONTAWESOME -->
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/5.2.0/css/all.min.css">
</head>
<body>
<!-- it is used to show the spining circle before loading the page-->
<div class="loader">
<div class="spinner-border text-primary" role="status">
<span class="sr-only">Loading...</span>
</div>
</div>
<!-- use for menu in narrow screen mode-->
<div class="site-wrap">
<div class="site-mobile-menu site-navbar-target">
<div class="site-mobile-menu-header">
<img style="margin-left: 10px; padding: 5px;" class="img-fluid" height="45" width="45" src="https://github.com/auvnitrkl/webAssets/blob/main/images/logo.png?raw=true"></a>
<div class="site-mobile-menu-close mt-3">
<span class="icon-close2 js-menu-toggle"></span>
</div>
</div>
<div class="site-mobile-menu-body"></div>
</div> <!-- .site-mobile-menu -->
<!-- NAVBAR -->
<header class="site-navbar" id="top">
<div class="container-fluid">
<div class="dark-div">
<button onclick="darkTheme()">DM</button>
</div>
<div class="mySidenav">
<a href="#timeline" class="t-mech">Timeline</a>
<button onclick="expand(0)" class="t-mech-btn">+</button>
<a href="#software" class="t-elec">Software</a>
<button onclick="expand(1)" class="t-elec-btn">+</button>
<a href="#electronics" class="t-soft">Electronics</a>
<button onclick="expand(2)" class="t-soft-btn">+</button>
<a href="#mechanical" class="t-pr">Mechanical</a>
<button onclick="expand(3)" class="t-pr-btn">+</button>
</div>
<div class="row align-items-center">
<div class="site-logo col-8"><a href="index.html">
TIBURON</a>
<a href="index.html"><img style="margin-left: 10px; padding: 5px;" class="img-fluid" height="45" width="45" src="https://github.com/auvnitrkl/webAssets/blob/main/images/logo.png?raw=true"></a>
<iframe src="https://www.facebook.com/plugins/like.php?href=https%3A%2F%2Fwww.facebook.com%2Ftiburonauv%2F&width=72&layout=button_count&action=like&size=small&show_faces=true&share=true&height=21&appId"
width="140" height="20" style="border:none;overflow:hidden" scrolling="no" frameborder="0" allowtransparency="true" allow="encrypted-media"></iframe>
</div>
<nav class="mx-auto site-navigation">
<ul class="site-menu js-clone-nav d-none d-lg-block">
<li><a href="index.html">Home</a></li>
<li><a href="team.html">The Team</a></li>
<li><a href="achievements.html">Achievements</a></li>
<li><a href="gallery.html">Gallery</a></li>
<li><a href="events.html">Technical Events</a></li>
<li><a href="robosub21.html">Robosub 2021</a></li>
<li><a href="sponsors.html">Sponsors</a></li>
<li><a href="contact.html">Contact Us</a></li>
</ul>
</nav>
<div class="col-4 site-burger-menu d-block d-lg-none text-right">
<a href="#" class="site-menu-toggle js-menu-toggle"><span class="icon-menu h3"></span></a>
</div>
</div>
</div>
</header>
<section class="site-navbar2" id="top2">
<div>
<img class="img-fluid" height="75" width="75" src="https://github.com/auvnitrkl/webAssets/blob/main/images/logo.png?raw=true">
</div>
</section>
<!-- HOME -->
<section class="home-section section-hero overlay slanted" id="home-section">
<div class="container">
<div class="row align-items-center justify-content-center">
<div class="col-md-8 text-center">
<h1><strong>ROBOSUB 2021</strong></h1>
<h2 style="color:cornsilk;">The AUV Team Of NIT Rourkela</h2>
<div class="mx-auto w-75">
<!--need to change this line -->
<h3 style=" color:rgb(191, 191, 191)">Group of committed engineers working to design the best tech to understand the marine ways.</h3>
</div>
<!-- <p class="mt-5"><a href="#inductions" class="btn btn-outline-white btn-md ">INDUCTION PROCESS</a></p> -->
</div>
</div>
</div>
<!-- VIDEO -->
<div class="video-container">
<video autoplay loop="true" controls muted>
<source type="video/mp4" src="https://github.com/auvnitrkl/webAssets/blob/main/videos/tiburon.mp4?raw=true">
</video>
</div>
<a href="#about-us-section" class="scroll-button smoothscroll">
<span class=" icon-keyboard_arrow_down"></span>
</a>
</section>
<!-- ABOUT US -->
<section class="site-section about-us-section container">
<div class="row">
<div class="col-lg-6 col-sm-12">
<div class="d-flex flex-row justify-content-center robosub-title align-items-center">
ROBOSUB 2021
</div>
</div>
<div class="col-lg-6 col-sm-12">
<div id="quick-access" class="d-flex flex-column justify-content-center align-item-center">
<div class="robo-opt m-auto">QUICK ACCESS</div>
<button><a href="./tdr/TDR_Tiburon.pdf" target="_blank">View TDR</a></button>
</div>
</div>
</div>
<div id="timeline" class="robosub-br">
<h1 class="super-heading ">Bot Timeline</h1>
</div>
<div class="m-2">
<!-- <h3>uncomment this part at end</h3> -->
<h3><span class="badge badge-primary">Tiburon 1.0</span></h3>
<div class="text-center">
<iframe src="https://gmail916123.autodesk360.com/shares/public/SH56a43QTfd62c1cd9681bbe14ddcb25cb5f?mode=embed" width="80%" height="480" allowfullscreen="true" webkitallowfullscreen="true" mozallowfullscreen="true" frameborder="0"></iframe>
<p> </p>
</div>
<h3><span class="badge badge-primary">Tiburon 2.0</span></h3>
<div class="text-center">
<iframe src="https://gmail916123.autodesk360.com/shares/public/SH56a43QTfd62c1cd96800fad94c6ee81cfd?mode=embed" width="80%" height="480" allowfullscreen="true" webkitallowfullscreen="true" mozallowfullscreen="true" frameborder="0"></iframe>
<p></p>
</div>
<h3><span class="badge badge-primary">Tiburon 3.0</span></h3>
<div class="text-center">
<iframe src="https://gmail916123.autodesk360.com/shares/public/SH56a43QTfd62c1cd968cecf0948b737743e?mode=embed" width="80%" height="480" allowfullscreen="true" webkitallowfullscreen="true" mozallowfullscreen="true" frameborder="0"></iframe>
<p></p>
</div>
<h3><span class="badge badge-primary">Tiburon 4.0</span></h3>
<div class="text-center">
<iframe src="https://gmail916123.autodesk360.com/shares/public/SH56a43QTfd62c1cd968c4c25a46168bd2a6?mode=embed" width="80%" height="480" allowfullscreen="true" webkitallowfullscreen="true" mozallowfullscreen="true" frameborder="0"></iframe>
<p> </p>
</div>
</div>
<!-- software----------------------------------------------------------------/////////////////////////////////////////////////////////////////////////////////////////////////////////////-->
<div id="software" class="robosub-br">
<h1 class="super-heading">Software Subsystem</h1>
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Perception Stack
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle d-flex flex-row text-wrap m-1">
Object Detection
</div>
<div class="robo-text-content d-flex flex-row text-wrap">
Apart from the preprocessing and postprocessing of the image stream that is done by
OpenCV(an open-source computer vision) library,
we are using Yolo(You Only Look Once, a Convolutional Neural Network(CNN) based on DarkNet architecture) for object detection. This allows
us to accurately identify the objects with complex
textures or shapes from a greater distance. Yolo is
trained on a large dataset of images and videos collected from test runs and past competitions. In addition, we have applied various image augmentations
to impart variations and increase the dataset size.
</div>
<!-- <div class="robo-text-subtitle">
</div> -->
<!-- <div class="robo-text-content">
<b>YOLO V5/V3</b> Although traditional image processing methods were successful in detecting objects underwater(with suitable adjustment in HSV values as per lighting) but it comes with certain drawbacks such as failing to detect from a greater distance and unreliable results in different lighting and background. Here we have found Yolov3(You Only Look Once) CNN ( a real-time object detection algorithm that identifies specific objects in videos, live feeds, or images.)to be effective. We realized if we can navigate to the object close enough using this neural network-based model, then we can apply suitable thresholding techniques or can use the same model to get the center and pass through the target(gate) or reach the task at hand. For training the CNN, images were collected from previous competitions. Data augmentation was also performed to introduce variation and increase the dataset. We used our own Labeler Toolbox to label the images. 1/5th of the images were reserved as the test set to run the inference and the model(based on open-source neural network framework Darknet) was trained on the rest of the data. Unfortunately, we could not test the model real-time underwater on Nvidia Jetson due to the pandemic, but we still achieved very encouraging results testing it on our local systems(25 fps 87%).
Together with new machine learning approaches, traditional image processing techniques, color correction algorithms, and other fine-tuned approaches, we hope to strengthen the perception system of
Makara 4.0
</div> -->
</div>
<div class="col-lg-6 col-sm-12 p-2">
<div class="col-lg-11 col-sm-12 d-flex flex-row justify-content-center p-4 m-auto">
<div id="demo0" class="carousel slide" data-ride="carousel">
<!-- Indicators -->
<ul class="carousel-indicators">
<li data-target="#demo0" data-slide-to="0" class="active"></li>
<li data-target="#demo0" data-slide-to="1"></li>
</ul>
<!-- The slideshow -->
<div class="carousel-inner">
<div class="carousel-item active" >
<img class="img-fluid img-shadow" src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s1.png?raw=true" width="100%" height="auto" alt="Los Angeles">
<div class="ss-cap ">Detecting multiple bins</div>
</div>
<div class="carousel-item ">
<img class="img-fluid img-shadow" src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s2.png?raw=true" width="100%" height="auto" alt="Chicago">
<div class="ss-cap ">Detecting gate</div>
</div>
</div>
<!-- Left and right controls -->
<a class="carousel-control-prev" href="#demo0" data-slide="prev">
<span class="carousel-control-prev-icon"></span>
</a>
<a class="carousel-control-next" href="#demo0" data-slide="next">
<span class="carousel-control-next-icon"></span>
</a>
</div>
</div>
<!--
<div class="col-lg-12 col-sm-12 p-2">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s1.png?raw=true" alt="Image" class="img-fluid img-shadow">
</div>
<div class="col-lg-12 col-sm-12 p-2">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s2.png?raw=true" alt="Image" class="img-fluid img-shadow">
</div>
-->
</div>
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Preprocessing Techniques
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle d-flex flex-row text-wrap m-1">
Glare Removal
</div>
<div class="robo-text-content d-flex flex-row text-wrap">
Sometimes the images obtained from the bottom camera are too bright and keep fluctuating due to sunlight reflection(happens when the arena isn't that deep and has a shiny bottom surface). To remove glares from these images we have used an algorithm to make them more visible and clear.
The algorithm keeps track of the most recent 'n' images and finds the median, mean variation of 'n' RGB values respectively for each pixel in the image. The final image obtained has less glare and the lighting is spread equally. As the glares are rapidly moving due to surface water motion, taking its median helps to distribute its intensity and makes the image more clear and
visible.
</div>
</div>
<div class="col-lg-5 col-sm-12 d-flex flex-row justify-content-center p-4 m-auto">
<div id="demo" class="carousel slide" data-ride="carousel">
<!-- Indicators -->
<ul class="carousel-indicators">
<li data-target="#demo" data-slide-to="0" class="active"></li>
<li data-target="#demo" data-slide-to="1"></li>
</ul>
<!-- The slideshow -->
<div class="carousel-inner">
<div class="carousel-item active" >
<img class="img-fluid img-shadow" src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s4.jpeg?raw=true" width="100%" height="auto" alt="Los Angeles">
<div class="ss-cap">Original Image</div>
</div>
<div class="carousel-item ">
<img class="img-fluid img-shadow" src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s5.jpeg?raw=true" width="100%" height="auto" alt="Chicago">
<div class="ss-cap">Processed Image</div>
</div>
</div>
<!-- Left and right controls -->
<a class="carousel-control-prev" href="#demo" data-slide="prev">
<span class="carousel-control-prev-icon"></span>
</a>
<a class="carousel-control-next" href="#demo" data-slide="next">
<span class="carousel-control-next-icon"></span>
</a>
</div>
</div>
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle d-flex flex-row text-wrap m-1">
Color Correction
</div>
<div class="robo-text-content ">
The Robosub arena watercolor is greenish and unclear and thus making things difficult for the object detection model as they are trained on a different environment.
To change the color and make it more natural blue, we have simply changed the weight of the RGB channel of the image (called <b> Stretching </b>) and obviously, the weight of the green channel needs to be reduced. Then to remove the fogginess in the image we used the <b> Contrast Limited Adaptive Histogram Equalization (CLAHE) </b> algorithm that helped to increase the image contrast and made images more
clear.
</div>
</div>
<div class="col-lg-5 col-sm-12 d-flex flex-row justify-content-center p-4 m-auto">
<div id="demo2" class="carousel slide" data-ride="carousel">
<!-- Indicators -->
<ul class="carousel-indicators">
<li data-target="#demo2" data-slide-to="0" class="active"></li>
<li data-target="#demo2" data-slide-to="1"></li>
<li data-target="#demo2" data-slide-to="2"></li>
</ul>
<!-- The slideshow -->
<div class="carousel-inner">
<div class="carousel-item active" >
<img class="img-fluid img-shadow" src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s6.jpeg?raw=true" width="100%" height="auto" alt="Los Angeles">
<div class="ss-cap">Original Image</div>
</div>
<div class="carousel-item ">
<img class="img-fluid img-shadow" src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s7.jpeg?raw=true" width="100%" height="auto" alt="Chicago">
<div class="ss-cap">Stretched Image</div>
</div>
<div class="carousel-item ">
<img class="img-fluid img-shadow" src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s8.jpeg?raw=true" width="100%" height="auto" alt="Chicago">
<div class="ss-cap">Stretched + CLAHE Image</div>
</div>
</div>
<!-- Left and right controls -->
<a class="carousel-control-prev" href="#demo2" data-slide="prev">
<span class="carousel-control-prev-icon"></span>
</a>
<a class="carousel-control-next" href="#demo2" data-slide="next">
<span class="carousel-control-next-icon"></span>
</a>
</div>
</div>
</div>
<div class="row">
<div class="col-lg-12 col-sm-12 p-2">
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Model Optimization
</div>
<div class="robo-text-content ">
The perception stack is mostly carried out by object detection methods using deep neural networks. However, there has always been a trade-off between accuracy and real-time fps. We have approached this problem with model optimization using quantization by implementing state-of-the-art quantization-aware training methods. This approach simply converts all float32 weights and activations to int8 thus drastically reducing model size as well as the inference time. This enables us to use any low-power computing devices and stack multiple models in our pipeline without compromising
performance.
</div>
</div>
</div>
<div class="row">
<div class="col-lg-12 col-sm-12 p-2">
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Stereo Vision
</div>
<div class="robo-text-content ">
In traditional stereo vision, two cameras, displaced horizontally from one another are used to obtain two differing views on a scene, in a manner similar to human binocular vision. By comparing these two images, the relative depth information can be obtained in the form of a disparity map, which encodes the difference in horizontal coordinates of corresponding image points. Computer vision algorithms are also applied to images for smoothening and rectification of the images. Thus by using stereo vision we can perceive the distance between the object and the camera plane while carrying out various tasks during the
mission.
</div>
</div>
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Navigation Stack
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle d-flex flex-row text-wrap m-1">
Auto-tuning PID
</div>
<div class="robo-text-content d-flex flex-row text-wrap">
The PID controllers used in our AUV have to tune it (that means to obtain the value of Kp, Ki, Kd constants ) manually so that it manoeuvres properly.
But this time we tried to automate things and decided to build a tuning algorithm using Ant Colony Optimization (ACO). To tune the constants of each degree of freedom, the algorithm first generates some random values and tries to minimize the error generated from a cost function for each generated value and then determine the best values obtained from this generation and use them as the parents for the next generation. This happens generation over generation till we get an optimal value and every generation is more optimal than its previous
one.
</div>
</div>
<div class="col-lg-6 col-sm-12 p-2">
<div class="col-lg-12col-sm-12 p-2">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s9.jpeg?raw=true" alt="Image" class="img-fluid img-shadow">
</div>
</div>
</div>
<div class="robo-text-subtitle d-flex flex-row text-wrap m-1">
Underwater Localization
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<!-- <div class="robo-text-subtitle d-flex flex-row text-wrap m-1">
Doppler Velocity Log
</div> -->
<div class="robo-text-content ">
Underwater localization has always been a
challenge for AUVs because of the incapability of
inertial sensors to calculate linear velocity accurately due to high noise at the accelerometer. In
order to tackle this challenge, the team will be using Doppler Velocity Log (DVL), which can provide updated velocities using various acoustic
measurements
</div>
</div>
<div class="col-lg-5 col-sm-12 d-flex flex-row justify-content-center p-4 m-auto">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s10.png?raw=true" class="img-fluid img-shadow" alt="...">
</div>
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<!-- <div class="robo-text-subtitle d-flex flex-row text-wrap m-1">
Doppler Velocity Log
</div> -->
<div class="robo-text-content ">
Sensor data from several sensors such as
DVL, Inertial measurement unit(IMU), point cloud
from stereo vision, even measurements from electronics components are sensor fused to achieve accurate localization
results.
</div>
</div>
<div class="col-lg-5 col-sm-12 d-flex flex-row justify-content-center p-4 m-auto">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/soft/s11.png?raw=true" class="img-fluid img-shadow" alt="...">
</div>
</div>
<!-- electronics--------------------------------------------------------//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// -->
<div id="electronics" class="robosub-br">
<h1 class="super-heading">Electronics Subsystem</h1>
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Electronics High-Level Overview
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle m-1">
Improvements in Circuit
</div>
<div class="robo-text-content limit-text">
Though the previous version of our electronics stack met all the requirements perfectly, but, it also has its issues that we have addressed in this year’s rebuild.
One major issue was power management, previously we used to power the thrusters using one battery and rest of the circuit using another. This caused the thruster power to run out quickly than the other which reduced our bots runtime and efficiency, so this year we switched to a single power source. We now use a single battery at a time and after it runs out we switch to another using hot swap controllers.This results in increased runtime and better power management.
This year, we have decided to change the Microcontroller platform from ATMEL to STM32 line of microcontrollers giving us the flexibility to use High processing power where required or Low power usage if needed.
Last year’s stack missed critical security features like Under and overvoltage protection, current spike and overcurrent protection. So we have added this protection using eFuses and TVS diodes. The thought of security leads us to the Kill switch, last year we used a REED switch as a kill switch which was extremely unreliable, and had a lot of false firing issues, this year we switched to a better and more reliable waterproof physical button as a kill switch.
Another major issue was the lack of any “Debugging data” making any sort of debugging a major pain in the buttocks, so we have introduced an independently powered (2x18650 cells) custom made debugging board, which allow us to keep track of Voltages of various nodes of the circuit along with current going to each thruster and temperatures of ESCs. This data will be sent to the main CPU using CAN protocol (which will also be the standard protocol all over the stack) then the CPU can use this data in intelligent decision making like shutting down parts of the stack in case of local errors and saving the rest of the circuit and AUV. the board will also have an EEPROM to store all this data for later use in analytics. </p>
Following the debugging line of thought, in our previous stack making any kind of repair was also difficult, so this time we have decided to replace a single PCB with various modularized PCBs making it easy to just replace the problematic board with a replacement much like any other commercial electronics gadget. Also, this allows our debugging board to execute the local shutdown of unimportant components in case of an error in the
board.
<!-- <br>
<b>shorten this para</b> -->
</div>
</div>
<div class="col-lg-6 col-sm-12 p-2">
<div class="col-lg-12col-sm-12 p-2">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/elec/e1.png?raw=true" alt="Image" class="img-fluid img-shadow">
</div>
</div>
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Power Management Board
</div>
<div class="text-center">
<iframe src="https://gmail912066.autodesk360.com/shares/public/SH919a0QTf3c32634dcf66e8a5c9df6d23fc?mode=embed" width="80%" height="480" allowfullscreen="true" webkitallowfullscreen="true" mozallowfullscreen="true" frameborder="0"></iframe>
</div>
<div class="robo-text-content">
As already discussed due to galvanic isolation, we powered the Thrusters with one battery and rest with the another. The Thruster battery discharged fast, decreasing runtime and efficiency. So, this year a smart battery management system was incorporated which used 2 16000 mAh LIPO battery packs but one battery at a time.After a battery discharges to a minimum preset level we switch batteries with help of a hot swap controller and a STM32f103c8t6. This board is also responsible for generation of different voltage levels,i.e. 12V,5V and 3.3V to meet the requirements of different components across our
circuit.
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Backplane/Main Board
</div>
<div class="robo-text-content">
All the different boards are connected to the main board using high density connectors which are controlled by STMH745ZI-Q. This board communicates with JETSON TX2.
A high-quality, reliable Inertial Measurement Unit (IMU) is a crucial part of any autonomous underwater vehicle (AUV). Previously our AUV had to rely on accelerometer and gyroscope signals to obtain velocity and position values, which was giving lots of error in determining the exact position. So to assist our previous IMU, we are installing Doppler Velocity Log (DVL) in Makara which serially communicates to the CPU. Erroneous velocity values that are impossible or unlikely are filtered out. For samples when no values are received, the program extrapolates until a real value is received.
Furthermore, to provide an accurate way of measuring depth, a pressure-based depth sensor was also added to our AUV which communicates with STM32 microcontroller
through I2C.
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Actuator Board
</div>
<div class="text-center">
<iframe src="https://gmail912066.autodesk360.com/shares/public/SH919a0QTf3c32634dcff1b8e29dec3a8e09?mode=embed" width="80%" height="480" allowfullscreen="true" webkitallowfullscreen="true" mozallowfullscreen="true" frameborder="0"></iframe>
</div>
<div class="robo-text-content">
The AUV has a support for 10 thrusters and also we have incorporated MOSFETs to drive solenoid valves for torpedoes. We use a mix of T200 and T100 thrusters, controlled using BlueESC. We are also planning to introduce current sensors to measure the current going to individual thrusters thus helping us to get an approximate idea for speed of thrusters as we can’t have an encoder on the
thrusters.
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Debugging Board
</div>
<div class="text-center">
<iframe src="https://gmail912066.autodesk360.com/shares/public/SH919a0QTf3c32634dcf977a80b01f1b5cb6?mode=embed" width="80%" height="480" allowfullscreen="true" webkitallowfullscreen="true" mozallowfullscreen="true" frameborder="0"></iframe>
</div>
<div class="robo-text-content">
In our previous circuit, we used LEDs to debug circuits which made troubleshooting difficult and time consuming. So, in the new circuit we introduced an independent debug/protection board with an onboard OLED screen powered with an independent power supply. This board contains a STM32f103c8t6 microcontroller which monitors voltage, current and temperature at all important nodes using LTC2990. The data acquired is stored in an EEPROM which can be used for later debugging. Our circuits have also been incorporated with TPS2660x efuses to ensure circuit protection. These efuses can also be controlled by the debugging board in case of any
emergency.
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Acoustic Board
</div>
<div class="robo-text-content">
In our previous bot, we tried making passive sonar for locating the pinger with hydrophone and RaspberryPi with external ADC but that did not turn out to be working. So This time we planned the whole Acoustic signal processing part based on FPGA. We designed our own Analog frontend board which Has Adjustable Bandpass filters, Amplifiers with Adaptive gain control, and High-speed ADCs which connect to the Eclypse Z7 FPGA board. We find out the pinger heading by finding the Time Delay of Arrival of signals. FPGA can communicate to onboard Computer via Ethernet or CAN bus
There is an onboard AVR microcontroller ATMEGA328p that controls the Frontend gain and filtering also controls various power rails within the board. This board has its own local power distribution to meet the special requirements of the board. The microcontroller also interfaces with the debugging board via CAN and can regulate high-performance and power-saving modes. And power down partially or fully in case of fault when signaled from debugging board.
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Old Circuit
</div>
<div class="text-center">
<iframe src="https://gmail912066.autodesk360.com/shares/public/SH919a0QTf3c32634dcf1c501fcd4f7e618b?mode=embed" width="80%" height="480" allowfullscreen="true" webkitallowfullscreen="true" mozallowfullscreen="true" frameborder="0"></iframe>
</div>
<!-- mechanical---------------------------------------////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////-->
<div id="mechanical" class="robosub-br">
<h1 class="super-heading">Mechanical Subsystem</h1>
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Vehicle Design
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle m-1">
Design and Material Selection
</div>
<div class="robo-text-content">
The frame is designed to be compact and modular to reduce drag force down to a minimum and provide efficient hydrodynamics. The frame design borrows some inspiration from Tesla's Cybertruck model. Aluminium is used as the material of the frame due to its lightweight and easy machining ability. The frame is trapezoidal (from the side view). The design of the frame is such that each elementary structure (hollow box and aluminium rods) can be separated from each other. It allows us a never before level of flexibility in adjusting dimensions depending upon the
requirements.
</div>
</div>
<div class="col-lg-6 col-sm-12 p-2">
<div class="col-lg-12col-sm-12 p-2">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/mech/m1.jpeg?raw=true" class="img-fluid img-shadow" alt="...">
</div>
</div>
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle m-1">
Hull Design
</div>
<div class="robo-text-content">
The hull design is aimed to provide high strength and rigidity while keeping the weight to a minimum, so we went for an S Glass box-shaped enclosure.
The hull is divided into two layers – one for electronics and circuitry and the other for battery enclosure. The material of the hull is chosen in such a way that it does not crack at a depth of about 10m. This is thoroughly tested by simulation in ANSYS Static Structural by applying hydrostatic pressure equivalent to a depth of 10m underwater.
Both its ends can be easily opened to access the electronics inside. The design also provides watertight sealing to the bot. In case of an accidental failure, the electronic components are placed on a perforated and tapered base plate. This ensures the safety of the circuit in case of leakage as the water first collects at the bottom of the base
plate.
</div>
</div>
<div class="col-lg-6 col-sm-12 p-2">
<div class="col-lg-12col-sm-12 p-2">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/mech/m2.jpeg?raw=true" class="img-fluid img-shadow" alt="...">
</div>
</div>
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle m-1">
Camera Hull Design
</div>
<div class="robo-text-content">
The camera hull is made of acrylic. So similar simulations were performed on the camera hull in ANSYS as done in the main hull to ensure structural
integrity.
</div>
</div>
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle m-1">
Dropper and Grabber
</div>
<div class="robo-text-content">
A very simple mechanism has been employed to drop the required objects from the vehicle. In our mechanism, a BO motor is mounted to a PVC pipe. When the motor rotates, the obstruction causing the object to stay is removed, and hence the object falls. Newton Subsea Gripper, developed by BlueRobotics, is
used.
</div>
</div>
</div>
<div class="row">
<div class="col-lg-6 col-sm-12 p-2">
<div class="robo-text-subtitle m-1">
Torpedo Design
</div>
<div class="robo-text-content">
The torpedoes are in the shape of a hollow cylinder of 10mm diameter and are 3D printed for high accuracy. A pressure-based system is used to launch the torpedoes by a tank with pre-filled pressure of 30Pa. The torpedo is connected to the tank through PU fitting pipe passing through a coil actuating solenoid, which on receiving the command opens for about 10 milliseconds, thus allowing the air to pass from the tank to the torpedo through the PU fitting
pipe.
</div>
</div>
<div class="col-lg-6 col-sm-12 p-2">
<div class="col-lg-12col-sm-12 p-2">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/mech/m3.jpeg?raw=true" class="img-fluid img-shadow" alt="...">
</div>
</div>
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Cooling System
</div>
<div class="row">
<div class="col-lg-12 col-sm-12 p-2">
<!-- <div class="robo-text-subtitle m-1">
TODO: Sub points if any
</div> -->
<div class="robo-text-content">
The thermal system consists of a quasi active cooling system.
It helps us provide better cooling without having a heavy toll on our batteries. The primary method of heat transfer is conduction. There are other minor sources of heat, which are cooled by convection due to the circulating air.
The cooling system primarily consists of heat pipes paired with internal cooling mechanisms present in some of the components. The heat pipes help in easy and continuous flow of heat out from the hull via the heat sinks present in the ICs and Jetson.
Jetson has an inbuilt cooling fan which benefits a lot from this cooling system by providing it with a viable heat sink in the otherwise closed environment.
In the earlier models, we did not emphasise a dedicated cooling system as heat released by electronic components of our previous renditions was manageable. In our new vehicle, we have used better and more powerful ICs and a Jetson, which are primary heat sources. Each is connected to heat sinks and further connected to heat pipes. Furthermore, the heat pipes are in contact with the water outside directly for further heat transfer. There is a cooling fan within Jetson that helps us circulate the air in the hull, which is in contact with heat sinks, and heat pipes help in faster cooling and allow circulation of cool
air.
</div>
</div>
</div>
<div class="robo-text-title d-flex flex-row text-wrap m-1">
Experiment Results
</div>
<div class="row">
<div class="col-lg-12 col-sm-12 p-2">
<!-- <div class="robo-text-subtitle m-1">
TODO: Software Name if any
</div> -->
<div class="robo-text-content">
Analysis of the battery hull and main hull using ANSYS Static structural was done considering pressures at a depth of 10m for all possible materials taken into consideration initially.Then appropriate materials were chosen thereby giving results of factor of safety of given structures well above 3( S-Glass was chosen for main hull and battery hull) . Thermal analysis of heat pipes used in the cooling mechanism was done using ANSYS - Steady state Thermal. The cooling system proved efficient to dissipate the heat generated by the electronic
components.
</div>
</div>
</div>
<div class="row align-items-center">
<div class="col-lg-4 col-sm-12 p-2 align-content-start">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/mech/m4.jpeg?raw=true" alt="Image" class="img-fluid img-shadow">
<p class="im-caption">Equivalent stress analysis of Battery hull</p>
</div>
<div class="col-lg-4 col-sm-12 p-2 align-content-start">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/mech/m5.jpeg?raw=true" alt="Image" class="img-fluid img-shadow">
<p class="im-caption" >Factor of safety results for battery hull</p>
</div>
<div class="col-lg-4 col-sm-12 p-2 align-content-start">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/mech/m6.jpeg?raw=true" alt="Image" class="img-fluid img-shadow">
<p class="im-caption">Factor of safety results for analysis of Main Hull</p>
</div>
<div class="col-lg-4 col-sm-12 p-2 align-content-start">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/mech/m7.jpeg?raw=true" alt="Image" class="img-fluid img-shadow">
<p class="im-caption">Equivalent Stress results for analysis of Main Hull</p>
</div>
<div class="col-lg-4 col-sm-12 p-2 align-content-start">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/mech/m8.jpeg?raw=true" alt="Image" class="img-fluid img-shadow">
<p class="im-caption">Temperature Gradient results for Heat pipe cooling ICs</p>
</div>
<div class="col-lg-4 col-sm-12 p-2 align-content-start">
<img src="https://github.com/auvnitrkl/webAssets/blob/main/images/tdr/mech/m9.jpeg?raw=true" alt="Image" class="img-fluid img-shadow">
<p class="im-caption">Temperature gradient results for heat pipe cooling JETSON</p>
</div>
</div>
<div class="end-page">
<h3><span class="badge badge-secondary">References</span></h3>
<div class="resources">
<ul>
<li><a href="https://www.intechopen.com/books/electronics-cooling/heat-pipes-for-computer-cooling-applications">heat-pipes-for-computer-cooling-applications</a></li>
<li><a href="https://celsiainc.com/heat-sink-blog/heat-pipe-thermal-conductivity">heat-pipe-thermal-conductivity</a></li>
<li><a href="https://www.boydcorp.com/resources/temperature-control/frequently-asked-thermal-questions.html">frequently-asked-thermal-questions</a></li>
</ul>
</div>
</div>
</section>
<footer class="site-footer slanted-footer">
<a href="index.html#home-section" class="smoothscroll scroll-top">
<span class="icon-keyboard_arrow_up"></span>
</a>
<div class="container">
<div class="row mb-5">
<div class="col-6 col-md-3 mb-4 mb-md-0">
<h3>Technical Events</h3>
<ul class="list-unstyled">
<li><a href="https://sauvc.org/">Singapore AUV Challenge(SAUVC)</a></li>
<li><a href="https://www.niot.res.in/SAVe/">Student Autonomous Underwater Vehicle(SAVe) Competition</a></li>
<li><a href="https://robosub.org/">RoboSub 2019</a></li>
</ul>
</div>
<div class="col-6 col-md-3 mb-4 mb-md-0">
<h3>Address</h3>
<ul class="list-unstyled">
<li>Room No: ME313 , Mechanical Department , Mechanical Sciences Building, NIT Rourkela</li>
<li><a href="mailto:tiburonnitr@gmail.com">Contact mail</a></li>
</ul>
</div>
<div class="col-6 col-md-3 mb-4 mb-md-0">
<h3>Faculty Advisor</h3>
<ul class="list-unstyled">
<li>Dr Haraprasad Roy</li>
<li><a href="mailto:royh@nitrkl.ac.in">royh@nitrkl.ac.in</a></li>
</ul>
</div>
<div class="col-6 col-md-3 mb-4 mb-md-0">
<h3>Contact Us</h3>
<div class="footer-social">
<a href="https://www.facebook.com/tiburonauv/"><span class="icon-facebook"></span></a>
<a href="https://www.twitter.com/auvnitrkl"><span class="icon-twitter"></span></a>
<a href="https://www.instagram.com/auvnitrkl"><span class="icon-instagram"></span></a>
<a href="https://www.linkedin.com/school/tiburonauv"><span class="icon-linkedin"></span></a>
<!--<a href="#"><span class="icon-linkedin"></span></a>-->
</div>
</div>
</div>
<div class="row text-center">
<div class="col-12">
<p class="copyright"><small class="block">
<!-- Link back to Colorlib can't be removed. Template is licensed under CC BY 3.0. -->
Copyright ©<script>
document.write(new Date().getFullYear());
</script> All rights reserved | Developed from
<a href="https://colorlib.com" target="_blank">Colorlib</a></a>
<!-- Link back to Colorlib can't be removed. Template is licensed under CC BY 3.0. -->
</small></p>
</div>
</div>
</div>
</footer>
</div>
<!-- SCRIPTS -->
<script src="js/jquery.min.js"></script>
<script src="js/bootstrap.bundle.min.js"></script>
<script src="js/isotope.pkgd.min.js"></script>
<script src="js/stickyfill.min.js"></script>
<script src="js/jquery.fancybox.min.js"></script>
<script src="js/jquery.easing.1.3.js"></script>
<script src="js/readmore.js"></script>
<script src="js/jquery.waypoints.min.js"></script>
<script src="js/jquery.animateNumber.min.js"></script>
<script src="js/slide-show.js"></script>
<script src="js/custom.js"></script>
<script src="js/dark.js"></script>
<script src="js/sideNavExpand.js"></script>
</body>
</html>