ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
2
DanyalTrial NumberTime (min:sec)Comments
3
Note I: 'stuck block' can be considered a failure, even if subsequent actions get the block to actually drop correct. I count them as place failures.
4
[This ambiguity is why we need to be careful when discussing "failure cases" because we can be lenient or strict.]
5
[We should make a distinction between recoverable failures and irrecoverable failures. That would help.]
6
Note II: "First set" refers to moving the first six blocks from one side of the board to another. The "second set" refers to moving those six back to the original side. Doing both sets completes a trajectory.
7
Note III: we are very strict in assigning grasp failures, basically any time we need to even do a second grasp "motion" is a failure. But those are corrected quickly by humans.
8
9
(NOT REPORTING THESE)
First set failureSecond set failureIrrecoverable
(He had no irrecoverable failures among all of his 34 trajectories)
10
Hand-over11:19
"Place failure" on the second set, did not drop correctly, Danyal corrected it. (I have a bad video, see Minho's videos.)
0
11
21:05
No failure cases. (I have a bad video, see Minho's videos.)
0
12
31:16
Had a 'stuck block' on first set but the next drop was able to hit it... otherwise no issues.
0
13
41:11
"Place failure" on the second set, third block. Danyal corrected it.
0
14
51:21
No failure cases, was a bit long due to repeated attempts to drop at correct spot.
0
15
61:17
Had a 'stuck block' on first set but Danyal nudged the block to drop correctly.
0
16
71:15No failure cases.0
17
81:04
"Grasp failure" on the second set (fifth block) where Danyal grabbed it and missed, but was able to get it on the next grip.
0
18
91:06No failure cases.0
19
101:02
"Place failure" on second set, third block, Danyal was able to correct for it by picking it up afterwards.
0
20
(REPORT THIS -- DUAL ARM)
Pick failuresPlace failuresTime (s)Attempts
21
No handover10:43
Had a 'stuck block / place failure' on first set, third pair, subsequent action was able to fix it.
100143123.583333333
22
20:38No failure cases.00038123.166666667
23
30:32No failure cases.00032122.666666667
24
40:35
"Grasp failure" on first set, first pair, subsequent action was able to grasp it correctly.
101035132.692307692
25
50:34No failure cases.00034122.833333333
26
60:31No failure cases.00031122.583333333
27
70:31
Had a 'stuck block / place failure' on the first set, but that was easily corrected, and it was also due to positioning of an existing block).
100131122.583333333
28
80:28
"Grasp failure" on second set, last pair, but was quickly corrected.
101028132.153846154
29
90:27No failure cases.00027122.25
30
100:31
Had a 'stuck block / place failure' on the first set, but it was easiliy corrected in a subsequent action.
100131122.583333333
31
(REPORT THIS -- SINGLE ARM)
1222.709615385
AVERAGE TO REPORT
32
One arm (Nov 18)10:55No failure cases.000
For this all trials had 12 attempts.
33
20:48No failure cases.000
Therefore we take average time of each trial, and divide by 12 to get the time per attempt.
34
30:58
No failure cases (could MAYBE argue a grasp failure on first set but really hard to justify even by our strict criteria)
000
Pretty straightforward.
35
40:56No failure cases.00055.3124.608333333
36
50:52No failure cases.000
37
60:58No failure cases.000
38
71:00No failure cases.000
39
80:55No failure cases.000
40
90:58No failure cases.000
41
100:53No failure cases.000
This is only because we know he did 12 attempts for each episode. Otherwise we'd have to do extra calculations.
42
Total attemptsNum successTimePer attempt
43
Two arm no H-over (Dec 06)
10:45
Place failure and grasp failure on first set, quickly corrected (especially grasp, this is harsh). Place failure on second set.
2101212941.5123.458333333
44
20:38
Place failure on second set, quickly corrected.
000011211
45
Sums + average240.8333333333
46
One arm no H-over (Dec 06)
11:04
No failure cases, though a bit silly in that he was able to get two blocks at once due to stickiness of the paint.
00000121261125.083333333
47
20:58
Place failure, first set, corrected. No failure cases (pooossibly grasp for first set but that'd be VERY harsh).
100011211
48
Note: this is not a count of the failure type but which *set* it was.
Sums + average240.9583333333
49
Not the same as what we have in the dvrk automated spreadsheet,
50
where we count the number of failure cases.
51
See the next columns for that.
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