Read automatically from the scorecard below — every claim can be checked against it.
31 % of the score comes from initial soc error (T10, 9.97 % × weight 0.100) — improving there moves the leaderboard position most.
Accuracy holds on the never-released cell (2.45 % vs 2.59 % on the open cells) — the model generalises rather than memorising the released data.
Cold is the weak spot: 3.57 % at −20 °C against 1.69 % at 25 °C. Below zero the cell's resistance rises and the voltage curve distorts; models without temperature awareness lose the most here.
A wrong initial SOC is not corrected: 9.97 % on test 10. Pure current integration can never recover from a bad start; voltage feedback is what fixes this.
Tolerates a biased current sensor (2.13 % on test 11) — the drift a bias causes is being corrected away.
Scorecard
Average RMSE (% SOC) per test case, weighted into the leaderboard score. Lower is better.
Estimation accuracyT1All cells(reference only — every other test is a subset of it)
2.56 %
0.0000
0.0000
T2Blinded cell (m448)
2.45 %
0.1000
0.2448
T3
Key cases
The runs that separate estimators — cold and hot cycles, the blinded cell, wrong initial SOC, a biased current sensor. Each plot says why it is there.
Selected drive cycles
The thirteen cycles the original MATLAB tool plotted, one per test: blinded vs non-blinded cell (tests 1–3), drive vs charge (test 4), the three payloads on one cycle (tests 5–6), standard vs non-standard highway profiles (tests 7–8) and the temperature extremes (test 9).
m80 UDDS at -20 °C
Whole cycle · RMSE 5.55 %, max 16.3 %
Actual SOC
UKF-Romulo
Error axis
Why this case: Test 9 — coldest condition: high resistance and a strongly non-linear OCV response — the hardest cycle for most estimators.
Wrong initial SOC (test 10)
Three cycles restarted with the estimator told the battery is at 90, 60 or 30 % while it is really full. Coulomb counting can never recover; voltage-informed estimators converge at a rate that is very visible here.
Wrong initial SOC 90 % — m80 LA92 at 25 °C
Whole cycle · RMSE 2.17 %, max 12.7 %
Actual SOC
UKF-Romulo
Error axis
Why this case: The cycle is joined at the point where the true SOC has already fallen to 90 %, so a model that assumes a full battery at start-up begins 10 % off. What matters is how fast the estimate converges onto the truth.
Current-sensor offset (test 11)
Three cycles re-run with ±0.3 A added to the measured current. Integration-based estimators drift steadily; the plot shows whether and how fast the model corrects the bias.
Current-sensor offset -0.3 A — m1000 US06 at -10 °C
Whole cycle · RMSE 3.49 %, max 5.8 %
Actual SOC
UKF-Romulo
Error axis
Why this case: A constant -0.3 A is added to the measured current before it reaches the model. Pure current integration drifts linearly; estimators that also use voltage should correct it.
RMSE vs. sensor offset
Test 11 — the same three cycles re-run with a constant current offset; 0 A is the unmodified run. Only ±0.3 A enters the score.
US06 at −10 °C
HWFET at 10 °C
LA92 at 40 °C
Test-case chartsThe scorecard as bar charts — tests 1–8, and RMSE against temperature (test 9).
Error by test case
Average RMS error (% SOC) for each blinded test case. Lower is better.
Error vs. ambient temperature
Test 9 — m80 cell, all blinded cycles at each chamber temperature.
All 144 cyclesEvery blinded drive cycle: per-cycle errors, and any of the plotted cycles in the time domain.
Cell
Temp.
Cycle
Duration
RMSE
MAE
MAXE
m80
-10 °C
UDDS
6.32 h
3.15
2.90
6.00
m80
-10 °C
HWFET
3.02 h
3.57
3.31
7.01
m80
-10 °C
LA92
4.74 h
3.95
3.58
8.60
m80
-10 °C
US06
2.30 h
3.89
3.42
6.87
m80
-10 °C
HWCUST
1.69 h
3.13
2.86
7.26
m80
-10 °C
HWGRADE
1.24 h
3.69
3.32
6.74
m80
-20 °C
UDDS
4.62 h
5.54
4.70
16.27
m80
-20 °C
HWFET
2.34 h
3.44
2.76
10.60
m80
-20 °C
LA92
3.54 h
5.08
4.02
15.55
m80
-20 °C
US06
1.86 h
2.27
1.77
8.66
m80
-20 °C
HWCUST
1.37 h
1.71
1.61
3.21
m80
-20 °C
HWGRADE
1.25 h
3.40
2.58
10.06
m80
0 °C
UDDS
8.75 h
1.68
1.42
4.38
m80
0 °C
HWFET
3.84 h
1.59
1.46
2.99
m80
0 °C
LA92
6.39 h
1.66
1.48
3.31
m80
0 °C
US06
2.93 h
2.26
1.98
3.45
m80
0 °C
HWCUST
2.08 h
2.92
2.69
4.03
m80
0 °C
HWGRADE
1.64 h
3.89
3.46
6.47
m80
10 °C
UDDS
11.32 h
1.79
1.29
4.37
m80
10 °C
HWFET
4.49 h
1.88
1.53
3.68
m80
10 °C
LA92
7.88 h
2.00
1.62
4.33
m80
10 °C
US06
3.42 h
1.45
1.21
3.73
m80
10 °C
HWCUST
2.42 h
1.47
1.15
2.92
m80
10 °C
HWGRADE
1.80 h
2.84
2.60
4.74
m80
25 °C
UDDS
16.13 h
1.26
0.97
3.44
m80
25 °C
HWFET
5.52 h
1.75
1.59
3.51
m80
25 °C
LA92
10.58 h
1.59
1.41
3.53
m80
25 °C
US06
4.22 h
1.59
1.45
3.08
m80
25 °C
HWCUST
2.86 h
2.06
1.95
3.30
m80
25 °C
HWGRADE
2.33 h
1.90
1.64
4.33
m80
40 °C
UDDS
12.40 h
1.53
1.26
3.27
m80
40 °C
HWFET
5.41 h
1.05
0.85
2.72
m80
40 °C
LA92
9.07 h
1.20
0.97
3.20
m80
40 °C
US06
4.17 h
0.92
0.70
2.92
m80
40 °C
HWCUST
2.95 h
0.85
0.64
2.09
m80
40 °C
HWGRADE
2.73 h
1.36
1.13
3.29
m448
-10 °C
UDDS
6.23 h
3.37
3.10
5.59
m448
-10 °C
HWFET
3.01 h
3.78
3.51
6.67
m448
-10 °C
LA92
4.59 h
4.10
3.67
6.71
m448
-10 °C
US06
2.25 h
4.03
3.44
6.46
m448
-10 °C
HWCUST
1.69 h
3.17
2.90
7.64
m448
-10 °C
HWGRADE
1.22 h
3.94
3.56
6.72
m448
-20 °C
UDDS
4.56 h
5.51
4.74
15.85
m448
-20 °C
HWFET
2.34 h
3.46
2.81
10.58
m448
-20 °C
LA92
3.46 h
4.59
3.75
14.21
m448
-20 °C
US06
1.80 h
2.04
1.65
7.72
m448
-20 °C
HWCUST
1.39 h
1.48
1.39
2.62
m448
-20 °C
HWGRADE
1.23 h
3.23
2.48
8.90
m448
0 °C
UDDS
8.60 h
1.66
1.39
4.09
m448
0 °C
HWFET
3.83 h
1.58
1.47
2.85
m448
0 °C
LA92
6.16 h
1.65
1.38
2.99
m448
0 °C
US06
2.87 h
2.44
2.16
3.68
m448
0 °C
HWCUST
2.09 h
3.13
2.88
4.38
m448
0 °C
HWGRADE
1.51 h
4.71
4.30
8.41
m448
10 °C
UDDS
11.08 h
1.98
1.41
4.83
m448
10 °C
HWFET
4.47 h
1.51
1.27
2.95
m448
10 °C
LA92
7.60 h
2.25
1.63
5.14
m448
10 °C
US06
3.35 h
2.12
1.69
4.18
m448
10 °C
HWCUST
2.42 h
1.95
1.68
3.35
m448
10 °C
HWGRADE
1.68 h
2.98
2.68
5.06
m448
25 °C
UDDS
15.59 h
1.76
1.60
3.44
m448
25 °C
HWFET
5.46 h
1.63
1.49
3.28
m448
25 °C
LA92
10.06 h
1.53
1.39
3.75
m448
25 °C
US06
4.07 h
1.69
1.49
3.71
m448
25 °C
HWCUST
2.86 h
1.90
1.81
3.23
m448
25 °C
HWGRADE
2.19 h
1.79
1.54
3.90
m448
40 °C
UDDS
12.09 h
1.78
1.50
3.89
m448
40 °C
HWFET
5.21 h
1.08
0.88
2.89
m448
40 °C
LA92
8.79 h
1.32
1.08
4.81
m448
40 °C
US06
4.06 h
0.92
0.69
3.82
m448
40 °C
HWCUST
2.92 h
0.77
0.58
1.79
m448
40 °C
HWGRADE
2.32 h
1.32
1.05
4.16
m448-N
-10 °C
UDDS
10.80 h
3.18
2.39
6.21
m448-N
-10 °C
HWFET
3.85 h
4.17
3.82
6.80
m448-N
-10 °C
LA92
6.66 h
4.72
3.88
8.43
m448-N
-10 °C
US06
2.70 h
5.43
4.57
8.77
m448-N
-10 °C
HWCUST
1.89 h
3.54
3.20
8.12
m448-N
-10 °C
HWGRADE
1.36 h
3.91
3.48
6.96
m448-N
-20 °C
UDDS
8.67 h
9.59
8.87
13.83
m448-N
-20 °C
HWFET
3.09 h
5.26
4.42
12.71
m448-N
-20 °C
LA92
5.21 h
9.11
8.52
12.99
m448-N
-20 °C
US06
2.23 h
3.41
2.96
9.10
m448-N
-20 °C
HWCUST
1.65 h
2.85
2.31
7.46
m448-N
-20 °C
HWGRADE
1.50 h
4.85
3.60
12.27
m448-N
0 °C
UDDS
12.79 h
1.69
1.13
4.91
m448-N
0 °C
HWFET
4.47 h
1.73
1.57
3.01
m448-N
0 °C
LA92
7.99 h
1.34
1.15
2.91
m448-N
0 °C
US06
3.22 h
2.43
2.13
3.68
m448-N
0 °C
HWCUST
2.31 h
3.06
2.79
4.25
m448-N
0 °C
HWGRADE
1.65 h
3.92
3.54
7.28
m448-N
10 °C
UDDS
13.92 h
1.77
1.07
4.79
m448-N
10 °C
HWFET
4.84 h
2.07
1.61
4.37
m448-N
10 °C
LA92
8.80 h
2.21
1.45
5.42
m448-N
10 °C
US06
3.56 h
1.99
1.58
3.97
m448-N
10 °C
HWCUST
2.53 h
1.89
1.63
3.33
m448-N
10 °C
HWGRADE
1.79 h
2.44
2.09
4.96
m448-N
25 °C
UDDS
15.61 h
1.47
1.28
3.26
m448-N
25 °C
HWFET
5.47 h
1.42
1.27
2.96
m448-N
25 °C
LA92
10.09 h
1.29
1.13
3.28
m448-N
25 °C
US06
4.08 h
1.29
1.16
2.36
m448-N
25 °C
HWCUST
2.86 h
1.87
1.78
3.20
m448-N
25 °C
HWGRADE
2.30 h
2.06
1.74
4.68
m448-N
40 °C
UDDS
16.75 h
1.71
1.41
3.87
m448-N
40 °C
HWFET
5.89 h
1.08
0.88
2.91
m448-N
40 °C
LA92
11.03 h
1.28
1.04
4.26
m448-N
40 °C
US06
4.44 h
0.87
0.67
3.49
m448-N
40 °C
HWCUST
3.13 h
0.75
0.59
1.69
m448-N
40 °C
HWGRADE
2.48 h
1.33
1.10
3.34
m1000
-10 °C
UDDS
5.91 h
3.72
3.41
5.85
m1000
-10 °C
HWFET
2.91 h
3.95
3.66
7.74
m1000
-10 °C
LA92
4.23 h
4.36
3.84
7.28
m1000
-10 °C
US06
2.09 h
3.93
3.36
6.38
m1000
-10 °C
HWCUST
1.67 h
3.35
3.06
8.00
m1000
-10 °C
HWGRADE
1.20 h
4.32
3.95
7.26
m1000
-20 °C
UDDS
4.31 h
5.94
5.01
16.75
m1000
-20 °C
HWFET
2.27 h
3.58
2.98
9.36
m1000
-20 °C
LA92
3.20 h
4.72
3.52
14.19
m1000
-20 °C
US06
1.74 h
1.74
1.07
7.52
m1000
-20 °C
HWCUST
1.37 h
1.44
1.34
2.52
m1000
-20 °C
HWGRADE
1.22 h
2.79
2.18
7.47
m1000
0 °C
UDDS
8.09 h
1.99
1.63
4.61
m1000
0 °C
HWFET
3.70 h
1.59
1.44
3.24
m1000
0 °C
LA92
5.60 h
1.76
1.49
3.37
m1000
0 °C
US06
2.66 h
2.35
2.11
3.72
m1000
0 °C
HWCUST
2.05 h
3.29
3.02
4.53
m1000
0 °C
HWGRADE
1.36 h
5.02
4.58
9.38
m1000
10 °C
UDDS
10.51 h
2.23
1.58
5.45
m1000
10 °C
HWFET
4.37 h
1.89
1.55
3.76
m1000
10 °C
LA92
7.06 h
2.17
1.63
4.90
m1000
10 °C
US06
3.11 h
1.25
1.03
3.45
m1000
10 °C
HWCUST
2.41 h
2.38
2.02
3.79
m1000
10 °C
HWGRADE
1.52 h
2.75
2.25
5.45
m1000
25 °C
UDDS
14.69 h
1.93
1.78
3.99
m1000
25 °C
HWFET
5.37 h
1.82
1.73
3.16
m1000
25 °C
LA92
9.30 h
1.71
1.59
4.40
m1000
25 °C
US06
3.79 h
2.04
1.82
4.43
m1000
25 °C
HWCUST
2.83 h
2.42
2.33
3.46
m1000
25 °C
HWGRADE
1.90 h
1.84
1.50
5.61
m1000
40 °C
UDDS
11.69 h
1.86
1.59
4.00
m1000
40 °C
HWFET
5.15 h
1.13
0.93
4.43
m1000
40 °C
LA92
8.33 h
1.32
1.10
3.57
m1000
40 °C
US06
3.89 h
0.93
0.72
3.73
m1000
40 °C
HWCUST
2.91 h
0.78
0.56
1.91
m1000
40 °C
HWGRADE
2.17 h
1.54
1.26
4.31
m80 UDDS at -20 °C
Whole cycle · RMSE 5.55 %, max 16.3 %
Actual SOC
UKF-Romulo
Error axis
One hour of padded data precedes every cycle in the evaluator and is excluded from the error metrics. Charts are down-sampled for display (peaks preserved); the full 1 Hz data is in the traces download below.
Score historyEvery evaluation attempt and every change to how this submission is scored; the current score is the last row.
How a unscented kalman filter worksThe canonical structure for this model family; the author's description below gives the specific architecture.
Unscented Kalman Filter — standardized view
Model-based state estimation
Downloads
PDF report (summary, every test case with the arithmetic, key plots, per-cycle table) · results as JSON · full 1 Hz traces of all 195 runs as a MATLAB v7 file (0.01 % SOC steps; readable with scipy.io.loadmat, see the readme variable inside).
UKF_Romulo.zip (44.3 KB) — deleted after evaluation
Visibility
Public
Evaluator
socbench-eval-0.1.0/matlab
Completed
Sep 1, 2026, 02:00 a.m. EDT
Submission ID
cmti97smo0001ii048at08mwl
Non-blinded cells
2.59 %
0.1000
0.2594
Operating conditionsT4Charging
2.83 %
0.1000
0.2829
T580 kg payload
2.37 %
0.0333
0.0790
T5448 kg payload, HVAC on
2.45 %
0.0667
0.1632
T6448 kg payload, HVAC off
2.86 %
0.0667
0.1907
T51000 kg payload
2.55 %
0.0333
0.0850
T7Standard drive cycles
2.54 %
0.1000
0.2544
T8Non-standard drive cycles
2.58 %
0.1000
0.2584
T9−20 °C
3.57 %
0.0167
0.0595
T9−10 °C
3.56 %
0.0167
0.0593
T90 °C
2.33 %
0.0167
0.0389
T910 °C
1.91 %
0.0167
0.0318
T925 °C
1.69 %
0.0167
0.0282
T940 °C
1.15 %
0.0167
0.0192
RobustnessT10Initial SOC error
9.97 %
0.1000
0.9973
T11Current sensor offset
2.13 %
0.1000
0.2129
Weighted error = Σ (weight × RMSE)
3.26 %
Max error 16.8 % — largest instantaneous |estimate − truth| over every blinded cycle (on m1000 UDDS at -20 °C — plotted under Key cases for new evaluations); not scoredComplexity 10 ±1 · Extreme+ (informational, not scored)
Governing equation
Vt=OCV(SOC)−R0I−i=1∑2VRC,i
V˙RC,i=−RiCiVRC,i+CiI
x^k=x^k−+Kk(Vmeas−Vt)
State carried in z
2 RC voltages + SOC; covariance P; current ECM parameters.
Ahnaf Rahman · McMaster University