Fault-Geometry-Aware Control Allocation for Emergency Quadrotor Landing under Single-Motor Loss of Effectiveness
Fault-Geometry-Aware Control Allocation for Emergency Quadrotor Landing under Single-Motor Loss of Effectiveness
This research compares three control allocators—bounded weighted least squares (PINV), a finite-candidate residual allocator (QP-lite), and a Cross-Entropy Method-tuned variant—for emergency quadrotor landing following single-motor loss of effectiveness. Using Crazyflie software-in-the-loop simulations, the study determines that the Cross-Entropy Method approach is selected for motors M1, M3, and M4, while PINV performs better for motor M2. The policy achieves 117/120 safe first-contact touchdowns on fresh seeds and 115/120 in a randomized-order integrated-supervisor evaluation. Sensitivity analysis reveals that one-factor-at-a-time plant perturbations can shift the nominal effectiveness threshold by up to 0.05495 for thrust coefficient and 0.04974 for mass, with the most frequently violated first-contact criterion changing from vertical speed to angular rate under certain conditions.
The study evaluates control allocation strategies for quadrotors under single-motor failure using software-in-the-loop simulations, identifying Cross-Entropy Method tuning as optimal for most motor configurations but highlighting sensitivity to plant-model mismatch in specific conditions.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- Environment
- SimulationOpen source S2
Reported numbers
safe first-contact touchdowns
117
Reported trials
View original evidence
g map, the policy achieves 117/120 safe first-contact touchdowns on fresh seeds and 115/120 in a separate randomized-order integrated-supervisor evaluation[cite: 11]. For the critical M2/PINV branch, Firth regression gives a nominal $\eta_{50}\approx0.4956$ with a within-model 95% bootstrap interval of approximately [0.4955, 0.4956][cite: 11]. One-factor-at-a-time $\pm10\%$ plant perturbations shift $\eta_{50}$ by up
Open source S4safe first-contact touchdowns
115
Reported trials
View original evidence
g map, the policy achieves 117/120 safe first-contact touchdowns on fresh seeds and 115/120 in a separate randomized-order integrated-supervisor evaluation[cite: 11]. For the critical M2/PINV branch, Firth regression gives a nominal $\eta_{50}\approx0.4956$ with a within-model 95% bootstrap interval of approximately [0.4955, 0.4956][cite: 11]. One-factor-at-a-time $\pm10\%$ plant perturbations shift $\eta_{50}$ by up
Open source S4total fresh seeds trials
120
Reported trials
View original evidence
g map, the policy achieves 117/120 safe first-contact touchdowns on fresh seeds and 115/120 in a separate randomized-order integrated-supervisor evaluation[cite: 11]. For the critical M2/PINV branch, Firth regression gives a nominal $\eta_{50}\approx0.4956$ with a within-model 95% bootstrap interval of approximately [0.4955, 0.4956][cite: 11]. One-factor-at-a-time $\pm10\%$ plant perturbations shift $\eta_{50}$ by up
Open source S4integrated-supervisor evaluation trials
120
Reported trials
View original evidence
g map, the policy achieves 117/120 safe first-contact touchdowns on fresh seeds and 115/120 in a separate randomized-order integrated-supervisor evaluation[cite: 11]. For the critical M2/PINV branch, Firth regression gives a nominal $\eta_{50}\approx0.4956$ with a within-model 95% bootstrap interval of approximately [0.4955, 0.4956][cite: 11]. One-factor-at-a-time $\pm10\%$ plant perturbations shift $\eta_{50}$ by up
Open source S4tested plant states for PINV vs CEM comparison
11
View original evidence
to 0.05495 for thrust coefficient and 0.04974 for mass; the most frequent violated first-contact criterion also changes from vertical speed in many conditions to angular rate for mass -10% and thrust-coefficient +10%[cite: 11]. At fixed $\eta=0.496$, PINV has a higher M2 safe count than frozen CEM in 9/11 tested plant states and ties in 2/11; the two ties are both 0/30[cite: 11]. The experiments therefore distinguis
Open source S5ties between PINV and CEM
2
View original evidence
to 0.05495 for thrust coefficient and 0.04974 for mass; the most frequent violated first-contact criterion also changes from vertical speed in many conditions to angular rate for mass -10% and thrust-coefficient +10%[cite: 11]. At fixed $\eta=0.496$, PINV has a higher M2 safe count than frozen CEM in 9/11 tested plant states and ties in 2/11; the two ties are both 0/30[cite: 11]. The experiments therefore distinguis
Open source S5failed M2 safe count in ties
0
Reported trials
View original evidence
to 0.05495 for thrust coefficient and 0.04974 for mass; the most frequent violated first-contact criterion also changes from vertical speed in many conditions to angular rate for mass -10% and thrust-coefficient +10%[cite: 11]. At fixed $\eta=0.496$, PINV has a higher M2 safe count than frozen CEM in 9/11 tested plant states and ties in 2/11; the two ties are both 0/30[cite: 11]. The experiments therefore distinguis
Open source S5remaining effectiveness threshold
0.496
View original evidence
e), and the same residual allocator with scoring parameters tuned offline by the Cross-Entropy Method (CEM)[cite: 11]. For this study, fault geometry is represented by the known failed-motor index[cite: 11]. At $\eta=0.496$ remaining effectiveness, matched development trials select CEM for M1, M3, and M4 and PINV for M2; the M3 selection is a one-trial margin (29/30 versus 28/30)[cite: 11]. After freezing this routin
Open source S3
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
arXiv:2611.08259v1 Announce Type: new Abstract: Partial single-motor loss of effectiveness (LoE) changes the remaining control-authority geometry of a quadrotor and can change which control allocator performs best during emergency landing[cite: 11]. We compare three allocators in Crazyflie software-in-the-loop simulation: bounded weighted least squares (reported as PINV), a finite-candidate residual allocator (QP-lit
Open source S2
g map, the policy achieves 117/120 safe first-contact touchdowns on fresh seeds and 115/120 in a separate randomized-order integrated-supervisor evaluation[cite: 11]. For the critical M2/PINV branch, Firth regression gives a nominal $\eta_{50}\approx0.4956$ with a within-model 95% bootstrap interval of approximately [0.4955, 0.4956][cite: 11]. One-factor-at-a-time $\pm10\%$ plant perturbations shift $\eta_{50}$ by up
Open source S4
to 0.05495 for thrust coefficient and 0.04974 for mass; the most frequent violated first-contact criterion also changes from vertical speed in many conditions to angular rate for mass -10% and thrust-coefficient +10%[cite: 11]. At fixed $\eta=0.496$, PINV has a higher M2 safe count than frozen CEM in 9/11 tested plant states and ties in 2/11; the two ties are both 0/30[cite: 11]. The experiments therefore distinguis
Open source S5
e), and the same residual allocator with scoring parameters tuned offline by the Cross-Entropy Method (CEM)[cite: 11]. For this study, fault geometry is represented by the known failed-motor index[cite: 11]. At $\eta=0.496$ remaining effectiveness, matched development trials select CEM for M1, M3, and M4 and PINV for M2; the M3 selection is a one-trial margin (29/30 versus 28/30)[cite: 11]. After freezing this routin
Open source S3
Source：arXiv Robotics — research abstracts · arxiv.org