{"id":"9f3e46d6-c2e5-4c16-b047-838afa2b5dae","arxiv_id":"1908.06301","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"An offline-calibrated propulsion database plus an online search algorithm automatically selects components and battery/airframe parameters for electric multicopters that meet target flight time, payload, altitude, and maneuverability.","lead":"This paper presents a two-stage computer method that picks motors, propellers, speed controllers, batteries, and frame size for an electric multicopter from a product database, based on target flight time, payload, altitude, and maneuverability. It precomputes and calibrates propulsion combinations offline, then searches them online in milliseconds, and validates the result against a real quadcopter test bench.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Appendix B's Eq. (40) is mathematically wrong: it writes sqrt(4K^2...-1) instead of sqrt(1+4K^2...); for typical parameters it yields a negative N*, invalidating the altitude conversion used in Eqs. (42) and (50).","rationale":"I read the paper as a practical design tool whose advertised capabilities include altitude-aware optimization. The weakest point is not the empirical constant alpha_air, which is a modeling assumption with a stated range; the Appendix B conversion is a derivational error in a core equation. The reader's verdict identified this sign error in the rationale but selected alpha_air as the weakest assumption; I disagree that alpha_air is more central, because the sign error is a proven inconsistency that directly breaks the altitude feature. The rest of the architecture (offline database, screening, objective function) is plausible and the single test-bench validation supports sea-level operation. But since the altitude requirement is part of the central claim, the manuscript as written is not correct until Eq. (40) is fixed. This does not change the overall CONDITIONAL verdict: the error is concrete, fixable, and localized, and the method may be sound after correction.","tokens_in":17626,"tokens_out":7043,"duration_ms":66064,"concrete_test":"Independently solve Eq. (38) for N* with the positive-root formula; plug in the MN3508 KV380 parameters from Section 3.3 (K_V=380, U_b=22.2, N*=5900, rho=1.2) and an altitude density, e.g., rho_hat=0.82 (about 4000 m). Compare the published Eq. (40) output with the corrected root. If Eq. (40) gives a negative N* or a thrust (via Eq. 42) differing by more than 10% from the corrected value, the altitude conversion is invalid. A supplementary check: run the online toolbox with identical requirements at sea level and at 4000 m and verify against a thrust bench.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the method handles altitude requirements rests on the air-density conversion in Appendix B. Eq. (38) is K_N rho N^2 + (1/K_V) N = U_b. Solving the quadratic and taking the positive root gives N* = [-1 + sqrt(1 + 4 K_V^2 K_N rho U_b)] / (2 K_V K_N rho). The paper's Eq. (40) has sqrt(4 K_V^2 K_N U_b rho - 1), replacing the +1 inside the radical by -1. This is not a positive root of the same equation. For the Section 3.3 example (K_V=380, U_b=22.2 V, N*=5900, rho=1.2), the corrected discriminant is about 2.54 while the published one is 0.54, so Eq. (40) gives a negative rotating speed. Because Eq. (42) and Eqs. (48)-(50) all use this N*, the online algorithm will produce incorrect full-throttle thrust and hovering current at any altitude differing from the database condition. The altitude input, explicitly listed as a design requirement, is therefore not handled correctly as written. This is an internal inconsistency, not a mere parameter uncertainty.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a two-stage design automation method for electric multicopters. An offline algorithm constructs a database of motor-ESC-propeller combinations using experimental or manufacturer test data, with thrust-current curves fitted by second-order polynomials. An online algorithm takes design requirements (hover time, payload, thrust ratio, propeller number, air density/altitude, battery type), computes battery mass and airframe diameter for each combination, screens combinations by hover-time error, and ranks the survivors with a weighted objective function. The method is validated with one test-bench case and compared with brute-force search. The stated advantages are high precision and sub-20-ms computation.","tokens_in":17881,"tokens_out":5539,"duration_ms":51300,"significance":"If the method works as described, it offers a practical design tool: the offline/online separation is a reasonable way to move expensive evaluations out of the interactive loop, and using component test data for calibration addresses a real limitation of purely model-based design. The core weight decomposition and battery-discharge equations are internally consistent, and the target hover time is used only as a screening criterion, so the approach is not circular. The paper also provides a public database and online toolbox, which strengthens reproducibility. However, the altitude-conversion error in Appendix B and the limited validation mean the precision and altitude-handling claims are not yet established as written.","major_comments":[{"comment":"The quadratic in Eq. (38), K_N ρ N^2 + (1/K_V) N = U_b, has positive root N* = [−1 + sqrt(1 + 4 K_V^2 K_N U_b ρ)] / (2 K_V K_N ρ). Equation (40) writes sqrt(4 K_V^2 K_N U_b ρ_hat − 1), with a minus sign inside the radical. For the manuscript's own example (Section 3.3: K_V=380, U_b=22.2, N*=5900, ρ=1.2), the argument is negative, so Eq. (40) gives no real positive rotating speed. Since Eqs. (42), (48), and (50) all use this N*_hat, the altitude design requirement (input ρ_hat) is not handled correctly as written. The sign should be corrected to plus and the altitude-conversion examples re-run.","section":"Appendix B, Eq. (40)"},{"comment":"The experimental verification consists of a single test-bench case comparing the website prediction (1.48 kg, 17.1 min, 0.5 kg payload, 4600 mAh) with one assembled multicopter (1.55 kg, 18 min, 0.55 kg, 5000 mAh). No repeated trials, measurement uncertainty, or environmental conditions are reported. Given that the online algorithm relies on fixed empirical constants (α_air=0.19, α_b=0.9, I_other≈0.5 A, and the fitted kt coefficients), a single close agreement is insufficient to substantiate the repeated claim of 'high precision.' Additional trials or a bounded sensitivity analysis are needed.","section":"Section 5.2, Table 2"},{"comment":"The battery mass is computed as the residual after fixing α_air=0.19, but the cited source [24] gives a range 0.08–0.40 for the airframe weight ratio. Because mbattery enters the hover-time calculation through Eq. (20), an actual frame at either end of this range would shift the predicted tdis substantially even if the propulsion database is exact. The paper does not report how sensitive the final design ranking is to α_air (or to α_b and I_other); please add a sensitivity study or, failing that, downgrade the precision claims accordingly.","section":"Section 4.1.1, Eqs. (13)-(16)"}],"minor_comments":[{"comment":"Equation (27) uses X_i for both the evaluation index and the normalizing parameter; the ratio should be written X_i / \\bar{X}_i to avoid confusion.","section":"Eq. (27)"},{"comment":"The phrase 'The eight most optimal eight multicopter design results' repeats 'eight,' and the conclusion states 'more than 2000 items' while Section 5.2 refers to 'more than 200 motors' and a database of 'more than 1500' combinations; please reconcile these numbers.","section":"Section 5.2 and Conclusion"},{"comment":"The phrase 'brutal search method' should be 'brute-force search method' for consistency with Algorithm 1.","section":"Throughout"},{"comment":"The tolerance threshold ε_t is introduced without a recommended value or selection rule; the authors should indicate a default and show how it affects the set of accepted designs.","section":"Eq. (21)"},{"comment":"The caption states 'Figs. 2(a)(b)(c) are of the common form and Figs. 2(a)(b)(c) are of the coaxial form'; the second list should refer to panels (d), (e), and (f).","section":"Figure 2 caption"}],"recommendation":"major_revision","confidential_remarks":"The manuscript depends heavily on the authors' prior work [14]–[16]; the editor may wish to clarify the incremental contribution relative to those papers, especially whether the offline database construction largely repeats [16]. The availability of the cited online database and toolbox URLs should also be verified."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper: the offline/online design-search architecture is genuinely useful and the validation, though thin, supports the core idea. But there is a sign error in Eq. (40) (Appendix B) that breaks the altitude-conversion step as written, and that step is load-bearing for one of the paper's headline capabilities.\n\nThe paper's contribution is the two-stage approach: precompute a calibrated database of motor-ESC-propeller combinations offline, then do a fast search online with a seven-index normalized objective function. That is a practical improvement over brute-force search and over the MILP-based methods it cites, and the published database plus the online toolbox make the work reproducible. The single F450 test-bench comparison is close (1.48 vs 1.55 kg, 17.1 vs 18 min), and the second-order thrust-current fit with R^2 > 0.99 is a legitimate offline calibration.\n\nThe soft spots are real but localized. The stress-test is right: Eq. (40) should have sqrt(1 + 4 K_V^2 K_N U_b \\hat{\\rho}), not sqrt(4 K_V^2 K_N U_b \\hat{\\rho} - 1). The quadratic from Eq. (38) gives the +1 version. As written, plugging the Section 3.3 example back into Eq. (40) with \\hat{\\rho} = \\rho recovers about 1420 RPM instead of 5900; for some inputs the discriminant goes negative and N* becomes imaginary. Since Eqs. (42) and (50) propagate N*, the online algorithm's altitude handling is incorrect. That is an internal inconsistency, not parameter uncertainty. It is fixable, but the paper must be revised and the altitude example re-run.\n\nLesser issues: alpha_air = 0.19 is a fixed empirical constant taken from a range of 0.08-0.40, and the hover-time prediction scales with it. The validation is a single case with no error bars or repeated trials. Offline database construction cost is asserted to be large but never quantified. The objective-weight choices are acknowledged as designer preferences, which is fine. None of this is fatal to the method's overall direction.\n\nWho should read it: practitioners building a multicopter design tool, and researchers working on component-level UAV sizing. It earns a serious referee slot, but the referee must require the Eq. (40) fix and an uncertainty or sensitivity check on alpha_air.","headline":"A useful two-stage multicopter design tool whose altitude-conversion step contains a sign error in Eq. (40) that needs a mandatory fix.","tokens_in":18470,"tokens_out":3132,"would_cite":false,"duration_ms":29254,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A two-stage database method converts multicopter requirements into a sorted, buildable design in under 20 ms.","keywords":["multicopter design optimization","propulsion system selection","component database","thrust-current curve fitting","flight endurance prediction","UAV sizing","objective function ranking"],"falsifier":"Weigh the airframe and battery of the validation quadcopter separately and compare the measured airframe fraction to 0.19; then recompute the predicted hover time with the measured fraction and see whether the discharge-time equation moves toward or away from the observed flight time.","tokens_in":17337,"feed_emoji":"🛸","tokens_out":7255,"duration_ms":72843,"temperature":0.7,"pith_summary":"This paper claims that an electric multicopter can be designed automatically from a short list of mission requirements: hovering time, payload mass, thrust ratio (maneuverability), number of propellers, air density, and battery type. The authors split the work into an offline pass, which builds a database of optimal motor-ESC-propeller combinations with calibrated thrust-versus-current curves, and an online pass, which searches that database and computes battery and airframe parameters for each candidate. Requirement-satisfying designs are then ranked by a weighted objective covering size, weight, performance closeness, efficiency, cost, and safety margin. The stated payoff is that what previously took hours or days of brute-force searching takes less than 20 ms on a large database, with experiment-calibrated precision.","feed_headline":"Multicopter design drops from hours to under 20 ms","feed_subtitle":"Enter flight time, payload, and maneuverability; get a ranked, buildable component list.","key_machinery":"The load-bearing object is the propulsion combination database $\\Phi_{mep}$. Each entry stores a motor, its chosen ESC and propeller, battery voltage, propeller diameter, KV value, mass, full-throttle thrust, current, speed, motor current limit, and the coefficients $k_{t0}, k_{t1}, k_{t2}$ of the fitted curve $I_e = k_{t2}T^2 + k_{t1}T + k_{t0}$ relating ESC input current to output thrust. This quadratic makes hovering current computable from a single thrust target, which feeds the discharge-time equation $t_{dis} = \\alpha_b \\, 60 \\rho_b m_{battery} / (U_b I_{bHover})$. The online screening criterion $|t_{dis} - \\hat{t}_{fly}| / \\hat{t}_{fly} \\le \\varepsilon_t$ keeps candidate combinations whose predicted hovering time matches the requirement, and the ranked objective $J = \\sum_{i=1}^7 k_i X_i / \\overline{X_i}$ finally orders designs by size, weight, requirement agreement, hovering efficiency, battery voltage and capacity, and safety margin.","core_discovery":"The paper's central claim is that multicopter design optimization can be decomposed into an offline product-selection problem and an online sizing problem. For each motor in a product database, the offline algorithm chooses the ESC and propeller that maximize a normalized score of full-throttle thrust, thrust efficiency, and propulsion-system mass, subject to safety and compatibility constraints; that winning combination, together with measured properties and a fitted quadratic thrust-current curve, becomes a row of a combination database. Online, each stored combination is tested against a design requirement by converting the required hovering thrust to battery current and predicting discharge time, and combinations whose predicted hovering time falls within a tolerance of the requested value proceed to battery sizing and airframe-diameter sizing from propeller geometry. Predicted performance is calibrated by experimental data, so that 'optimal' means best among tested, buildable combinations rather than best in an idealized model. The paper validates the pipeline on a quadcopter close to the F450 class and reports that the online stage completes in less than 20 ms on a large database.","pith_inferences":["Beyond the paper, the fixed airframe fraction could be turned into an input or a lookup table by material and size; doing so would remove the main parameter that controls battery mass in the residual-weight equation.","Beyond the paper, the same offline/online split transfers to coaxial layouts or other air vehicles as long as the propulsion performance model and the screening equations are replaced accordingly.","A direct test of the curve-fit assumption would be to compute residual errors of the quadratic fit across all database entries; if residuals grow for very large or high-KV motors, the fit order or a piecewise model would need extending."],"forward_implications":["A designer can specify hovering time, payload, thrust ratio, propeller count, altitude, and battery type and receive a component list, battery parameters, airframe diameter, and expected performance in a single automated run rather than by trial and error.","The same calibrated database works across altitudes because the paper derives density conversions for full-throttle thrust and hovering current, so test-bench data taken at one air density can be reused.","Online computation cost is set by the size of the preselected propulsion database, not by the product of all motor, ESC, propeller, battery, and airframe choices, which is what removes the hours-to-days search times of earlier approaches.","The tool's output is built to be assemblable: components must pass safety voltage and current limits and manufacturer compatibility, and the airframe diameter is chosen with geometric and aerodynamic clearance between propellers."],"supporting_citations":[{"why":"Supplies the performance-evaluation model and measured propulsion data that the offline calibration and hovering-time estimates build on.","marker":"[14]"},{"why":"Provides the analytical method for choosing an optimal propulsion combination that the offline algorithm generalizes and extends.","marker":"[15]"},{"why":"Supplies manufacturer test-bench data for motor-ESC-propeller combinations, which is the source of the experimentally calibrated database.","marker":"[23]"},{"why":"Gives the airframe weight fraction range and the 0.19 average used to compute battery mass as residual weight.","marker":"[24]"},{"why":"Defines multicopter body-system structure, layout types, propeller interference, and the airframe-radius clearance factor.","marker":"[2]"},{"why":"State-of-the-art mixed-integer optimization baseline whose reported multi-second run times the online algorithm is compared against.","marker":"[18]"},{"why":"Genetic-algorithm multicopter design baseline providing the multiobjective evaluation context and a speed comparison point.","marker":"[19]"}],"fun_headline_variants":["Drone design shrinks from hours to a 20-ms query","Pick parts offline, size battery in milliseconds","Requirements to ranked parts list in 20 ms","Multicopter design: offline part pick, millisecond battery sizing","Offline component database makes drone design a 20-ms task"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Every design inherits a fixed airframe weight fraction of 0.19, so battery mass is whatever remains after subtracting payload and propulsion mass from the total weight implied by the thrust ratio; if the real frame is lighter or heavier than that assumed share, predicted hover time shifts systematically.","fun_headline_variants_meta":{"raw":{"variants":["Drone design shrinks from hours to a 20-ms query","Pick parts offline, size battery in milliseconds","Requirements to ranked parts list in 20 ms","Multicopter design: offline part pick, millisecond battery sizing","Offline component database makes drone design a 20-ms task"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000773,"raw_usage":{"total_tokens":3433,"prompt_tokens":970,"completion_tokens":2463,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":586,"completion_tokens_details":{"reasoning_tokens":2380}},"tokens_in":586,"tokens_out":2463,"duration_ms":18353,"temperature":1.0,"reasoning_tokens":2380,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:50:06.387973+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Weigh the airframe and battery of the validation quadcopter separately and compare the measured airframe fraction to 0.19; then recompute the predicted hover time with the measured fraction and see whether the discharge-time equation moves toward or away from the observed flight time.","supporting_citations":[{"cited_title":"A practical performance eval- uation method for electric multicopters,","cited_arxiv_id":null,"evidence_quote":"Supplies the performance-evaluation model and measured propulsion data that the offline calibration and hovering-time estimates build on."},{"cited_title":"An Analytical Design Optimiza- tion Method for Electric Propulsion Systems of Multicopter UAVs with Desired Hovering Endurance,","cited_arxiv_id":null,"evidence_quote":"Provides the analytical method for choosing an optimal propulsion combination that the offline algorithm generalizes and extends."},{"cited_title":"T-motor official website,","cited_arxiv_id":null,"evidence_quote":"Supplies manufacturer test-bench data for motor-ESC-propeller combinations, which is the source of the experimentally calibrated database."},{"cited_title":"Electric multirotor UAV propulsion system sizing for performance prediction and design optimiza- tion,","cited_arxiv_id":null,"evidence_quote":"Gives the airframe weight fraction range and the 0.19 average used to compute battery mass as residual weight."},{"cited_title":"Quan, Introduction to Multicopter Design and Control","cited_arxiv_id":null,"evidence_quote":"Defines multicopter body-system structure, layout types, propeller interference, and the airframe-radius clearance factor."},{"cited_title":"Multicopter design opti- mization and validation,","cited_arxiv_id":null,"evidence_quote":"State-of-the-art mixed-integer optimization baseline whose reported multi-second run times the online algorithm is compared against."},{"cited_title":"Multirotor design optimization using a genetic algorithm,","cited_arxiv_id":null,"evidence_quote":"Genetic-algorithm multicopter design baseline providing the multiobjective evaluation context and a speed comparison point."}],"review_version":1}