{"id":"9344331a-2f98-45c5-beba-5be71d8471e9","arxiv_id":"2507.17132","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A Lagrangian dynamic model plus genetic algorithm optimizes leg segment dimensions, cutting swing-phase peak torques and energy consumption by 20 to 28 percent in simulation.","lead":"This paper designs a three-segment leg for a construction robot and uses a genetic algorithm to optimize its dimensions, claiming over 20% reductions in peak joint torques and energy consumption. The work combines standard Lagrangian dynamics with optimization applied to a new leg geometry, with validation only in software simulation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (35) defines 'energy consumption' as a signed sum of instantaneous power, so the reported 22-28% reductions may be an artifact of cancellation between positive and negative power intervals; recomputing Table 4 with an unsigned metric is required before the energy claim can stand.","rationale":"The reader's weakest_assumption identifies the missing mapping from the nine dimensions to masses and inertias as the main obstacle. That is a genuine reproducibility gap, and I agree the paper should supply t, the density, and the inertia formulas. But the more load-bearing problem for the central claim is that the energy objective in Eq. (35) uses signed power. The energy reductions in Table 4 are values of this signed objective, so the headline 'over 20% reduction in energy consumption' may hold only because positive and negative work cancel in the sum. This is not a question of missing constants; it is a question of whether the reported quantity means what the paper says it means. A concrete numerical check -- recomputing Table 4 with an absolute-value or positive-part energy metric -- would settle it without any new experiments. The reader's conditional verdict already requires adding missing details and a hardware test; I would keep that conditional status but add this specific metric check as a necessary condition. Because the concern is real but not yet confirmed, and because the peak-torque portion of the claim may still be sound, I do not move the verdict to reject; I keep the conditional assessment.","tokens_in":11236,"tokens_out":17923,"duration_ms":184101,"concrete_test":"Recompute the energy values in Table 4 using the same optimized and initial dimensions, trajectory, and discretization, but replace Eq. (35) with Qi_abs = sum_j |tau_j^i * thetadot_j^i| * T_j^i, and separately with the positive-part version Qi_pos = sum_j max(0, tau_j^i * thetadot_j^i) * T_j^i. If the reported 22.6-28.5% energy reductions do not persist under either unsigned metric, the energy-consumption claim is an artifact of signed-power cancellation and must be revised; if the reductions remain above 20%, this concern is settled.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing issue is the energy metric itself, not the missing wall thickness. In Section V.B, Eq. (35) defines joint energy consumption as Qi = sum_j tau_j^i * thetadot_j^i * T_j^i, with no absolute value, positive-part, or regenerative-loss model. Over the two swing phases, the planned joint velocities are symmetric: the hip velocity is nonnegative on the first leg of the trajectory and nonpositive on the second, while the gravitational and inertial torque components do not necessarily change sign with velocity. Thus positive and negative power intervals can cancel in the sum. Minimizing this signed quantity can be achieved by making the torque oppose the velocity during parts of the motion, which is not a physical reduction in actuator energy consumption. The headline claim in the abstract and Section V.D, 'reduces peak joint torques and energy consumption by over 20%', rests on Table 4, which is computed from this signed metric. Even if every dynamic equation and mass value were correct, the energy numbers would not be a valid energy budget. The missing wall thickness t in Eq. (37) and the absent mass/inertia formulas in Section V.A make the model harder to reproduce, but the signed-power flaw attacks the reported numbers directly under the paper's own assumptions.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a design and dimensional optimization study of a three-joint (root/hip/knee) leg for a hexapod construction robot. Joint-space trajectories are planned with quintic polynomials, a Lagrangian dynamic model is derived for three uniform-rod leg segments, and a genetic algorithm is used to optimize nine geometric variables (length, width, and height of each segment) with peak joint torque and joint energy consumption as weighted objectives, subject to foot-reach and bending-stiffness constraints. The authors report 24-28% reductions in peak torques (Table 3) and 22-28% reductions in energy consumption (Table 4), and present ADAMS simulations showing reduced joint driving power. The paper concludes that the optimized leg improves dynamic performance.","tokens_in":11451,"tokens_out":9468,"duration_ms":103130,"significance":"If the quantitative claims were correct, the work would offer a straightforward, reusable template for dimensioning legged-robot links: an explicit dynamic model, a transparent genetic-algorithm loop, and an independent multibody check. The dynamic equations are derived from first principles, and the trajectory polynomials are given in closed form, which are strengths. However, the energy-consumption metric in Eq. (35) is a signed sum of mechanical power and is therefore not an energy budget, and the optimization is not fully reproducible because the wall thickness in Eq. (37) and the mass/inertia formulas connecting the nine dimensions to m_i and I_i are missing. These issues bear directly on the headline 20% torque and energy reductions, so the central claim is not yet established as stated.","major_comments":[{"comment":"Equation (35) defines Q_i as the sum of tau_i^j * thetadot_i^j * T_i^j without an absolute value, a positive-part operator, or a regenerative-loss model. For the planned trajectory, the hip velocity is nonnegative on the first swing segment (Eq. 2) and nonpositive on the second (Eq. 5), while the knee velocity has the opposite pattern (Eqs. 3 and 6); the gravitational torque components in Eqs. (30) and (33) do not share this sign symmetry. Positive and negative mechanical power intervals can therefore cancel in the sum, so Q_i is net mechanical work rather than actuator energy consumption. The 22.60-28.47% reductions in Table 4 and the abstract claim of over 20% energy reduction rest on this signed metric. The ADAMS curves in Section VI report driving power but no unsigned energy integral, so they do not repair the metric. I request recomputation with Q_i = sum |tau_i^j * thetadot_i^j| * T_i^j (or with an explicit regenerative-loss model) and rerunning the optimization under the corrected objective.","section":"V.B, Eq. (35)"},{"comment":"The wall thickness t in Eq. (37) is never defined numerically or as an optimization variable, so the bending-stiffness constraint EIz >= (1-mu)EIz0 cannot be evaluated. More generally, the paper states that the hollow rectangular simplification preserves mass, center of mass, and moment of inertia, but it provides no formulas mapping the nine design variables (l_i, w_i, h_i, and the wall thickness t) to the segment masses m_i and inertias I_i used in Eqs. (18)-(33). Tables 1 and 2 give masses for the initial and optimized designs, but the reader cannot check how m_i and I_i are updated when the dimensions change. Since every torque and energy value flows through m_i and I_i, this missing mapping is load-bearing. Please supply the closed-form mass/COM/inertia relations, the value or bounds of t, and the resulting I_i values for both designs.","section":"V.A, Eq. (37)"}],"minor_comments":[{"comment":"The text says the trajectory plot uses initial lengths l1=200 mm, l2=400 mm, and l3=400 mm, while Table 1 lists the initial coxa length as 140 mm and the femur/tibia lengths as 460 mm; please reconcile this discrepancy or clarify that the trajectory illustration uses a different nominal set.","section":"Section III vs Table 1"},{"comment":"The symbols alpha2 and alpha3 are used where a2 and a3 are intended; please standardize the notation for the center-of-mass distances.","section":"Eqs. (19)-(20)"},{"comment":"The energy values in Table 4 are labeled N-m, but Eq. (35) integrates power over time; joules (J) or N-m-s would be clearer and dimensionally correct.","section":"Table 4"},{"comment":"The genetic-algorithm description gives only population size and iteration count; please also report crossover and mutation probabilities, selection scheme, and convergence or stall criteria so that the optimization is reproducible.","section":"Section V.C"},{"comment":"The conclusions state that energy consumption decreases by 24% to 28%, but Table 4 reports a hip-joint reduction of 22.60%; please correct this numerical inconsistency.","section":"Section V.D"},{"comment":"The manuscript states twice that it is under review at Mechanics Based Design of Structures and Machines; such statements should be removed from the submitted version.","section":"Manuscript front matter"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the journal's scope, and the dynamic-modeling portion is largely transparent. The main risk is the signed energy metric in Eq. (35), which directly undermines the headline energy reductions; the missing mass/inertia mapping is a secondary but still load-bearing reproducibility issue. I would ask the editor to require the unsigned-energy recomputation and the explicit mass/inertia formulas before further consideration."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nQuick take: this is a workmanlike leg-design optimization paper, not a breakthrough. What's new is the specific objective formulation and the reported 20-28% reductions in peak torque and energy for a construction-robot leg. The Lagrangian derivation looks internally consistent on spot-check, and the ADAMS cross-check in a different software package is a reasonable sanity check. It deserves a serious referee, but the energy claim has a real problem.\n\nThe load-bearing issue is the energy metric. Eq. (35) defines joint energy as a signed sum of instantaneous power with no absolute value or positive-part. Over the two swing phases the hip velocity changes sign and the torque doesn't necessarily follow, so positive and negative intervals cancel. Minimizing that signed quantity is not the same as minimizing physical energy consumption; it can reward torque opposing velocity. The abstract's \"energy consumption by over 20%\" rests on Table 4 computed from this metric. The torque reductions in Table 3 are less tainted—peak torque is well-defined—but the energy numbers should be recomputed with an unsigned metric before they stand.\n\nSecondary issues: the optimization is under-specified. Wall thickness t in Eq. (37) is undefined, the mass/inertia formulas linking the nine dimensions to the masses in Tables 1 and 2 are missing, and variable bounds are not given. The paper's own conclusion admits the ideal-rod simplification, but that's a stated limitation, not a hidden one. The stance phase is never addressed, even though the abstract mentions load-bearing capability—though the optimization is explicitly swing-phase, so that's a scope limitation, not an error.\n\nThe torque comparison itself is credible as a simulation result under the uniform-rod model. The ADAMS simulation uses the same geometry, so it's a cross-check, not independent validation. A hardware test would be needed before trusting the numbers for a real robot.\n\nBottom line: the paper is a reasonable engineering study that would benefit from fixing the energy metric and filling in the optimization details. It's a straightforward application of known tools—Lagrangian dynamics, quintic trajectories, GA—so the novelty is modest. But the Lagrangian core is sound and the torque results, if the signed-power flaw doesn't propagate, are plausible.\n\nRecommendation: send it to peer review with a request for a major revision addressing the energy metric and the missing details. It's not a desk-reject, but it shouldn't be accepted as-is.","headline":"A standard leg-optimization study with an internally consistent Lagrangian model, but the energy claim is undercut by a signed-power metric and the optimization is under-specified.","tokens_in":12004,"tokens_out":618,"would_cite":false,"duration_ms":8207,"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":"Optimizing nine geometric leg dimensions cuts peak joint torque and energy consumption by over 20%.","keywords":["construction robots","legged robots","dynamic modeling","Lagrangian mechanics","dimensional optimization","genetic algorithm","joint torque","energy consumption"],"falsifier":"Build or measure a leg segment with the optimized dimensions in the intended material, measure its mass and moment of inertia about the joint, and compare with the uniform-rod values used in Tables 1 and 2; if the real values differ enough to recompute the torque curves and the peak torques no longer drop by at least 20% for all joints under the planned trajectory, the claim would be falsified. A simpler check is to compute the sensitivity of the predicted torques to the unspecified wall thickness $t$ in the stiffness constraint and see whether plausible values of $t$ change the optimized dimensions.","tokens_in":10990,"feed_emoji":"🤖","tokens_out":7780,"duration_ms":74723,"temperature":0.7,"pith_summary":"This paper proposes a dynamics-based dimensional optimization method for the legs of a construction-oriented hexapod robot, using an ant-inspired articulated leg with coxa, femur, and tibia. The authors build a Lagrangian dynamic model of the leg during its swing trajectory, define peak joint torque and joint energy consumption as evaluation metrics, and run a genetic algorithm over nine geometric dimensions (length, width, and height of each segment) to find a better leg shape. Their central claim, supported by the model calculations and by dynamic simulations, is that the optimized dimensions reduce the peak torque at each joint by 24-29% and the energy consumption by 23-28%, a reduction of over 20% for all joints. If true, this indicates that geometric tuning alone can substantially improve the load capacity and endurance of legged construction robots without changing their configuration or control logic.","feed_headline":"Leg optimization cuts joint torque and energy by over 20%","feed_subtitle":"A Lagrangian model plus genetic algorithm tunes nine leg dimensions to improve construction-robot efficiency.","key_machinery":"The carrier of the argument is the Lagrangian dynamic model of the leg, in which the coxa, femur, and tibia are treated as uniform rods with mass concentrated at half length ($a_i = l_i/2$). From this model the paper derives closed-form expressions for the joint torques as functions of the nine geometric variables (length, width, and height of each segment) and of the joint angles, velocities, and accelerations along a quintic-polynomial trajectory. A genetic algorithm then minimizes a weighted sum of normalized peak torque and normalized energy consumption, with penalty terms for violating constraints on the farthest foot placement distance and the bending stiffness $EI_z = E [hw^3 - (h-2t)(w-2t)^3]/12$ of the hollow rectangular cross-sections. The same dynamic model is reused after optimization to compute the predicted torques and energies that appear in the final comparison.","core_discovery":"The paper's central discovery is that the geometric dimensions of a three-segment articulated leg, not just its lengths, have a strong effect on joint torque and energy consumption, and that optimizing all nine dimensions together yields a leg with substantially lower dynamic loads. Under a uniform-rod dynamic model with Lagrangian mechanics, the optimized leg reduces the peak torque of the root, hip, and knee joints by 28.53%, 24.47%, and 24.16%, respectively, and reduces energy consumption by 28.47%, 22.60%, and 24.14%. The same trend appears in the authors' multibody dynamic simulation, where joint driving power drops by roughly 20%. The optimization respects constraints on farthest foot placement distance and bending stiffness, so the lighter, shorter segments are claimed not to sacrifice the robot's reach or structural resistance.","pith_inferences":["The paper leaves the wall thickness $t$ of the hollow cross-sections unspecified, so the mapping from the nine optimized dimensions to the reported segment masses cannot be checked from the published data alone; this is our inference from the absence of the value, not a claim the paper makes.","The reported reductions are computed for a single two-second swing trajectory; whether the same dimensions also reduce torque and energy in stance, during walking gaits, or on uneven terrain is not tested.","The uniform-rod model omits joint friction and air resistance, so the percentage energy savings in a physical robot could differ from the 22-28% predicted in simulation.","The same optimization pipeline could be applied to other leg configurations and to additional objectives such as structural stress or actuator peak power, which the paper does not address."],"forward_implications":["If the claim holds, the same leg configuration can be used with smaller, cheaper actuators because peak torque demand drops by roughly a quarter.","The method gives a design-time tool: from a planned trajectory, a designer can compute optimal segment dimensions for any articulated legged robot before building hardware.","The optimization implies that leg designs which tune only segment lengths are leaving some energy savings on the table, since width and height also affect dynamic loads.","The constraint framework for reach distance and bending stiffness provides a template for incorporating structural limits into other robot dimension optimizations."],"supporting_citations":[],"fun_headline_variants":["Ant-inspired leg design cuts robot torque and energy by over 20%","Leg dimension optimization reduces construction robot energy by 22-28%","Lagrangian model tunes legs to cut joint torque and energy consumption","Optimized leg geometry slashes peak joint torques by up to 28%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire torque and energy prediction rests on treating each leg segment as a perfectly uniform rod with mass at its midpoint and assuming the simplified hollow rectangular cross-sections preserve the real mass, center of mass, and moment of inertia, which the paper does not demonstrate.","fun_headline_variants_meta":{"raw":{"variants":["Ant-inspired leg design cuts robot torque and energy by over 20%","Leg dimension optimization reduces construction robot energy by 22-28%","Lagrangian model tunes legs to cut joint torque and energy consumption","Optimized leg geometry slashes peak joint torques by up to 28%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000306,"raw_usage":{"total_tokens":1740,"prompt_tokens":920,"completion_tokens":820,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":536,"completion_tokens_details":{"reasoning_tokens":742}},"tokens_in":536,"tokens_out":820,"duration_ms":10098,"temperature":1.0,"reasoning_tokens":742,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T14:57:28.561553+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Build or measure a leg segment with the optimized dimensions in the intended material, measure its mass and moment of inertia about the joint, and compare with the uniform-rod values used in Tables 1 and 2; if the real values differ enough to recompute the torque curves and the peak torques no longer drop by at least 20% for all joints under the planned trajectory, the claim would be falsified. A simpler check is to compute the sensitivity of the predicted torques to the unspecified wall thickness $t$ in the stiffness constraint and see whether plausible values of $t$ change the optimized dimensions.","supporting_citations":[],"review_version":1}