{"id":"be693228-7489-4cef-8483-41bdb5c7b599","arxiv_id":"2605.23240","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":1,"one_line_summary":"The STL motion-planning problem is reformulated as a shortest-path problem over a graph of convex sets to generate smooth Bézier-spline trajectories satisfying logical, timing, smoothness, and velocity constraints.","lead":"The paper develops a framework that turns Signal Temporal Logic rules into a timed automaton, combines it with a convex decomposition of space, and solves a shortest-path problem on graphs of convex sets to produce smooth Bézier trajectories for robots. Smart generalists might read it to see how high-level task rules with timing can be turned into executable robot motions without manual tuning.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"No significant objection identified","rationale":"The reader's weakest_assumption directly restates the explicit scoping in the strongest_claim. Because the polynomial scaling and soundness are both stated only after the automaton and decomposition are given, the scoping itself is the load-bearing precondition; no further internal inconsistency appears in the abstract or claim.","tokens_in":1812,"tokens_out":249,"duration_ms":27532,"concrete_test":"Extract the precise statement of the soundness theorem (likely Theorem 1 or equivalent in §4) and the definition of the joint transition system; confirm that every accepting run in the product corresponds to a continuous trajectory whose Bézier pieces lie inside the prescribed convex sets at the clock valuations required by the timed automaton.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is explicitly conditioned on the timed automaton and convex decomposition already being fixed. Under that precondition the joint transition system yields a GCS whose shortest-path solution induces a Bézier trajectory; soundness follows from the product construction and the standard GCS convex relaxation. No hidden assumption about automatic construction of the automaton or decomposition, nor any gap between the discrete product and continuous-time STL satisfaction, is required by the stated theorem.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper proposes combining timed automata for STL specifications with graphs of convex sets (GCS) derived from a convex decomposition of configuration space. The STL planning problem is recast as a shortest-path problem on the resulting GCS; its solution yields a smooth Bézier-spline trajectory satisfying the STL formula, velocity bounds, and smoothness requirements. Soundness of the formulation is claimed, and computational complexity is shown to be polynomial in configuration-space dimension and Bézier degree once the timed automaton and convex decomposition are fixed. A compact timed-automaton construction for an expressive STL fragment is also given, supported by numerical experiments on low-dimensional cases, a 3-D quadrotor, a 30-DoF humanoid, and a UR-3 hardware demonstration.","tokens_in":1863,"tokens_out":567,"duration_ms":19030,"significance":"If the soundness result holds, the work supplies a scalable, convex-optimization-based route to continuous-time STL motion planning that produces executable smooth trajectories for high-DoF systems. The explicit conditioning on fixed automata and decompositions, together with the polynomial scaling claim for the GCS relaxation, distinguishes the contribution from prior discrete or sampling-based STL planners and could influence practical deployment in robotics tasks with temporal requirements.","major_comments":[{"comment":"Abstract and § on complexity: the polynomial scaling statement is explicitly conditioned on the timed automaton and convex decomposition already being fixed; the manuscript should state the explicit dependence of GCS size (hence solve time) on the number of automaton states and regions so that the overall complexity claim is unambiguous.","section":"Abstract / complexity analysis"},{"comment":"The soundness argument rests on the product construction correctly lifting discrete transitions to continuous-time STL satisfaction via Bézier segments; the manuscript should supply the precise lemma showing that satisfaction of the timed-automaton acceptance condition plus region occupancy implies STL satisfaction over the entire continuous trajectory (including at segment boundaries).","section":"Soundness section"}],"minor_comments":[{"comment":"Notation for the joint transition system (states, transitions, labels) should be introduced with a single compact diagram or table early in the paper to aid readability.","section":"Preliminaries / joint transition system"},{"comment":"The hardware experiment on the UR-3 arm would benefit from a brief description of the low-level controller used to track the Bézier trajectory and any observed tracking error.","section":"Experiments"},{"comment":"Bézier degree appears as a tunable parameter; a short discussion of its effect on both feasibility and computational cost would strengthen the presentation.","section":"Method"}],"recommendation":"minor_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the detailed and constructive report. The two major comments are addressed point-by-point below. Both can be resolved by targeted additions and clarifications that strengthen the presentation without altering the technical claims.","responses":[{"response":"We agree that an explicit statement of the dependence improves clarity. The GCS is constructed with one vertex per pair (automaton state, convex region), so its size is linear in |Q| × |R|. In the revision we will add this relation immediately after the existing conditioning sentence in both the abstract and the complexity section, making the overall scaling unambiguous while preserving the polynomial claim in dimension and degree once |Q| and |R| are fixed.","revision_made":"yes","referee_comment":"[Abstract / complexity analysis] Abstract and § on complexity: the polynomial scaling statement is explicitly conditioned on the timed automaton and convex decomposition already being fixed; the manuscript should state the explicit dependence of GCS size (hence solve time) on the number of automaton states and regions so that the overall complexity claim is unambiguous."},{"response":"The current soundness argument sketches the lifting via the product construction and Bézier interpolation, but does not isolate the required implication as a standalone lemma. We will insert a concise lemma (with a short proof) that states: if a trajectory satisfies the timed-automaton acceptance condition and each Bézier segment lies entirely inside its assigned convex region, then the continuous-time signal satisfies the original STL formula at every time, including segment boundaries. This addition will be placed in the soundness subsection.","revision_made":"yes","referee_comment":"[Soundness section] The soundness argument rests on the product construction correctly lifting discrete transitions to continuous-time STL satisfaction via Bézier segments; the manuscript should supply the precise lemma showing that satisfaction of the timed-automaton acceptance condition plus region occupancy implies STL satisfaction over the entire continuous trajectory (including at segment boundaries)."}],"tokens_in":1436,"tokens_out":419,"duration_ms":10993,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The paper's main move is to represent the STL formula with a timed automaton, pair it with a convex decomposition of the workspace, and form a joint transition system. Solving a shortest-path problem over the resulting graph of convex sets then yields a Bézier spline that satisfies the specification, velocity bounds, and smoothness by construction. The authors establish soundness from the product and show that the convex relaxation scales polynomially with dimension and degree under the fixed-input precondition. They also supply a compact template construction for an expressive STL fragment and run experiments on low-dimensional cases, a 3-D quadrotor, a 30-DoF humanoid, and a UR-3 arm in hardware. Those results are useful for seeing that the output trajectories are executable.","headline":"The paper turns STL specs into smooth Bézier trajectories by building a timed-automaton and GCS product and solving a shortest-path problem on it, with soundness and polynomial complexity once the automaton and decomposition are already fixed.","tokens_in":2361,"tokens_out":236,"would_cite":false,"duration_ms":12281,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":{"model":"grok-4.3","evidence":[],"headline":"STL motion planning via GCS + timed automata is orthogonal to RS forcing chain","alignment":"orthogonal","rationale":"Paper's central construction is the joint transition system (timed automaton × convex decomposition) turned into a GCS shortest-path problem whose solution yields a Bézier spline; soundness follows from the product run inducing an accepting TA trajectory. This machinery lives entirely in automata theory, convex optimization, and trajectory parameterization. RS theorems (reality_from_one_distinction, Jcost uniqueness via washburn_uniqueness_aczel, AlexanderDuality.alexander_duality_circle_linking forcing D=3, 8-tick periodicity, φ-ladder constants) derive spacetime and constants from a single distinction and reciprocal cost; none of those structures appear here. No shared cost function, ratio symmetry, periodicity, or parameter-free derivation.","tokens_in":62232,"confidence":"high","tokens_out":186,"duration_ms":9312,"cache_read_input_tokens":38528,"cache_creation_input_tokens":0},"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Coupling timed automata for STL specs with convex decompositions of configuration space turns motion planning into a shortest-path problem over graphs of convex sets that yields smooth Bézier trajectories.","keywords":["signal temporal logic","motion planning","graphs of convex sets","timed automata","bézier splines","convex optimization","robot trajectories"],"falsifier":"A computed shortest-path solution in the GCS that produces a Bézier trajectory violating the original STL formula or the velocity bounds.","tokens_in":2702,"feed_emoji":"","tokens_out":704,"duration_ms":13541,"temperature":0.7,"pith_summary":"The paper develops a planning method for continuous-time robot motions that must satisfy Signal Temporal Logic formulas while obeying smoothness and velocity limits. It converts the STL requirement into a timed automaton and merges it with a convex breakdown of the robot's workspace to create a joint transition system. This system is then modeled as a graph of convex sets whose shortest path encodes both task progress and geometric feasibility. Solving the resulting convex relaxation produces a Bézier spline trajectory that meets all constraints. The approach is shown to be sound and to scale polynomially with dimension once the automaton and decomposition are given.","feed_headline":"Shortest paths on convex graphs yield STL-satisfying trajectories","feed_subtitle":"Timed automata for the specification couple with space decomposition so that convex optimization produces smooth paths meeting logic and kin","key_machinery":"The joint transition system obtained by coupling a timed automaton for the STL specification with a convex decomposition of the configuration space, which is then encoded as a graph of convex sets whose shortest-path solution produces the desired Bézier spline.","core_discovery":"An STL specification is represented by a timed automaton, which is coupled with a convex decomposition of the configuration space to form a joint transition system encoding both task progress and region occupancy. Based on this joint transition system, the STL motion-planning problem is reformulated as a shortest-path problem over a GCS, whose solution induces a smooth Bézier-spline trajectory satisfying the STL specification, smoothness requirements, and velocity bounds. The formulation is sound and the convex relaxation scales polynomially with the configuration-space dimension and the Bézier degree once the timed automaton and convex decomposition are fixed.","pith_inferences":["The same joint-system construction could be reused for other temporal logics that admit timed-automaton encodings.","If convex decompositions can be updated online, the polynomial scaling opens a route to receding-horizon STL planning.","The quality of the initial convex decomposition directly limits both feasibility and path optimality.","Hardware experiments indicate that the resulting splines are directly usable as low-level references without further smoothing."],"forward_implications":["The induced Bézier trajectory satisfies the STL specification, required smoothness, and velocity bounds.","Once the automaton and decomposition are fixed, the convex relaxation scales polynomially with configuration-space dimension and Bézier degree.","A compact timed-automaton construction is available for an expressive STL fragment via dedicated templates and Boolean composition.","The method produces executable trajectories on benchmarks, a 3-D quadrotor, a 30-DoF humanoid, and a physical UR-3 arm."],"fun_headline_variants":["Convex graphs host shortest paths for STL robot trajectories","Timed automata encode STL specs in convex set graphs","Joint transition system turns STL into GCS shortest path","Bézier trajectories from GCS paths satisfy timed logic"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The timed automaton correctly captures the STL specification and the convex decomposition accurately represents the configuration space and its connectivity.","fun_headline_variants_meta":{"raw":{"variants":["Convex graphs host shortest paths for STL robot trajectories","Timed automata encode STL specs in convex set graphs","Joint transition system turns STL into GCS shortest path","Bézier trajectories from GCS paths satisfy timed logic"]},"model":"grok-4.3","cost_usd":0.006512,"raw_usage":{"total_tokens":3076,"prompt_tokens":727,"num_sources_used":0,"completion_tokens":58,"cost_in_usd_ticks":65124500,"prompt_tokens_details":{"text_tokens":727,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2291,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":727,"tokens_out":58,"duration_ms":15136,"temperature":1.0,"reasoning_tokens":2291,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-05-25T04:32:57.967005+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A computed shortest-path solution in the GCS that produces a Bézier trajectory violating the original STL formula or the velocity bounds.","supporting_citations":[],"review_version":1}