{"id":"8334563c-e583-4069-abbb-d8982e732c59","arxiv_id":"1908.01826","paper_version":2,"verdict":"UNVERDICTED","confidence":"HIGH","novelty_score":3.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Jenny 5 is an open-source, low-cost, humanoid-style robot with a tracked platform, two 7-DOF arms, and a head, whose full design and code are publicly documented.","lead":"This paper describes the hardware, software, and source code of Jenny 5, a low-cost open-source humanoid robot. It is a build and status report for the robotics community rather than a study, so it offers a reference design for hobbyists and researchers working on similar platforms.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reproducible-robot claim is load-bearing but unverified: the paper itself says the assembly manual is under construction (§2) and BOM quantities are approximate (Table 2), so a third-party build from the repository alone is not established.","rationale":"I agree with the reader's verdict and largely with its weakest-assumption identification. The paper makes no hypothesis and reports no measurements under controlled conditions; it is a project description, so UNVERDICTED is the correct categorical response. The strongest claim is about availability and rebuildability of an open-source design. That claim has a real soft spot, and the manuscript itself points to it: the assembly manual is explicitly under construction and BOM quantities are approximate. A reader cannot currently verify that the design is complete enough to rebuild. I do not regard this as a reason to reject the paper: the source code and CAD are genuinely public, the license is permissive, and the description is internally coherent. The concern is that the paper's value as a reproducible open-source contribution is conditional on documentation that is not yet present. The test I propose would settle this directly by attempting a documentation-only build of one arm. Since the reader already treats the paper as UNVERDICTED documentation, my concern does not shift the verdict; it strengthens the rationale for not treating the rebuildability claim as established.","tokens_in":15410,"tokens_out":2971,"duration_ms":48604,"concrete_test":"Independently attempt a documentation-only replica of at least one subsystem, preferably an arm: (1) resolve every URL in §2 and confirm the BOM spreadsheet is public; (2) parse the BOM and the OpenSCAD source to list every unique part, and check that the BOM contains a defined quantity for each part used in the CAD; (3) following only the available text, CAD, and Fritzing files, assemble a single arm, recording every missing or ambiguous step. If the arm can be assembled without author contact and the BOM quantities reconcile with the CAD, the rebuildability concern is resolved; if the attempt stops at a missing manual step or an unquantified part, the paper should state that the design is not yet documented for third-party reproduction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central contribution is an open-source, low-cost, rebuildable robot; the load-bearing condition is that a third party can reconstruct Jenny 5 from the provided CAD, firmware, and BOM without access to the author. The paper itself flags the two places where this condition is weakest. Section 2 states that the assembly manual 'is under construction and will not be completed until the robot is completed'; Table 2 says quantities 'are not always exact' and §5.1 directs the reader to the assembly manual for details, which is the same unavailable document. The BOM is a Google spreadsheet, but no itemized cost calculation is given, so the '<2500 USD' claim has no auditable basis. CAD is in OpenSCAD and custom PCBs in Fritzing, but there is no published assembly sequence, wiring diagram, calibration procedure, or known-good build log. Without these, 'fully open-source' is true in the license/source sense but the stronger engineering claim—that a competent third party can rebuild the robot from the repository—is not demonstrated. This is not an internal inconsistency in the technical description; it is a gap in the evidence supporting the contribution's central value. If the manual and BOM remain incomplete, the paper functions as a project snapshot rather than a reproducible design.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper describes Jenny 5, an open-source, quasi-humanoid mobile robot developed by the author. The robot consists of a tracked mobile platform, a two-motor telescoping leg, two 7-DOF arms with grippers, a 2-DOF head, and an onboard laptop as its computing brain. The paper's central claim is that the complete design—OpenSCAD CAD files, Arduino firmware (Scufy), a PC control library (Scufy Lib), custom PCB layouts (Fritzing), an HTML5/WebSocket control interface, a bill of materials (BOM), and an assembly manual—is freely available under an open license, and that the robot can be built for less than 2500 USD in materials. The text provides a component-by-component hardware description, electronics list, software command reference, and some design calculations (e.g., arm torque and speed). The paper is written as a project snapshot and does not present experimental validation, measured performance data, or a comparison with other robotic platforms.","tokens_in":15646,"tokens_out":2961,"duration_ms":33325,"significance":"If the repository and documentation are complete and accurate, the paper would provide a useful open-source reference design for a low-cost 20-motor humanoid research platform. The author has made a commendable effort to share CAD, firmware, library, PCB, and client code under the MIT license, and the included torque and speed calculations give a concrete quantitative basis for the design's capabilities. The paper also ships pointers to publicly accessible repositories and datasheets, which aids reproducibility in principle. The main significance, therefore, hinges on whether a third party can actually rebuild the robot from the provided materials; this is exactly the point that the manuscript itself leaves unverified.","major_comments":[{"comment":"The central claim of the paper is that Jenny 5 is a fully open-source, low-cost, rebuildable robot, but the manuscript explicitly states that the assembly manual 'is under construction and will not be completed until the robot is completed' (§2) and that the BOM quantities 'are not always exact' with details deferred to that same unavailable manual (§5.1). Because a third-party build depends on exactly this documentation—assembly sequence, wiring, calibration, and a known-good build log—the load-bearing reproducibility claim is not currently supported by the evidence in the paper. The manuscript should either provide a completed assembly manual and an itemized, exact BOM, or the claims should be narrowed to describe a documented design snapshot rather than a complete reproducible build.","section":"Section 2 and Section 5.1"},{"comment":"The '<2500 USD' materials cost claim has no auditable basis: the BOM is a Google Spreadsheet but no itemized cost breakdown or per-part pricing is given, and Table 2 lists quantities only as 'Many', 'Several', or 'About 2 kgs'. Since low cost is a central advertised property of the robot, the paper should include a concrete cost table (with unit prices and totals, or a date-stamped price list from the referenced suppliers). Without this, the low-cost claim is not verifiable.","section":"Section 1 and Section 5.1/Table 2"},{"comment":"The arm lifting specification in Table 1 ('<1kg when fully extended') appears inconsistent with the statement in §5.5 that experiments lifted 'more than 3.5kg at 40cm' and that a motor can 'lift about 9kgs at 70cm' as a theoretical upper bound. The paper should reconcile these numbers by explicitly giving the torque-vs-distance relationship (e.g., a two-point torque estimate) and clarifying which value is measured, which is theoretical, and how the spring assistance affects the specification. Also, the torque calculation of '91 N.cm' at a 70 cm arm length has a unit error: 6370 N.cm divided by 70 cm is 91 N, not 91 N.cm.","section":"Section 3 and Section 5.5"}],"minor_comments":[{"comment":"There is a grammatical typo: 'it means that we can about 7.5 complete rotations in 1 second' should read 'we can make about 7.5 complete rotations in one second.' The following speed calculation is otherwise consistent.","section":"Section 5.5"},{"comment":"The entry 'About 2 kgs' for PLA is informally worded; use 'approximately 2 kg' for consistency with technical writing.","section":"Table 2"},{"comment":"These figures contain multiple panels (lateral/frontal view; design/implementation) but the sub-panels are not labeled (a), (b), etc., which makes it hard to reference specific views in the text.","section":"Figure 4 and Figure 8"},{"comment":"The code example contains a hard-coded timeout 'NUMSECONDSTO WAITFORCONNECTION' and a 'Game over' message in strings; these are acceptable in source code but should be formatted consistently in the paper (e.g., as a listing with proper font).","section":"Section 9.2.9"}],"recommendation":"major_revision","confidential_remarks":"The paper is essentially a hardware project report without experimental evaluation. If the journal typically expects validated systems or comparisons with prior open-source robots, the fit is marginal. The main revision path is to either complete and verify the documentation so that the reproducibility claim holds, or to explicitly rescope the paper as an 'open-source design description with documentation in progress.' The torque/speed calculations are a strength and should be kept, but they need unit corrections and a clear statement of measured vs. computed values."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a project documentation preprint, not a research paper. It describes an open-source robot (tracked base, two 7-DOF arms, head) and points to the online CAD, firmware, and libraries. There is no hypothesis, no evaluation, and no dataset. The real contribution is the repository, and the paper works best as an index to that repository.\n\nWhat is actually new: the specific Jenny 5 design and source files are not in prior cited work, so in a narrow sense it is a new open-source platform. The architecture is conventional, and the paper makes no claim to a new method. The documentation is honest: the author flags that some specifications are based on experience, gives a plausible torque calculation, and reports a real lift test (3.5 kg at 40 cm) with a video link. The MIT-licensed source release is genuinely open, and that is worth crediting.\n\nSoft spots: the load-bearing claim is that a third party can rebuild this from the repository. The paper itself undercuts that. Section 2 says the assembly manual is 'under construction and will not be completed until the robot is completed'; Table 2 says quantities are 'not always exact' and points to that same unavailable manual for details. There is no wiring diagram, no calibration procedure, no known-good build log, and the '<2500 USD' cost claim has no itemized basis. So 'fully open-source' is true in the license/source sense, but the stronger engineering claim is not demonstrated. That is not fatal for a project snapshot, but it limits the paper's value as documentation until the manual exists.\n\nBottom line: for a robotics researcher, this is not a scientific result; it is a pointer to an open-source project. If you want to build or adapt a hobby/research platform, it is worth a look. As a paper, it is not a candidate for serious peer review in a research venue. I would send it to a workshop or a systems/hardware short-paper track rather than a main track, and I would not require revisions beyond adding the missing documentation before treating it as reproducible.","headline":"A transparent build report for an open-source hobby/research robot; useful as documentation, but not a research paper, and the reproducibility claim is unverified.","tokens_in":16079,"tokens_out":2912,"would_cite":false,"duration_ms":26884,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper reports a working, fully open-source almost-humanoid robot and claims its complete build data is public.","keywords":["open-source robot","humanoid robot","tracked mobile platform","robotic arm","stepper motor","Arduino firmware","computer vision","OpenSCAD"],"falsifier":"Attempt a from-scratch rebuild using only the posted files and listed materials; if a builder cannot assemble a working tracked robot with two independently movable 7-DOF arms, a compressible leg, and a responsive head without consulting the designer, the core claim would be falsified. A quicker check is to load the CAD project and verify that every stated degree of freedom, including the leg, shoulder, elbow, gripper, and head, appears as a movable joint in the rendered model.","tokens_in":15251,"feed_emoji":"🤖","tokens_out":6130,"duration_ms":60790,"temperature":0.7,"pith_summary":"This paper reports that Jenny 5, a 40 kg tracked robot with two 7-degree-of-freedom arms, a compressible leg, and a two-axis head, is fully open source: the CAD files, electronics designs, Arduino firmware, C++ libraries, and web client are all publicly released under the MIT license. It argues that a capable, almost-humanoid research robot can be built from hobbyist tools and commonly available materials for under $2500, excluding the onboard computer. The text documents the mechanical design, electronics, control software, and two working vision behaviors (face centering and person following). For a researcher or hobbyist, the value would be a complete reference implementation that can be modified, rebuilt, or used as a starting point for robot manipulation and mobile robotics experiments.","feed_headline":"A complete low-cost humanoid robot is now open source","feed_subtitle":"CAD, firmware, electronics, and a bill of materials are all public so anyone can build or modify the robot.","key_machinery":"The central object is Jenny 5 v3 itself, an almost-humanoid robot whose design is meant to be the reusable artifact. The load-bearing mechanism is the complete open-source hardware and software stack: parameterized geometry files that let joint angles be edited and re-rendered; the asynchronous Scufy firmware, which turns each motor and sensor into a command-response controller over serial; the Scufy Lib C++ library that wraps those firmware messages in events; and the motor-control boards with their library for the platform and leg. Together these components make every degree of freedom addressable by software, which is what lets the paper's claims about rebuildability and programmability be tested by a third party.","core_discovery":"On its own terms, the paper establishes that the current iteration (v3) of Jenny 5 is a working open-source robot with a rubber-tracked platform driven by two DC gearmotors, a leg powered by two 750 N linear actuators, two 7-DOF arms with magnetic rotary encoders at each joint, grippers, a two-axis head with a camera and ultrasonic sensor, and a laptop acting as the brain. It claims the complete design is freely available: parameterized CAD models, custom PCB layouts, the Scufy Arduino firmware and the Scufy Lib PC library, a WebSocket server with an HTML5 client that can be driven by a smartphone's gyroscope or voice, and a bill of materials. The paper also reports measured performance, such as a roughly 11-second full up-down arm move and the ability to lift over 3.5 kg at 40 cm, and identifies stability and back-drive limitations that currently constrain the design.","pith_inferences":["If the files are as complete as claimed, Jenny 5 could serve as a reproducible baseline platform in robotics education, letting many labs start from identical hardware; the paper does not itself draw this conclusion.","Because the CAD geometry is parameterized in terms of joint angles, one could generate the robot's configuration space as a set of rendered poses and use those images for simulation-to-real transfer; this is an extension beyond the text.","The documented trade-off between gearbox reduction, back-drive resistance, and speed suggests a testable design rule: replacing the 50:1 gearboxes with 27:1 ones should reduce actuation time while preserving similar torque limits, a change the paper lists as planned future work.","The stated instability when the leg is extended implies that adding a controllable center-of-mass shift would be a natural next step toward faster and safer locomotion; the paper lists body movement as a future task."],"forward_implications":["A third party with access to the public files and the listed materials can assemble a 40 kg tracked robot with two 7-DOF arms and a head for under $2500, excluding the onboard computer.","Each layer, CAD, firmware, PC library, and web client, can be reused independently; the Scufy firmware in particular is a general stepper, servo, and sensor controller rather than a Jenny 5-specific program.","The measured arm performance, lifting more than 3.5 kg at 40 cm and taking about 11 seconds for a full up-down sweep, gives a concrete baseline for robotic-arm design comparisons.","The face-centering and person-following behaviors demonstrate that the robot can execute closed-loop, vision-guided motion using its existing sensors and webcams."],"supporting_citations":[{"why":"Project website that indexes the CAD, assembly manual, and bill of materials.","marker":"[1]"},{"why":"Public source-code organization where all repositories are stored.","marker":"[2]"},{"why":"Scufy Arduino firmware implementing the command set for motors and sensors.","marker":"[3]"},{"why":"Scufy Lib C++ library that turns firmware responses into events on the PC side.","marker":"[4]"},{"why":"RoboClaw control library used for the platform and leg motor drivers.","marker":"[5]"},{"why":"OpenSCAD, the CAD environment in which the parameterized robot geometry is written.","marker":"[7]"},{"why":"RoboClaw motor-controller boards that drive the platform and leg actuators.","marker":"[14]"},{"why":"AS5147 magnetic rotary encoders that give joint positions for the arms and head.","marker":"[16]"},{"why":"Stepper-motor torque specifications used in the arm lift-capacity calculation.","marker":"[24]"}],"fun_headline_variants":["Jenny 5: open-source humanoid with complete design files","Build your own humanoid: Jenny 5's CAD and code are public","Open-source robot Jenny 5 offers full build details","Jenny 5: a humanoid robot you can assemble from open files"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the publicly released CAD files, electronics, firmware, and bill of materials are complete and consistent enough for someone else to rebuild the robot; the paper's own note that the assembly manual is still under construction and that some quantities are approximate makes this the fragile point.","fun_headline_variants_meta":{"raw":{"variants":["Jenny 5: open-source humanoid with complete design files","Build your own humanoid: Jenny 5's CAD and code are public","Open-source robot Jenny 5 offers full build details","Jenny 5: a humanoid robot you can assemble from open files"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000193,"raw_usage":{"total_tokens":1287,"prompt_tokens":817,"completion_tokens":470,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":433,"completion_tokens_details":{"reasoning_tokens":394}},"tokens_in":433,"tokens_out":470,"duration_ms":4897,"temperature":1.0,"reasoning_tokens":394,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:00:54.426444+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Attempt a from-scratch rebuild using only the posted files and listed materials; if a builder cannot assemble a working tracked robot with two independently movable 7-DOF arms, a compressible leg, and a responsive head without consulting the designer, the core claim would be falsified. A quicker check is to load the CAD project and verify that every stated degree of freedom, including the leg, shoulder, elbow, gripper, and head, appears as a movable joint in the rendered model.","supporting_citations":[{"cited_title":"github.io, last accessed on 2019.06.10","cited_arxiv_id":null,"evidence_quote":"Project website that indexes the CAD, assembly manual, and bill of materials."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Public source-code organization where all repositories are stored."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Scufy Arduino firmware implementing the command set for motors and sensors."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Scufy Lib C++ library that turns firmware responses into events on the PC side."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"RoboClaw control library used for the platform and leg motor drivers."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"OpenSCAD, the CAD environment in which the parameterized robot geometry is written."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"RoboClaw motor-controller boards that drive the platform and leg actuators."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"AS5147 magnetic rotary encoders that give joint positions for the arms and head."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Stepper-motor torque specifications used in the arm lift-capacity calculation."}],"review_version":1}