{"id":"3c37a0ee-b6e3-4f28-999f-90fb0c452d54","arxiv_id":"2506.13704","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A unified haptic-guided teleoperation framework uses one leader arm to control both a mobile base and its manipulator, improving trajectory adherence and task speed in 20-user trials.","lead":"This paper presents a system where a single haptic-feedback robotic arm pilots both a mobile robot and its onboard manipulator arm, switching between navigation and grasping automatically. In a 20-person user study, haptic guidance reduced trajectory deviations and manipulation times, though the 'no cognitive load increase' claim is only weakly supported.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Primary 'accuracy' metric is circular: Eq. 6 generates haptic force from the same DWA reference trajectory used to measure deviation, so reduced deviation reflects guidance compliance, not task accuracy; no independent outcome metric is reported.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the primary accuracy metric is coupled to the haptic intervention. I agree with that assessment and would not change the CONDITIONAL verdict. I also considered the unsupported cognitive-load claim and the suspicious repetition of z = 2.012 for two different tests; these are real issues, but the circular accuracy metric is the most load-bearing because it directly undermines the headline claim of improved task accuracy. Even if the statistics were fully corrected, the conclusion would still fail without an outcome metric independent of the guidance trajectory. The proposed test would settle this by checking whether haptic guidance improves objective task performance, not merely adherence to a potentially suboptimal reference. Until such evidence is provided, the paper should remain conditional rather than fully accepted.","tokens_in":9413,"tokens_out":3674,"duration_ms":41877,"concrete_test":"Recompute the accuracy comparison using only logged task-outcome metrics that do not involve the DWA reference trajectory: grasp success rate, number of collisions, and end-effector-to-object alignment error at the moment control switches to manipulation, comparing Conditions 1 and 2. If haptic guidance does not significantly improve these independent metrics, the abstract's accuracy claim should be revised from 'task accuracy' to 'trajectory adherence to a reference.'","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's main accuracy result is that 'deviations along the y-axis were significantly reduced with haptic guidance (p < 0.05, z = 2.012)' during tele-manipulation. This deviation is computed relative to the DWA reference trajectory, and that same trajectory is the input to the haptic force in Eq. 6: F_fmr = K_fmr(P_fmr^(t+40) - P_fmr). The force therefore actively pushes the operator toward the reference, so reduced deviation primarily confirms that the force was applied, not that the task was executed more accurately. The paper itself states that the initial reference trajectory is 'often suboptimal' (Section V, Experiment Design), so adherence to it cannot be equated with task quality. No independent task-outcome metric is reported: there is no grasp success rate, no collision count, no end-effector placement error at grasp time, and no final object-placement error. Consequently, the central claim that the framework 'significantly improves task accuracy' is not supported by the evidence as presented, even if the reported z-statistics are correct.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents HARMONI, a teleoperation framework in which a single 7-DoF leader arm (Franka Panda) controls both a mobile base (Hunter 2.0) and a follower arm mounted on it, with haptic guidance derived from DWA-planned reference trajectories. The system unifies tele-navigation and tele-manipulation with autonomous mode switching, and the authors report a user study with 20 participants under real-world conditions including known, semi-known, and unknown obstacles and visual/audio distractors. The central claims are that haptic guidance significantly improves task accuracy (reduced y-deviation) and efficiency (reduced manipulation time) without increasing cognitive load. The paper also reports heart-rate measurements and a mental-rotation correlation analysis.","tokens_in":9651,"tokens_out":4750,"duration_ms":50738,"significance":"The engineering contribution is substantial: integrating navigation and manipulation control through a single leader arm with haptic feedback addresses a real operational bottleneck in mobile manipulation teleoperation, and the real-robot user study with distractors goes beyond simulation-only evaluations. If the evaluation were supported by independent task-outcome metrics and a validated workload measure, the paper would be a useful addition to the shared-control literature. The current evidence, however, is weakened by the coupling between the guidance law and the primary accuracy metric, and by the absence of any direct workload assessment.","major_comments":[{"comment":"The primary accuracy metric is deviation from the DWA reference trajectory, but the haptic force in Eq. (6) is computed from that same reference trajectory: F_fmr = K_fmr(P_fmr^(t+40) - P_fmr). The paper itself states that the initial reference trajectory is \"often suboptimal\" (Section V, Experiment Design). Reduced deviation therefore likely reflects operator compliance with the haptic cue rather than improved task accuracy. No independent task outcome is reported (no grasp success rate, collision count, end-effector placement error, or final object placement error). The abstract's claim that the framework \"significantly improves task accuracy\" is thus not supported by the presented evidence. Please report task-level outcome metrics or reframe the claim as improved trajectory adherence to the planner's guidance.","section":"Section V, Force Cue Processing; Eq. (6)"},{"comment":"The abstract claims the framework improves performance \"without increasing cognitive load,\" but no validated workload measure (e.g., NASA-TLX or equivalent) is used. Heart rate (HR) is a physiological stress indicator, not a direct cognitive-load measure, and the mental-rotation correlation rs(18) = -0.055 is unrelated to workload. The conclusion's more modest wording \"without increasing operator stress\" is better supported, but even that rests only on HR. Either add a validated workload assessment or remove the cognitive-load claim from the abstract.","section":"Abstract; Section V, Stress Response"},{"comment":"The statistical reporting is incomplete and internally inconsistent. No descriptive statistics (means, standard deviations) or effect sizes are given for the main significant results; the text reports only p-values and z-values. Table I is hard to interpret: the columns are labeled \"Condition 1\" and \"Condition 2\" while the text says the test compares Condition 3 (haptic guidance) against Condition 2 (no haptic guidance), and several cells contain only p-values without indicating which condition favored. Additionally, the text reports \"Analysis of 236 trials\" and \"11 of 18 participants\" for Condition 3, which does not reconcile with the stated 20 participants and four trials per condition. These issues make it difficult to assess the reliability and generality of the reported effects.","section":"Section V, Results; Table I"}],"minor_comments":[{"comment":"The term τT appears in Eq. (1) but is never defined; please clarify its meaning.","section":"Eq. (1)"},{"comment":"The PD control law in Eq. (2) uses +Kd·qfra_dot, which is an unusual sign for a damping term; verify it is not a typo and describe how the derivative term is computed.","section":"Eq. (2)"},{"comment":"The sentence \"Participants operated the LRA using only visual feedback from provided screens, without direct line of sight to\" is incomplete and should be finished.","section":"Section V, Task Setup"},{"comment":"The sentence \"This study, approved by the University of Lincoln's ethics committee 1\" is a sentence fragment; please integrate the ethics reference into a complete sentence.","section":"Section IV"},{"comment":"The notation states \"All poses P ∈ R6\" but Eq. (6) uses γ for orientation; please define the orientation representation consistently throughout.","section":"Section III"},{"comment":"The layout and column headings of Table I should be redesigned to make clear which conditions are being compared and what each p-value refers to; the current presentation is confusing.","section":"Table I"},{"comment":"The haptic force in Eq. (6) has nonzero components only along x and γ, yet the reported significant improvement is in y-deviation; the causal mechanism by which the x/γ force reduces lateral deviation should be explained in the text.","section":"Section V, Force Cue Processing"}],"recommendation":"major_revision","confidential_remarks":"The engineering platform is promising and the real-robot study is a strength, but the current evaluation does not support the headline accuracy and cognitive-load claims. I would ask the authors to either supply independent task-outcome metrics (which may require re-analysis of logged data or a new experiment) and a validated workload measure, or substantially temper the claims. The inconsistencies in trial counts and Table I should also be resolved before resubmission."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the genuinely new thing here is using a single 7-DoF leader arm for both tele-navigation and tele-manipulation with autonomous switching, and that is a legitimate, if incremental, integration. The cited prior work mostly uses separate devices for the two tasks, so the paper fills a real gap. The 20-participant real-robot study is also more than we usually see for this kind of system, with counterbalancing, distraction conditions, and physiological monitoring.\n\nWhat the paper does well: the control architecture is clearly laid out, the switching mechanism is described concretely, and the evaluation is reasonably careful for an engineering paper. It does not oversell the navigation-time result—it explicitly reports no significant time benefit there. The null mental-rotation correlation is a decent inclusivity result.\n\nWhere it gets soft: the main accuracy result is circular. Haptic force in Eq. 6 is computed from the DWA reference trajectory, and accuracy is measured as deviation from that same trajectory. So reduced deviation mostly confirms the force was applied, not that the task was done better. The paper even admits the reference trajectory is often suboptimal (Section V), which makes the circularity worse: obeying a suboptimal reference is not task quality. No independent outcome metric appears—no grasp success rate, no placement error, no collision count. The abstract's 'significantly improves task accuracy' is therefore not supported as written.\n\nThe 'without increasing cognitive load' claim is similarly unsupported. Heart rate and mental rotation are not a validated workload measure. HR was lower for 13 of 20 participants; that is a trend, not a firm result. Statistical reporting is thin too: repeated z = 2.012, missing entries in Table I, no effect sizes or multiple-comparison corrections, and no data or code to check the trajectories.\n\nNet: the integration is a solid engineering contribution and the study is a reasonable first evaluation, but the evidence needs strengthening before the broader claims can stand. A revised abstract, a separation of guidance compliance from task accuracy, and an independent outcome metric would fix most of the problem.\n\nRecommendation: worth sending to peer review. A serious referee can ask for achievable revisions rather than a rewrite.","headline":"Useful integration and real-robot study, but the main accuracy metric is partly self-referential and the cognitive-load claim outruns the data.","tokens_in":10209,"tokens_out":1646,"would_cite":true,"duration_ms":20091,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A single haptic-guided leader arm can steer both a mobile robot and its manipulator, and users perform better with it.","keywords":["haptic guidance","shared control","tele-manipulation","tele-navigation","mobile manipulator","virtual fixtures","user study","cognitive load"],"falsifier":"Compare haptic guidance against a condition that shows the same reference trajectory on screen as a visual path but gives no force feedback; if visual-only guidance produces the same reduction in deviation and task time, the haptic channel is not the cause. Alternatively, measure an outcome independent of the reference, such as grasp success rate or placement accuracy, and see whether haptic guidance improves it.","tokens_in":9226,"feed_emoji":"🕹️","tokens_out":3495,"duration_ms":31084,"temperature":0.7,"pith_summary":"This paper argues that one 7-DoF leader arm can serve as a haptic interface for both tele-navigation and tele-manipulation of a mobile manipulator, replacing the separate controllers typically used for each task. In a user study with 20 participants under real-world conditions, haptic guidance significantly reduced trajectory deviation and task duration during tele-manipulation, while most participants showed lower heart rates, suggesting reduced stress. The authors claim this unified haptic shared control improves operator accuracy and efficiency without increasing cognitive load, and they show it remains effective under visual distractions. If true, this would simplify teleoperation hardware and training for complex mobile manipulation tasks in hazardous or remote environments.","feed_headline":"One haptic arm guides both robot navigation and grasping","feed_subtitle":"20-person trial shows haptic feedback cuts manipulation errors and time without raising stress.","key_machinery":"The central mechanism is a unified leader-follower control law with a binary mode switch: equation 1 selects torque contributions from either the follower arm (manipulation) or the mobile robot (navigation), with null-space damping for safety. Navigation uses a velocity mapping from leader end-effector displacement, and haptic force is generated by a virtual fixture pulling the leader toward the DWA-planned reference pose (equation 6); manipulation uses PD torque mirroring of joint positions. This lets one 7-DoF arm provide both 2-DoF navigation guidance and 6-DoF manipulation guidance with autonomous mode switching.","core_discovery":"The central claim is that haptic-guided shared control, delivered through one leader arm kinematically similar to the follower arm, improves operator accuracy and efficiency during tele-manipulation of a mobile manipulator while not increasing cognitive load. The system computes haptic forces from the DWA planner's reference trajectory for navigation and uses impedance control to guide the operator's hand; when the object becomes graspable, control switches automatically to manipulation, with the follower arm mirroring the leader's joint positions. In a 20-participant study with audio and visual distractions, haptic guidance significantly reduced y-axis deviation during tele-manipulation (p < 0.05) and decreased task duration (p < 0.05), and 13 of 20 participants had lower heart rates, indicating reduced stress.","pith_inferences":["The accuracy metric is coupled to the guidance law, so the reported improvement may partly measure compliance with the planner rather than task success; an outcome-based metric would separate these.","The single-leader design may trade off simultaneous control, since the operator cannot adjust the base and the arm at the same time, which could matter for tasks needing coordinated motion.","The benefit might generalize to other leader devices and dynamic environments, but the static-map setup means the results do not yet establish that.","A direct comparison against dual-device control (joystick plus arm) would quantify how much of the gain comes from unification versus haptic guidance alone."],"forward_implications":["Teleoperators of mobile manipulators could use a single leader arm instead of separate joystick and arm controllers, reducing equipment and training overhead.","Haptic cues are most beneficial for precise manipulation subtasks, where they cut errors and time; navigation benefits less significantly.","The system's autonomous navigation-to-manipulation switch offloads graspability decisions from the operator, letting them focus on alignment.","Guidance helps operators maintain smoother trajectories even under visual distraction and reduces physiological stress for most users.","The framework is usable by operators with diverse mental-rotation abilities, suggesting it does not require strong spatial skills."],"supporting_citations":[{"why":"Shows that a 7-DoF identical twin master can teleoperate a 7-DoF follower arm, providing the kinematic-similarity precedent for the leader-follower design.","marker":"[30]"},{"why":"Haptic-guided grasping to minimise torque effort, used as the basis for force feedback in the tele-manipulation control.","marker":"[7]"},{"why":"Human-in-the-loop optimisation for grasping informs the haptic cue generation and shared-control approach for manipulation.","marker":"[8]"},{"why":"Haptic-guided shared control grasping for collision-free manipulation supports the guidance strategy used for the follower arm.","marker":"[9]"},{"why":"Motion scaling and haptic guidance effects on workload are the basis for evaluating operator load and performance in this study.","marker":"[11]"},{"why":"Shared planning and control for mobile robots with haptic feedback is the foundation for the navigation guidance mechanism.","marker":"[28]"},{"why":"Automatic switching teleoperation using virtual fixtures provides the concept for the autonomous navigation-to-manipulation mode switch.","marker":"[20]"},{"why":"Dual-user haptic teleoperation of a mobile manipulator is the alternative approach the paper contrasts with, motivating the unified single-leader design.","marker":"[31]"}],"fun_headline_variants":["Unified haptic control improves teleop precision and speed","Single haptic arm steers robot and grasp actions","Haptic guidance unifies tele-navigation and manipulation","Haptic arm guides navigation and grasping in one system","Single haptic controller boosts teleop accuracy without stress"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The measure of accuracy is deviation from a reference trajectory that the paper itself calls \"often suboptimal,\" so lower deviation may show that operators follow haptic cues rather than that the task is done better.","fun_headline_variants_meta":{"raw":{"variants":["Unified haptic control improves teleop precision and speed","Single haptic arm steers robot and grasp actions","Haptic guidance unifies tele-navigation and manipulation","Haptic arm guides navigation and grasping in one system","Single haptic controller boosts teleop accuracy without stress"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001039,"raw_usage":{"total_tokens":4329,"prompt_tokens":858,"completion_tokens":3471,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":474,"completion_tokens_details":{"reasoning_tokens":3394}},"tokens_in":474,"tokens_out":3471,"duration_ms":22500,"temperature":1.0,"reasoning_tokens":3394,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T00:26:43.229380+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare haptic guidance against a condition that shows the same reference trajectory on screen as a visual path but gives no force feedback; if visual-only guidance produces the same reduction in deviation and task time, the haptic channel is not the cause. Alternatively, measure an outcome independent of the reference, such as grasp success rate or placement accuracy, and see whether haptic guidance improves it.","supporting_citations":[{"cited_title":"Haptic-guided teleoperation of a 7-dof collabo- rative robot arm with an identical twin master,","cited_arxiv_id":null,"evidence_quote":"Shows that a 7-DoF identical twin master can teleoperate a 7-DoF follower arm, providing the kinematic-similarity precedent for the leader-follower design."},{"cited_title":"Haptic-guided grasping to minimise torque effort during robotic telemanipulation,","cited_arxiv_id":null,"evidence_quote":"Haptic-guided grasping to minimise torque effort, used as the basis for force feedback in the tele-manipulation control."},{"cited_title":"Human-in-the-loop optimisation: Mixed initiative grasping for optimally facilitating post-grasp manipulative actions,","cited_arxiv_id":null,"evidence_quote":"Human-in-the-loop optimisation for grasping informs the haptic cue generation and shared-control approach for manipulation."},{"cited_title":"Haptic-guided shared control grasping: collision- free manipulation,","cited_arxiv_id":null,"evidence_quote":"Haptic-guided shared control grasping for collision-free manipulation supports the guidance strategy used for the follower arm."},{"cited_title":"The impact of motion scaling and haptic guidance on operators’ workload and performance in teleoperation,","cited_arxiv_id":null,"evidence_quote":"Motion scaling and haptic guidance effects on workload are the basis for evaluating operator load and performance in this study."},{"cited_title":"Shared planning and control for mobile robots with integral haptic feedback,","cited_arxiv_id":null,"evidence_quote":"Shared planning and control for mobile robots with haptic feedback is the foundation for the navigation guidance mechanism."},{"cited_title":"An automatic switching approach to teleopera- tion of mobile-manipulator systems using virtual fixtures,","cited_arxiv_id":null,"evidence_quote":"Automatic switching teleoperation using virtual fixtures provides the concept for the autonomous navigation-to-manipulation mode switch."},{"cited_title":"Dual-user haptic teleoperation of complementary motions of a redundant wheeled mobile manipulator considering task priority,","cited_arxiv_id":null,"evidence_quote":"Dual-user haptic teleoperation of a mobile manipulator is the alternative approach the paper contrasts with, motivating the unified single-leader design."}],"review_version":1}