cs.CROct 19, 2026

PhantomTac: Hijacking Robotic Tactile Systems with Deploy-and-forget Tags

Authors: Binbin Zhang, Yifeng Zhang, Xiaomeng Chen, Dongyao Chen

Organizations: Shanghai Jiao Tong University Shanghai, China · Singapore Management University Singapore

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Tactile perception is crucial for precise robotic manipulation, providing an important foundation for dexterous, closed-loop force control. Among various tactile sensing techniques, magnetic tactile sensors stand out thanks to their fine-grained, three-dimensional force perception capability and strong robustness against visual occlusion. In this study, we discover a physical loophole called PhantomTac in magnetic tactile sensors. The attacker can alter the magnetic field, thus spoofing the force interpreted by magnetic tactile sensors. The attacker can stealthily deploy such an attack by simply attaching a thin (the thickness is around 0.5 mm), battery-free magnetic tag to the surface of an object. Hence, the attack tag can be easily concealed by a barcode sticker. Such an attack scheme is "deploy-and-forget" -- the attacker can deploy the tag and leave. When a robotic hand is picking up an object, it may severely damage the object. Specifically, spoofing the normal force can crush the object, whereas spoofing the shear force can lead to a loose grip, thus dropping the object. Compared with directly sabotaging the robotic hand, PhantomTac is much more concealed as the attacker does not need access to the device. We performed extensive reverse-engineering efforts to characterize the sensing mechanism of commercial-off-the-shelf tactile sensors and proposed a practical tag design. End-to-end evaluations conducted on real robots and robotic arms demonstrate that our attack can bypass existing anti-interference mechanisms and effectively hijack their closed-loop control processes, achieving an attack success rate exceeding 95% when the grasping-position offset is bounded within 8 mm. Finally, based on the limitations of existing anti-magnetic interference mechanisms, we propose a multimodal-consistency defense mechanism to mitigate such physical-layer security threats.

Figures & tables

Figure 1. Principle and structure of magnetic tactile sensors. (a) demonstrates the neutral, compressed, and shear states. (b) Structure of a COTS magnetic tactile sensor.
MetricOpticalCapacitiveResistiveMagnetic
Fine-grained force sensing\ding 51\ding 55\ding 55\ding 51
Easy multi-dimensional force decoupling\ding 51\ding 51\ding 55\ding 51
Low cost & overheads\ding 55\ding 55\ding 51\ding 51
Robust static & dynamic sensing\ding 51\ding 51\ding 55\ding 51
Table 1. Comparison of tactile sensing modalities.
Figure 2. Experiment on Anti-Interference Mechanisms of Magnetic Tactile Sensors. The platform for the sample is 10 mm away from the bottom surface of the sensor. (a) presents the placement of the MLX90393 Module; whereas (b)—(e) denote the placement of various magnetic objects. Overview of existing robotic tactile sensing technologies and their underlying sensing principles.
Figure 3. Placement of PaXini and XELA Sensors. (a) XELA uSPa 46. (b) PX6AX-GEN3-MC-M2020-Elite. Example of a robotic tactile sensing system showing the sensor structure and the physical interaction process during contact.
Figure 4. Comparison of magnetic field intensity between common magnetic objects in daily life and small magnets. Illustration of magnetic tactile sensing mechanisms showing the relationship between magnetic fields and tactile sensor responses.
Figure 5. Comparison of interference on magnetic tactile sensors caused by different objects. Comparison of different tactile sensing modalities highlighting the characteristics and advantages of magnetic tactile sensors.
Figure 6. An overview of PhantomTac ’s attack model. Overview of the PhantomTac attack framework showing the attacker, deploy-and-forget magnetic tags, robotic tactile sensors, and the resulting manipulation of robotic perception.
Figure 7. Setup for the reverse engineering process. Logical diagram illustrating the reverse engineering process of PhantomTac, including the magnetic tactile sensor, measurement setup, and the procedure used to characterize sensor responses for constructing the attack.
Figure 8. Three-dimensional force feedback of magnetic tactile sensor. The diameter of the magnetic interference source is 3 mm. Three-dimensional force feedback visualization of a magnetic tactile sensor, showing the sensor response under magnetic interference from a 3 mm diameter magnetic source.
Label TypeCommon ApplicationsThickness (mm)
Thermal Paper ( Avery Dennison, 2021 )Express shipping labels0.142±10%0.142\pm 10\%
Coated Paper ( Avery Dennison, 2024 )Product barcodes0.183±10%0.183\pm 10\%
Synthetic Paper ( Avery Dennison, 2010 )Waterproof QR code labels0.145±10%0.145\pm 10\%
PET Label ( Avery Dennison, 2022 )Electronic product back labels0.115±10%0.115\pm 10\%
Table 2. Common Sticker Thicknesses.
Figure 9. The penetration capability of 2.5 mm magnet. Experimental evaluation of the penetration capability of a 2.5 mm diameter magnet, showing the effect of magnetic interference through different distances or material layers on the tactile sensing response.
Figure 10. Disguised appearance of PhantomTac . (a)—(e) present QR code, blank paper, shipping label, cartoon sticker, and product sticker, respectively. Examples of disguised appearances of PhantomTac tags. The subfigures show the attack device embedded or concealed as a QR code, a blank paper label, a shipping label, a cartoon sticker, and a product sticker.
Figure 11. Blind test results with magnets of different diameters. (a) Blind test results of 3 mm diameter magnets. (b) Blind test results of 2.5 mm diameter magnets.
Figure 12. Fabricate attack tags based on the contour comparison of magnetic tactile sensor feedback force. Fabrication process of PhantomTac attack tags based on contour comparison of magnetic tactile sensor force-feedback patterns, showing how sensor response contours are analyzed and used to construct effective attack tags.
Figure 13. (a) Test setup for robotic grasping (Dofbot-pi4B). Note that the magnetic tactile sensor was mounted on the inner side of the gripper. (b) Test setup for humanoid robot (Unitree G1). Note that the magnetic tactile sensor was mounted on the back of the hand.
Figure 14. End-to-end attack results on the robotic manipulator and the humanoid robot. (a) Positive and negative normal force injection in the grasping task. (b) Positive and negative normal force injection in the collision detection task. Note that S1, S2, and S3 denote the different stages of a benign operation. Sp′1S_{p}^{\prime}1 , Sp′2S_{p}^{\prime}2 , and Sp′3S_{p}^{\prime}3 denote stages under positive force injection, whereas Sn′1S_{n}^{\prime}1 , Sn′2S_{n}^{\prime}2 , and Sn′3S_{n}^{\prime}3 present negative ones.
Figure 15. Result of robustness test for contact position variation. Robustness evaluation results of PhantomTac under contact position variation, showing how changes in the contact location affect the attack performance of the magnetic tactile sensing system.
Figure 16. Generalizability experiment setup. (a) PX6AX-GEN3-DP-S2015-Elite. (b) PX6AX-GEN3-IP-S1610-Elite. Experimental setup for evaluating the generalizability of PhantomTac across different magnetic tactile sensors, showing two sensor models used in the evaluation: PX6AX-GEN3-DP-S2015-Elite and PX6AX-GEN3-IP-S1610-Elite.
More in this topic

Robotic Tactile Sensing

Robotic tactile sensing equips machines to detect contact, force, slip, shape, and material properties for physical interaction.

See all 101 papers

Explore similar work

CardsList
Aug 21, 2026

GhostTac: Manipulating Tactile Sensors without Physical Contact

Tactile sensors are integral components of modern robotic systems, enabling robots to perceive and interact with the physical environment through tactile feedback. Despite their importance, the physical-layer security of tactile sensors has received little attention in prior work. In this paper, we present GhostTac, to the best of our knowledge, the first contactless attack that manipulates tactile sensing via electromagnetic interference (EMI). We identify that EMI exploits the nonlinear rectification and limited bandwidth amplification effects, allowing carefully crafted EMI signals to be converted into a persistent DC offset that bypasses on-board filtering and induces stable measurement deviations. Building on this mechanism, GhostTac enables fine-grained and controllable manipulation of sensor outputs by reshaping the spatial distribution and manipulating the magnitude at the targeted location. Such interference can induce unintended and harmful robot behaviors, such as causing a domestic robot to exert excessive force, resulting in physical damage or human injury. We evaluate GhostTac on 10 sensor modules and 2 dexterous hands, covering 15 tactile sensors of different types, and demonstrate consistent attack effectiveness across all tested devices. We further present three case studies on tactile grasping, slip detection, and material classification to illustrate practical impacts in real robotic tasks. We envision that our findings shed light on a new physical attack vector against tactile sensing in robotic systems.

Kun Wang, Xuancun Lu, Ruochen Zhou +6

Zhejiang University Hangzhou, China · Hong Kong University of Science and Technology Hong Kong, China

Adversarial AttacksTactile Sensing
AskPDF
May 21, 2026

TacO: Benchmarking Tactile Sensors for Object Manipulation

Vision-based learning from demonstrations has achieved remarkable success in enabling robots to perform manipulation tasks and high-level semantic reasoning, yet it remains insufficient for complex, contact-rich manipulation. While there is broad agreement that tactile sensing improves manipulation, there is no empirical guidance on which tactile sensors are best suited for which manipulation tasks. In this paper, we provide a systematic, task-driven evaluation of tactile sensors for robot manipulation and propose a framework for selecting and evaluating sensors based on manipulation policy performance. Separate manipulation policies are trained for tactile sensors of four distinct modalities: visual, acoustic, magnetic, and resistive, across three tasks: pick-and-place with unknown mass, object reorientation, and plug insertion. For each task, an analysis of how sensor properties such as spatial resolution, shear sensing, and tactile representation, and the inherent material friction affect task performances is done. Rather than tactile sensing being universally beneficial in the same way, our results show that the usefulness of tactile information depends strongly on sensor modality, material properties, and the specific manipulation tasks. All of the tactile sensors, code, data, and hardware setup will be publicly available on the project website.

Anya Zorin, Zilin Si, Myungsun Park +11

UC San Diego United States · CMU · UC San Diego +2

Tactile SensingContact-Rich Robotic ManipulationRobot Manipulation BenchmarksTactile Sensing for Robotic Manipulation
AskPDF
Apr 30, 2026

FlexiTac: A Low-Cost, Open-Source, Scalable Tactile Sensing Solution for Robotic Systems

We present FlexiTac, a low-cost, open-source, and scalable piezoresistive tactile sensing solution designed for robotic end-effectors. FlexiTac is a practical "plug-in" module consisting of (i) thin, flexible tactile sensor pads that provide dense tactile signals and (ii) a compact multi-channel readout board that streams synchronized measurements for real-time control and large-scale data collection. FlexiTac pads adopt a sealed three-layer laminate stack (FPC-Velostat-FPC) with electrode patterns directly integrated into flexible printed circuits, substantially improving fabrication throughput and repeatability while maintaining mechanical compliance for deployment on both rigid and soft grippers. The readout electronics use widely available, low-cost components and stream tactile signals to a host computer at 100 Hz via serial communication. Across multiple configurations, including fingertip pads and larger tactile mats, FlexiTac can be mounted on diverse platforms without major mechanical redesign. We further show that FlexiTac supports modern tactile learning pipelines, including 3D visuo-tactile fusion for contact-aware decision making, cross-embodiment skill transfer, and real-to-sim-to-real fine-tuning with GPU-parallel tactile simulation. Our project page is available at https://flexitac.github.io/.

Binghao Huang, Yunzhu Li

Columbia University

Tactile SensingTactile Sensing for Robotic ManipulationRobotic Tactile Sensing
AskPDF