NTT's Paper Accepted for IROS 2026, the World's Highest Conference on Robotics and Intelligent Systems

NTT's paper was accepted for IROS 2026(IEEE/RSJ International Conference on Intelligent Robots and Systems), the world's largest and most influential international conference on robotics and intelligent systems, to be held in Pittsburgh, Pennsylvania, USA from September 27 to October 1. The accepted papers are as follows.

Abbreviated names of the laboratories:
HI: Human Informatics Labs., NTT, Inc.

■Development and Validation of a Wearable Sensing System for Quantifying Skin Softness for Remote Edema Assessment

Koki Ebina(HI), Mitsuhiro Goto(HI), Masato Fukuda(HI), Aya Saitoh(Niigata University), Noriko Yagyuda(Niigata University), Shigekuni Kondo(HI), Sayuri Sakai(Niigata University)

Telemedicine can mitigate the uneven distribution of medical resources, yet remote palpation remains challenging because tactile information is difficult to transmit. This paper investigates human-mediated remote palpation, in which a local proxy, such as a visiting nurse or family member, performs palpation while quantitative signals acquired by wearable sensors are shared remotely. We present a wearable sensing system capable of estimating tissue stiffness. The system measures palpation force and indentation using a nail-mounted force sensor and a depth camera, with infrared (IR) retroreflective marker tracking used for indentation measurement. In calibration experiments, the proposed method achieved a mean absolute percentage error of 18.4% in estimating Young's modulus, a measure of tissue stiffness. As a representative application, the system was evaluated for lower-limb edema assessment using medical training edema models. A machine-learning-based classifier achieved an average accuracy of 72.1% and a quadratic weighted kappa (QWK), a metric for agreement in ordinal classification, of 0.915 in five-level severity classification, demonstrating performance comparable to palpation by nurses. Overall, these results demonstrate that the proposed system enables robust and minimally intrusive quantitative edema assessment, supporting human-centered wearable telemedicine applications.