EpiPad: Enabling Multi-Finger and Force-Sensitive Skin Interface for Smartwatches

With the growing popularity of smart wrist-worn devices, their small touchscreen size fundamentally limits the expressiveness of user input. Although various interaction methods have been proposed to extend input capabilities, few provide a seamless transition from traditional touchscreens and touchpads. To address this challenge, we present EpiPad, which for the first time enables multi-finger and force-sensitive input for smartwatches through mmWave-enabled on-skin touch interfaces. Leveraging a compact wrist-mounted mmWave array, EpiPad transforms the forearm's skin surface into an extended input area that preserves familiar gestures such as tapping and sliding, while enriching them with multi-finger and force interactions common in commodity touchpads. To achieve stable force estimation, we develop a novel signal extraction method based on mmWave phase dynamics and implement a recognition pipeline composed of lightweight machine learning models. Experiments with 20 users show that, after a short one-time calibration of 3 minutes 54 seconds, EpiPad achieves a gesture recognition accuracy of 92.8%, force detection accuracy of 90.2%, and ranging precision of 10.2 mm on unseen users. A real-time evaluation validates the system's reliability for real-world deployment, and a follow-up survey further confirms that EpiPad offers a low learning cost, rich gesture diversity, and an input experience closely aligned with traditional touch interfaces.