UltraGelBot: Autonomous Gel Dispenser for Robotic Ultrasound

Robotic Ultrasound Systems (RUSs) have gained in- creasing prominence in recent years to alleviate prob- lems concerning operator-dependability in free-hand ul- trasound examinations [1]. These robotic systems op- erate either in teleoperated [2] or autonomous [3], [4], [5] mode. These systems can increase ultrasound ac- cessibility in underserved regions and ensure the safety of the operator during the COVID-19 pandemic [2], [6]. However, in these robotic systems, the gel is still applied by the human attendant. This human intervention between the procedure results in the acquisition of sub- optimal images due to inappropriate acoustic coupling [7]. Further, the procedure time also increases as RUS needs to be halted several times in between the proce- dure for manual gel application. Moreover, the presence of humans near the patient’s vicinity did not achieve the complete safety of the operator, as promised by RUS. Thus, an automated method for dispensing of ultrasound gel is the immediate requirement of RUS. In this paper, we proposed a new end-of-arm tool for RUS, referred to as ‘UltraGelBot’. This robot can detect and dispense the gel autonomously. A deep learning model is developed to detect the gel region from real- time images acquired using an onboard camera. A mo- torized mechanism is developed, which uses this feed- back and dispenses the gel. In addition, sonographers often utilize multiple ultrasound probes, such as linear, curvilinear, etc., to visualize different anatomies up to a specific depth. The existing gripping methods for RUS are specific to the design profiles of the probe. Thus, this gripper features a modular mechanism to grip ultrasound probes with varying profiles. This mechanism will also allow the operator to easily change different probes during the procedure without using mechanical tools. The performance of the UltraGelBot was compared to that of manual gel application for ultrasound of artery in human’s forearm. In addition, for assessment of user experience, the NASA Task Load Index (TLX) was used.