SHBC1228

J.H.LOW¹, P.M. KHIN¹, Q.Q. HAN¹, C.H. YEOW¹

National University of Singapore¹

Due to the impact of Covid-19, automated food handling system is crucial to ensure hospital kitchens can continue to function with minimal impact from the safe management measures at workplaces where the number of human workers are limited. This project presents a versatile soft robotic gripper system that can handle delicate food items using various programable grip poses.

Our core technologies include customized 3D-printed pneumatic finger actuator, reconfigurable gripper base and object detection model based on deep learning algorithm. The finger actuator was directly 3D-printed with a type of thermoplastic elastomers, and this material allows the proposed gripper to gently grip delicate food such as tofu. The gripper base consists of three different motors which allow the gripper to handle a wide range of dishes using various grip poses: scoop, pinch and claw. An object recognition system was developed to identify food samples.

Bento preparations with five food items (dry noodle, broccoli, tofu, sausage, sliced cake) were conducted successfully with the gripper attached onto a collaborative robot arm. Our system achieved higher grip success rates (93.3%) compared to other grippers (10% – 70%) and do not cause damages to any targeted items. One set of our gripper system can prepare 120 to 150 set of meals with five dishes in one hour.

Our gripper system has been demonstrated to handle various dishes successfully, making it a possible collaborative solution with human workers for meal assembling in hospital kitchens. We are working with WHC and CGH to further evaluate this system.