Abstract:
Human with different conditions such as disease, allergy, gender, ethnicity, or religion need different types of food. Nowadays, people prefer to choose diet based on nutrition value over price such as composition of food or the impact on body, so personalized or customized food could potentially be a big market in few years. With this aspect, the assembly robot becomes a good choice in the food industry to produce the small amount of packed meal with highly adjustable ingredients to match with the requirements. With the automation system, the cooking process is faster, cheaper and safer, yet coming with good traceability in unusual incidents. However, the robot that overall works practically is still in an early stage of development. One issue is a design of the end-effectors that should be interchangeable to match with different types of food. In this thesis, a highly flexible food assembling framework has been realized and evaluated. The robot with designed end-effector could recognize and localize objects to assemble Japanese lunch box. A modular end-effector using soft gripper and vacuum pad in the single unit was designed, prototyped, and installed. The performance has been evaluated by the pick-and-place process with various scenarios. For a single type of food, the system could successfully pick-and-place karaage, onigiri, and bologna into a meal box with a 90% success rate. However, when picking and placing multiple types of food, the success rate is reduced lower than 50%.
