FingerVision Inc. aims to expand the scope of application of robots and machines and solve various social issues through the practical application of "vision-based tactile" technology. We have signed a joint development agreement with Yoshinoya Co. Ltd. to automate and save labor in the dishwashing process at gyudon stores by using robots.
At Yoshinoya, which has approximately 1,200 gyudon stores, employees manually soak dirty dishes after use in water, then take them out one by one, placing them on racks for washing machines, and finally using the dishwasher to clean them. We aim to use robots to automate the process of extracting a wide variety of dirty dishes from this dirty water and racking them.
Yoshinoya’s issue
At Yoshinoya, one employee efficiently manages multiple tasks such as customer service, cooking, serving, and cleaning up. Among these, the cleaning process of dishware
- is difficult to directly link to added value to customers
- results in dirty hands and arms, leading to the risk of chapped hands and other injuries caused by broken dishware.
These problems have existed for many years and there is a strong demand for automation in the cleaning process, which has not hitherto been satisfied.
Specifically, employees manually soak each dirty item of dishware after use in a sink filled with water to remove some of the dirt, and then transfer the dishes to the dishwasher rack for cleaning. Over time, the water that collects in the immersion sink becomes cloudy and dirty, and employees have to put their hands and arms into this dirty water to pick up the dishes. The dirty water contains a large amount of oil, which can cause chapped hands. It can also hide broken dishware, leading to the risk of injury. Consequently, there is a strong need for automation.
In addition, many off-the-shelf fully automatic dishwashers are large machines, which poses challenges when it comes to store space and the different layouts in different stores. These machines also have a high cost for fast food chains, rendering it difficult to make decisions when horizontally deploying them across multiple stores.
Solutions using FingerVision technology
There are three problems that are technically difficult to overcome when implementing a robot that meets the needs for automation of cleaning processes:
There are many types of dishware that need to be handled.
If there is a lot of dirt, such as oil, the dishware will be slippery and hard to remove from the immersion sink when picked up.
It is extremely difficult to recognize dishes submerged in dirty water from outside the water using an external camera.
In response to these technological challenges, FingerVision's vision-based tactile robot hand can be used to handle various types of dishware in a versatile manner, and avoid dropping dishes through slip detection, while also being able to detect different types of dishware even in dirty water. By estimating the dishware’s position, orientation, distance, depth and other characteristics, it becomes possible to significantly increase the range of autonomous control of the robot.
Based on the acquired image information, it is possible to separate information from various modalities, effectively exploiting the full strengths of the vision-based tactile robot hand.
Experimental results of project development
This year, Yoshinoya and FingerVision are proceeding with basic development focusing on the most difficult process in a restaurant, that of removing soiled dishes from dirty water. We chose to focus on the most difficult restaurant task based on the assumption that if we can overcome this level of difficulty, the other components of the automation process will be easier to achieve, making the overall probability of the project’s success far higher.
We successfully conducted experiments at a physical store. Using our technology, the time required for racking by employees at the JR Kanda Station restaurant during peak hours (11 a.m. to 2 p.m.) was reduced from 523 seconds to 289 seconds. Assuming that a standard store operates 24 hours per day, these results represent a 101.7% improvement in human-hour productivity.
Future prospects
By expanding the scope of automation, Yoshinoya and FingerVision aim to further improve productivity and enhance the working environment in restaurants, and in the near future, create a situation where restaurants all over the world no longer need to wash dishes manually. Keeping the keywords "people" and "technology" always in mind, we are accelerating development and improvement with a vision of maximizing the value of "people" by introducing "technology."
Press release：PRTIMES http://prtimes.jp/main/html/rd/p/000000018.000095912.html