Quick Service Restaurant Order Assistant

Many quick-service restaurants, especially in the fast-food industry are facing a similar challenge: large amounts of orders hit employees with time being a limited and valuable factor. These orders can reach high levels of complexity and contain a vast variety of products. As a result, selecting and packing the right menu items turns into a crucial and error-prone task that requires a constantly high level of focus.

Mistakes made typically result in costly fixes, a lowered customer satisfaction and reputation damages in the long run.

With the Quick Service Restaurant Order Assistant (QSROA) we attempt to solve this problem once and for all.
The QSROA aims to help employees through the order handling process and verifies the status of the current order. Using a camera surveying the packing area, the assistant continuously collates the detected menu items with the current order. QSROA thereby guides the employee not only towards missing menu items but also calls attention to incorrect items and verifies the status of the order.

With the QSROA we attempt to bring added value to the market based on the idea of entering cooperation with business partners. Accordingly, we do not present a fixed product to the market but offer customers a service that is individually tailored to their needs. This approach proves to be beneficial to the customer as that chain-specific challenges can be taken into account.

The QSROA can be broken down into the following main components:

Hardware:       - an edge device with an attached camera

Software:        - real-time menu item detection running on the edge device

- web-based interface collating detected items with the current order

Service:           - software maintenance

- updating network corresponding to seasonal menu-changes

The functional sequence is described below.

Once a new order is placed in the system, the monitor[2] displays its requested menu items to the employee. The job of the employee[4] is now to collect the required menu items and place them in the packing area[3]. The camera[1] captures the items and sends the image to the Jetson Nano, where a custom-trained neural network is working to detect the items in the image. The menu items are then transferred to the new status "checked off". As soon as all the items have been detected, the order can be marked as complete, and the employee may pack the ordered menu and proceed with the following menu request.

Order View