Navigation in a Supermarket


No more Einkaufsmuffel!

The intelligent shopping cart supports people to find products or navigate from A to B.


Demo


 
 

  •  

Functionality


 
 

  • Active
    Navigation

  • By means of a presorted shopping list, the customer can be navigated from one product to another. For this purpose, a (nearly) optimal path considering all the products on the shopping list is computed.
  • Using a touch-screen, the customer can also select individual products from his or her shopping list. Thereupon, the customer is navigated to the corresponding product on the shortest path.

  • targeted search

  • Passive
    Navigation

  • The customer moves through the supermarket in the usual way. His or her current position in the market is displayed on the screen of the instrumented shopping cart.
  • If he or she gets close to any products that are specified on his or her shopping list, their positions are marked on the map displayed on the screen.

  • unobtrusive reminders

  • Macro-Navigation

    As long as the customer is far away from the aimed product, he or she is navigated via visual hints on the scren of the instrumented shopping cart - either from a bird's eye view (overview map) or from a first-person view (3D model).

  • Micro-Navigation

    As soon as the product gets within the customer's range of vision, visual hints (like e.g., projected arrows) can be displayed in the environment. These hints can be combined with (customer-adaptive) advertising.

Infrastruktur


 
 

  • Current Infrastructure

  • RFID tags: Passive RFID tags are positioned below the floor covering of the supermarket and on the base parts of the product shelves.
  • RFID antennas: An RFID antenna is mounted at the base part of the instrumented shopping cart, which detects the tags on the floor and those at the shelves.
  • Yamamoto model: A 3D model of the supermarket is created using the 3D modeling tool Yamamoto. Both the locations of the different shelves and those of the RFID tags are stored in this model.

  • Other Possible Transponders for Positioning

  • active RFID tags
     
  • Infrared beakons
     
  • Bluetooth
     
  •  
  • The measurements of the different transponders can be weighted individually depending on their accuracy. In this way, they can be combined using dynamic Bayesian Nets in order to determine the current position of the shopping cart. This position detection incorporates not only the currently received signals but also the history of the previous measurements.