Smartcart project

Latest update on 20 March 2016


Shopping assistance
Consumer, connected cart, stock shelves, and edge gateway

Smartcart is the project name for a smart and connected shopping cart system including hardware, software, documentation, and support for retail business.

Elevator pitch

Smartcart adds value to retail areas, serving customers and employees

Use cases

Blueprint backgroundConsumer browsing store shelves and producing metadata for retail analysis

Shopper approaches shelves

Shopper places stock in cart

  • Name: Shopper movement
  • Description: A shopper moves about a retail floor viewing stocked products
  • Actors: Consumer of stocked products, connected smartcart container
  • Preconditions: The consumer takes a smartcart when entering the store
  • Basic flow: Both cart and shelf trade IDs, record the time spent in proximity, and when the consumer finishes shopping the smartcart stores and forwards behavior data
  • Alternate flows: A cart determines its human relation and invites the consumer's mobile devices to connect to the smartcart web application
  • Exception flows: Improper placement of beacon and receiving devices causes interference requiring a measure of self correction
  • Post conditions: The consumer replaces the smartcart in a well defined location and the smartcart onboard movement circuit senses a terminal state
  • Name: Shopper selection
  • Description: A shopper searches for, finds, and takes a product from the shelf
  • Actors: Consumer of stocked products, smartcart container, product shelf
  • Preconditions: Inanimate actors are RFID tagged or connected via corresponding technology
  • Basic flow: A cart reads the RFID tag of a selected product and may issue a telecommand (offer a rebate or advise on missing recipe ingredients) to store shelves when the product is removed or the consumer approaches a related shelf
  • Alternate flows: If the last item is selected, the shelf may announce that more stock is available
  • Exception flows: Measured weight or similar sensory criteria may indicate a flawed RFID reading
  • Post conditions: A cart reads and temporarily stores identity of selected product
Retailer reviewing the consumer dashboardBlueprint background

Retailer reviews time spent

Retailer merges databases

  • Name: Retailer reporting
  • Description: The retailer learns of shopping patterns through point of sale requests and real time notifications and may take action
  • Actors: Branch manager, edge gateway
  • Preconditions: The manager has a online device
  • Basic flow: Management and their authorized staff receive information about shopping behavior, and may offer help where most needed or reshuffle shelves
  • Alternate flows: Management request product stock estimates and cognitive analysis from IBM Watson or similar cloud services
  • Exception flows: Management receives an anomaly alert in contrast with analysed patterns, and may investigate
  • Post conditions: The manager informs the gateway administrator of unwanted notifications
  • Name: Database merge
  • Description: The retailer merges cart location, duration, selection, stock on hand, and vendor databases for reorder and future promotions
  • Actors: Branch manager, gateway administrator
  • Preconditions: Data is available in compatible formats (JSON, BSON, CSV, SQL)
  • Basic flow: Humans (via H2M communication), carts, and shelves send data to the edge gateway which stores it in a database, in turn providing the branch manager with the ability to merge databases for the purpuse of data mining of shopping patterns
  • Alternate flows: A database merge leads to information calling for a vendor review (like price versus demand comparison)
  • Exception flows: A database merge fails, requiring intervention by the edge gateway administrator
  • Post conditions: A new database is produced and a log entry is made on the database server


Smartcart is unique, as shown by searching for 'smart connected cart' and finding that no existing projects implement or extend a Physical Web as Google has defined it. It also has a complementary human user interface for both retailer and consumer.


The project integrates legacy (Bluetooth) and modern (ZigBee, Z-Wave) communication transports to provide for a seamless shopping experience. It employs proven low power IoT messaging protocols (like MQTT) and showcases environmental efficiency by harvesting energy from wheel based carts with dynamo generators.

Modern computer controlled shopping cart


By promoting a pilot system, Smartcart will attract attention from market analysts and groups buying infrastructure with added value (like charger equipped restaurant tables.)

  • Target
  • Costco
  • Wallgreens
  • Home Depot
  • Tesco
  • Auchan
  • Carrefour
  • Sainsbury's
  • Aldi
  • Metro
  • Migros
  • Bauhaus

Smartcart is retail agnostic and can complement clothing, hardware, book, food, garden, textile, pet, or other stores where shopping carts are used.

Privacy wise

Important! Electronic substitutions for plastic customer loyalty cards suffer from privacy considerations affecting both the shopper (who is annoyed by perceived data theft) and retailer (who is accountable to vendor contracts and identity laws.) Smartcart tracks the cart rather than the consumer and implicitly preserves customer loyalty while delivering on data collection of shopping patterns.


Smartcart is a potential student challenge at educational events and hackathons. Rapid prototypes and live demonstrations inspire and motivate conference delegates.

Biweekly progress reports and construction advice will be published in two or more online locations. A full length article will be published by a renowned publisher. The project plan is designed to benefit and add value to the community of IoT developers and their understanding of how connected devices can work toegether.

Project plan

Project Timeline in simulated Gantt format Project component dependencies hand drawn



The project plan includes a timeline estimate and release roadmap.

Design doc

Curie embedded cart transmitting Eddystone URI beacon framesEdge router performing prioritized message collection and processing

Cart and shelf communication

Edge parsing telemetry data

The preliminary design includes a set of UML diagrams to be published in a general detailed design document located in the online archives.

Work flows

Implementing a Edison beacon

Implementing a nRF52 beacon

Many software components rely on Node.js which is hand coded with custom package.json files at times. The MRAA and UPM libraries provide high level support for sensor platforms, and the Intel XDK is used for developing shelf bound IoT devices or when constructing user interfaces for H2M communication with the system. Apache Cordova provides a abstraction layer for those human user interfaces. Lastly, the Intel System Studio for Microcontrollers integrates Eclipse to provide an IDE for Intel Quark D2000 and SE equipped devices.


A hardware bill of materials is provided, and software can be freely downloaded from the project fileserver. Documentation for testers, system administrators, service technicians, managers, and users is available in the online archives. Staff carries out production deployment on a case by case basis to allow for custom configurations.

Pilot launch

A first stage pilot launch serves as a proof of concept. In the last week of development, a limited scale prototype arrangement (no energy harvesting dynamo and large cart substituted with hand cart) is tested in a gateway connected six beacon building matrix. A shopping event is simulated.

Small hand carried red shopping cart

Classic red handled shopping cart

A second stage pilot launch incorporates multiple large carts and commercial blind tests. It concludes the code and project completion date and depends on cooperation from local retailers.


Management device

  1. 1 Dell Edge 5000 (#1001)
  2. Beacon appliances (#1002)
  3. …or Edison devices (#1003)
  4. …or nRF52 devices (#1004)

Cart device

  1. 1 Intel Curie (#1101)
  2. 1 Hand carried cart (#1102)
  3. 1 Rolling metal cart (#1103)

Energy source

  1. Wheelspin energy harvester (#1201)
    1. 1 5V voltage regulator (#1202)
    2. 1 6V dynamo generator (#1203)
    3. 1 2A Li-ion/polymer battery (#1204)

Shelf device

  1. 1 Intel Edison (#1301)

H2M device

  1. 1 Dell Venue 8 Pro 5000 tablet (#1401)
  2. 1 QR code reading smartphone (#1402)

The bill of materials accomodates one test scenario which is scalable to retail spaces of any practical size. Smartcart is an inexpensive shopping supplement to existing infrastructure.


The project was founded by Michael Schloh von Bennewitz and staffed by Michael and Flaki.


The project is self funded and hosted by Europalab Networks. We are interested in working with other groups, and will enter the project in calls to bid. A business plan is available to investors considering funding the project.


Smartcart is licensed under the GNU Public License version 3, and contains Embedded CACert Root Certificates. For more information please read the CACert root distribution license. Components of Smartcart contain software from other projects which are accordingly sublicensed.