FiWare

CardioID Technologies

Safety on the road has always been of utmost concern to authorities and governments around the globe, where countless lives have been wasted due to carelessness, negligence, or sheer bad luck. CardioWheel is composed of several hardware components: a) a steering wheel with conductive fabrics upholstered, which are the sensor elements that enable the acquisition of the ECG signal from the driver, b) an acquisition unit, entitled CardioMetrics, which is connected with the steering wheel sensor elements, and processes the ECG signal, computes the instantaneous heart-rate, c) a base-station, which receives, via RF, information from the acquisition unit, and sends it aggregated with other complementary information to the cloud.The system includes an analog front end (AFE) with lead-on-detection (LOD), and a new inertial sensor for acquiring all the motion dynamics of the steering wheel, namely linear, and angular accelerations, as well as absolute angles. The system embedded on the wheel connects via RF to a gateway, which transmits the information to our cloud infrastructure, in conjunction with GPS data.

Goals and challenges

Regarding software development, and use of FIWARE technology, this project encompassed three main perspectives:
* Firmware, running on our base station, to relay information to our servers.
* API Backend, to handle authentication, incoming and outgoing data, storage, and processing.
* Dashboard, to make data and insights available to the users.

The choice of FIWARE

The core of the cloud service is the API backend. The system, hosted on FIWARE Cloud using the available virtual machines, makes use of a reverse proxy, which shields the various individual components from direct access from the Internet, and enforces HTTPS authentication and encryption. The API is based on a RESTful paradigm, making use of JSON messages for data
exchange. Serving both the device uplink and the visualization dashboard, the API is backed by an SQL database, storing long term information related to the devices, vehicles and users around which the system is built, and by the Orion Context Broker, which stores short term contextual information. The context broker stores current positional information of vehicles (location, speed,
altitude), as well as driver wellbeing metrics such as heart rate and drowsiness. Additionally, the context broker is configured to notify the Cygnus service, which in turn persists the contextual information into a Hadoop Distributed File System (HDFS). The HDFS is also used to store the ECG recordings obtained from the SD Card. The API itself is implemented in Python using the Falcon Web Framework. This API was designed also taking into account the integration with third party, namely fleet management systems, that will consume the information provided by our system.

The use of FIWARE technology

The dashboard aggregates all data collected in a user-friendly way. For each vehicle it is possible to visualize their routes, the instantaneous heart rate measurements and vehicle speed. This information is very helpful to understand how geolocation can correlate with heart rate and also to understand how the drivers are interacting with the steering wheel throughout the route. The
instantaneous heart rate is one of the parameters that is crucial for the fatigue detection and is being analysedanalyzed in real-time.

Benefits and accomplishments

As result of SOUL-FI project CardioID was able to mature the initial prototype of CardioWheel to a TRL6 – technology demonstrated in relevant environment (industrially relevant environment in the case of key enabling technologies)”, in particular with the pilot performed in conjunction with Barraqueiro. We have established a collaborative agreement with Rodogest, SA through its bus long
route operator Barraqueiro Transportes, the largest private passengers road transportation company in Portugal. This company have demonstrated interesting and provided technical means to evaluate our “drive continuous monitor health status” for theirs drivers.
The pilot was deployed in one long route bus, which is used by several drivers, and will extend to one more bus during August. This pilot enabled CardioID to acquire real data, totalizing 60715 km of monitoring.