FAST & RELIABLE TRACKING
Real-Time
VACCINE AND
PHARMACEUTICAL
Monitoring System
(Source: Imperial College)
There is a growing need for ensuring the safe storage and distribution of pharmaceutical products that are sensitive to environmental conditions, such as most COVID-19 vaccines which need to be stored and transported at ultra-low temperatures. For these pharmaceutical products, there can be dire consequences if the necessary temperature and atmospheric conditions are unmet.
​
Based on the Raspberry Pi Zero, our smart device utilizes various onboard sensors to track and monitor the environmental conditions of important pharmaceutical consignments. Featuring an Adafruit NXP 9 Degrees of Freedom sensor to analyse translational and rotational changes the system monitors the stability of the container and is able to detect whether the container has been flipped or dropped. Also featured is an Adafruit Si7021 temperature and humidity sensor that will enable clients to ensure environmental conditions during storage are optimal. Lastly, a BMP280 air pressure sensor ensures that any pressurised containers are securely sealed and are not compromised.
The device samples the conditions at a user adjustable-rate to construct periodic reports, which not only contain raw data but also key statistical analysis of the latest readings, to assist in determining whether any anomalies are worthy of raising concern over. Suspect anomalous measurements are processed and communicated immediately upon detection to enable a fast response where it may be needed by the client. The data is communicated over WiFi via an MQTT broker in JSON format and utilised by our back-end for data storage purposes and relayed to our front-end, where clients can view the status of their payload 24 hours a day 7 days a week in real-time.
FAST
A smart IoT embedded device designed to track and monitor fragile and delicate pharmaceutical goods on the move. Designed to analyze and report immediate anomalies in the environment
Prototype:
EASY
It's easy to track your consignment with a clear dashboard displaying all important information including temperature variations and pressure changes.
SECURE
All data is encrypted with SSL and immediate notifications of anomalous sensor readings are sent wirelessly to the dashboard.
Dashboard:
Promotional Video
A short promotional marketing video for our device:
PromotinalVid
The Hardware
Raspberry Pi Zero W
Based around the Raspberry Pi Zero W with a 1GHz CPU and 512MB RAM, our system processes data at a blistering rate and with integrated Wi-Fi, able to transmit data wirelessly for IoT functionality.
​
Onboard the system undertakes statistical analysis on the various sensor readings and uses different algorithms to process acceleration and gyroscopic data.
Si7021 Temperature and Humidity Sensor
The precision temperature and humidity sensor is able to detect changes with an accuracy of ±0.4°C and ±3%RH. With a Dynamic Range of -40 to +125°C and 0-100% respectively our system is able to detect and respond to any stimulus and is able to operate in even the most extreme environments.
BMP280 Pressure Sensor
The pressure sensor is accurate to ±0.12hPa with an integrated temperature sensor for added redundancy and accuracy. The system is also able to error check and verify certain readings.
Accelerometer & Gyroscope
The NXP 9 Degrees of Freedom Breakout Board includes 2 sensors integrated onto one board. With the FXOS8700 3-axis accelerometer/magnetometer and the FXAS21002 3 axis gyroscope the system is able to measure the acceleration of the device and gather angular rotation data of the system. With this data, the system is able to detect the orientation of the device (and thus the consignment of goods), whether it is in motion or even if it has been dropped.
​
Future plans involve implementing a complementary filter to mesh the gyroscopic data and acceleration due to gravity to get a more accurate positional reading of the device.
The Software
The pursuit of a polished product doesn’t end with the hardware. That’s why we have written our own code to harmonise with our sensors to provide a complete service for our clients.
​
One major aspect of our code is the communication with the MQTT broker which acts as a middleman between our device and the back-end of our service. The data that is published is sent via a secure SSL connection with the broker and the device is subscribed to the same topic that it publishes to - allowing two-way communication between the user at the front-end and the device. Any incoming communication from the front-end will be detected by a message handling function.
​
Now that we have a connection to our MQTT broker, we need an ear on the bus looking for readings, and functions to deal with the incoming data. Each sensor has its own associated “get” function which utilises the smbus2 and BUSIO Python modules to communicate with its respective hardware using the i2c protocol. The functions read data that is broadcast on the bus and convert the raw input into usable data which is returned for subsequent functions to take advantage of.
​
The data received from the sensor is checked against predefined thresholds to determine if an anomalous reading has been detected. If the data is an anomaly, an anomaly report is sent immediately, packaging the relevant information into JSON format and publishing it to the corresponding MQTT sub-topic. If the readings all lie within suitable boundaries, the data is appended to relevant arrays and accumulated. Once this is achieved we carry out statistical operations to determine important values such as the min, max, average and standard deviation in the data. All of this information is periodically packaged into a report and published in JSON format to the corresponding MQTT sub-topic.
Dashboard
The back-end service pulls the published data from the relevant sub-topics from the broker and appends new reports to a database. The reports are handled according to their nature and are displayed differently depending on whether they are standard periodic updates or anomaly reports. The front-end web app displays the data from this database in a clean and user-friendly layout. It would be possible to have the client use the web app to input desired values for the thresholds, allowing the device to be used for any required payload without the need for hardcoding the values.
Our proprietary software dashboard makes it easy to monitor and manage the tracking data provided by the system. It's easily accessible 24/7 from the following:
Future POTENTIAL
Future and planned features for the system:
A planned feature is to implement theft detection and enhanced motion tracking using a combination of the existing accelerometer, magnetometer and gyroscopic data as well as data from an added GPS module and meshing the collated data using various filter and algorithms to get accurate, reliable and precise tracking.
A necessary future feature is added cellular connectivity to allow for GPS tracking and precise location monitoring and a path for 5G integration.
A potential to expand into the enterprise and consumer market space targeting both large online retailers and small scale individuals and providing the option to add an extra layer of insurance when tracking goods and packages sold online.
