Improving IoT Sensor Data Quality using Kalman Filter: The case of Cold Chain monitoring for Indian Mangoes

The sensor data comprising Temperature, Carbon-dioxide, and other miscellaneous gases are the parameters collected during cold chain operation and indicate the fruit quality. However, due to low-cost sensors, the data is often noisy and requires cleaning and estimation at the edge node before being sent over to the cloud. This paper proposed a method that improves the data correspondingly, enhancing the sensor’s performance. A point-to-point wireless network setup of IoT devices is used for collecting two datasets during an experiment on improving the shelf life of Indian mangoes. Each dataset comprises sensor responses for the quality parameters under plasma treatment and controlled environment. In each case, devices are placed in three locations: chamber space, middle-box, and lowest-box. Using this data, we showed how Kalman Filter could be used to improve the data quality of the sensors when compared with the ground truth measurements.