Wireless Channel Estimation for Low-Power IoT Devices Using Real-Time Data

Abstract

The Internet of Things (IoT) is gaining immense popularity in executing automation activities via wireless connectivity in the modern era. The IoT networks are designed using mostly low-power IoT (LP-IoT) devices that are battery-operated and have limited computational power. The wireless communication amongst these LP-IoT devices is affected due to the undesirable factors affecting the wireless channel, such as physical obstructions, the distance between devices, wireless network interference, and power limitations of IoT devices. These factors result in attenuation, distortion and phase-shift of the signals arriving at the receiver device. To encounter the effects of the factors affecting the wireless channel in LP-IoT communication, we estimate the wireless channel at the transmitter device before transmission. An effective channel estimation guarantees reliable transmission, improves the throughput rate, and extends the life of the entire IoT network. This study presents two models relevant to LP-IoT communication in IoT networks. The first model is the LP-IoT communication model, which provides a theoretical representation of the wireless channel for the LP-IoT network. The second model is the channel estimation model, where we apply the Least Squares (LSE) and Maximum Likelihood (MLE) techniques to estimate the LP-IoT wireless channel. We analyse the squared error obtained through the LSE and minimise it to reach a Target Error Threshold (TET), where the estimation results are considered accurate. We developed a novel outlier removal method (OUT-R) to eliminate outliers in LP-IoT wireless channel data to achieve this. After outlier removal, we implement the Kalman Filter (KF) method to further improve the channel estimation accuracy. The observation data needed for this investigation has been obtained from real-time measurements in a controlled Line of Sight (LoS) indoor setting using LP-IoT devices. The findings of this study indicate that the suggested method may meet the specified error threshold TET, yielding accurate channel estimation for LP-IoT communication in IoT networks.

Keywords

• Wireless channel estimation
• low-power IoT devices
• RSSI estimation
• waspmote
• least squares
• Kalman filter