Machine learning driven aggregation aware bitmap MAC protocol for energy efficient data transmission in WSNs.

An Aggregation-Aware Energy-Efficient Bit-Mapping Medium Access Control Protocol (AABMP) is proposed for efficient data transmission in Wireless Sensor Networks and the Internet of Things. The main objective of the proposed method is to effectively manage the data transmission from sensor nodes to cluster heads in an energy-efficient manner. AABMP aggregates the data by estimating the mean value of a sliding window of the previous few samples and calculates the deviation of the current reading with respect to mean value. The deviation aware method is proposed for the transmission decision from sensor node to cluster head node. The deviation is predicted for real-time dataset using various machine learning methods. Based on performance evaluations, the most effective ML method is integrated with the bit-mapping MAC protocol in an energy-efficient manner. The aggregation-aware ML approach reduces the number of transmitted packets by efficiently identifying redundant data. The proposed approach is evaluated using the Intel LAB dataset ( https://db.csail.mit.edu/labdata/labdata.html ) of 55 real sensor nodes. The proposed method demonstrates practical applicability for energy-efficient data transmission in IoT-oriented WSN deployments. For the efficient analysis, the precise number of transmitted packet is estimated based on transmission probability. The algorithm is proposed for the estimation of transmission probability. For performance analysis, we have compared multiple ML methods to determine the most suitable one. This optimal method is then integrated with the bit-mapping-based energy-efficient piggybacking scheme. AABMP is also compared with existing MAC protocols. The results demonstrate its superiority in terms of energy savings across both worst-case and best-case scenarios.