Abstract
Wireless sensor communications are distributed, self-disciplined, and on-demand, integrating the benefits of different infrastructure and communication technologies. Recent development in wireless user application requires on-demand localization and target tracking competence to improve communication reliability. This manuscript proposes a learning-based localization (LL) method for minimizing localization errors due to variations in received signal strength assessment. The learning process analyzes the signal in terms of the communication modes adopted by the sensors to identify localization error. Errors in transmission, channel noise, and angle of deviation are addressed using the Gaussian transform (GT) function for the received signal. The errors are recurrently identified with the displacement of the sensor to improve location precision. Distinct from the conventional learning process, this learning is assisted with constraints obtained from a hidden layer to improve the rate of region coverage. The performance of the proposed localization method was evaluated through simulations and was found to improve localization coverage with decreased detection time, energy, and localization error. The experimental results proved the consistency of the proposed LL-GT method by minimizing localization error by 39.47%, energy requirement by 13.73%, and detection time by 26.08% and by improving the coverage by 15.15%.