Target tracking is the application of WSN in which sensor nodes continuously monitors and reports the positions of moving objects to the base station with minimum latency. However, tracking a target using visual sensors is more difficult due to the random movements of visual objects. As WSN continuously
monitors the environment, energy efficiency of the network gets degraded. Therefore, a novel deep learning technique called Minkowski Jarvis-Patrick-Clustered Multilateration Shift Invariant Deep Convolutional Neural learning (MJPCMSIDCNL) is introduced for improving the target tracking accuracy with lesserenergy consumption. The proposed MJPCMSIDCNL technique comprises many layers such as input, three hidden and output layers. Initially, the IoT devices are used as sensor nodes in the input layer to sense and collect the data. Then the residual energy and received signal strength of each sensor node are measured in the first hidden layer to improve the energy-efficient target tracking in WSN. In the second hidden layer, the Minkowski Jarvis-Patrick-Clustering technique is applied to partition the network into different groups based on the energy and received signal strength of the sensor node. Finally, target object tracking is performed at the third hidden layer using the True-range multilateration method. As a result, the exact location is identified at the output layer with higher accuracy and minimum time consumption. The simulation is conducted with various performance metrics such as energy consumption, target tracking accuracy and target tracking time, with respect to a number of sensor nodes. The results and discussion demonstrate that the proposed MJPCMSIDCNL technique increases the target tracking accuracy and minimizes the energy consumption as well as target tracking time than the existing techniques.