Research Article
Energy Efficiency of Ultra-Low-Power Bicycle Wireless Sensor Networks Based on a Combination of Power Reduction Techniques
Table 6
Designed parameters of the ANN based on the PSO algorithm.
| Parameters | Value | Type |
| Number of inputs | 3 | RSSI | Number of hidden layers | 2 | Hyperbolic tangent sigmoid activation function | Number of outputs | 1 | Linear activation function/distance | Number of neurons in each hidden layer (1 and 2) | 18 | For 1 (obtained from the PSO algorithm) | 18 | For 2 (obtained from the PSO algorithm) | Learning rate () | 0.5941 | Obtained from the PSO algorithm | Number of iterations | 1000 | | Target error (goal) | 0.001 m | | Number of collected data samples | 1125 | RSSI samples for outdoor environment | 1125 | RSSI samples for indoor environment |
|
|