196
5 Optimization Algorithm and RFID System Physical Anti-Collision
Fig. 5.32 The spatial structure of five RFID tag network
network can effectively model the nonlinear relationship between the RFID multi-tag
3D coordinate distribution and the corresponding RFID tag reading distance. The
wavelet neural network model can predict the reading distance of RFID multi-tag
networks effectively.
In order to verify the superiority of the wavelet neural network, the GA-BP neural
network and the PSO neural network are compared with the wavelet neural network,
respectively, in this paper. The time cost is calculated 200 times to get the average
time cost of different networks by using the DELL computer (Inter(R) Core(TM)
i5-2450M CPU @ 2.50GHz).
From Fig. 5.33, comparing with PSO neural network and GA-BP neural network,
the prediction relative error of wavelet neural network is smaller. The wavelet neural
network can better model the nonlinear relationship between the 3D coordinates of
RFID multi-tag network and the corresponding RFID tag reading distance. From
Tables 5.8 and 5.9, the average prediction relative error of PSO neural network is
2.77% and the average prediction relative error of GA-BP neural network is 3.44%.
Comparing with PSO neural network and GA-BP neural network, the average prediction relative error of wavelet neural network is only 0.71%. Besides that the time cost
of wavelet neural network is 2.17s. The time cost of wavelet neural network is also
less than other two methods. Compared with corresponding method, the proposed
method has two advantages. On the one hand, the prediction value of the proposed
method is closer to the experimental value, so the relative error is smaller. On the
other hand, the proposed method has better real-time performance than PSO neural
network and GA-BP neural network in the detection system designed in this paper.
5 Optimization Algorithm and RFID System Physical Anti-Collision
Fig. 5.32 The spatial structure of five RFID tag network
network can effectively model the nonlinear relationship between the RFID multi-tag
3D coordinate distribution and the corresponding RFID tag reading distance. The
wavelet neural network model can predict the reading distance of RFID multi-tag
networks effectively.
In order to verify the superiority of the wavelet neural network, the GA-BP neural
network and the PSO neural network are compared with the wavelet neural network,
respectively, in this paper. The time cost is calculated 200 times to get the average
time cost of different networks by using the DELL computer (Inter(R) Core(TM)
i5-2450M CPU @ 2.50GHz).
From Fig. 5.33, comparing with PSO neural network and GA-BP neural network,
the prediction relative error of wavelet neural network is smaller. The wavelet neural
network can better model the nonlinear relationship between the 3D coordinates of
RFID multi-tag network and the corresponding RFID tag reading distance. From
Tables 5.8 and 5.9, the average prediction relative error of PSO neural network is
2.77% and the average prediction relative error of GA-BP neural network is 3.44%.
Comparing with PSO neural network and GA-BP neural network, the average prediction relative error of wavelet neural network is only 0.71%. Besides that the time cost
of wavelet neural network is 2.17s. The time cost of wavelet neural network is also
less than other two methods. Compared with corresponding method, the proposed
method has two advantages. On the one hand, the prediction value of the proposed
method is closer to the experimental value, so the relative error is smaller. On the
other hand, the proposed method has better real-time performance than PSO neural
network and GA-BP neural network in the detection system designed in this paper.