Mobile Netw Appl (2014)19:524-533 531 18, 16 十一ARCP PRCP 14 ARCP ·PRCP BRCP CRCP 20 ECRCP RCH 10 -ECRCP-No Estimation 8 6 0.050.10.15020.250.30.350.4 500100015002000250030003500400045005000 Ratio Number of tags in scanning area Fig.5 Execution time under different value of ratios Fig.7 Execution time under different number of tags in scanning area 2.3).the kinds of category IDs in the slot are small and 8.4 Execution time comparison collision slots cannot be fully utilized.If a is too large,the collision slots are difficult to be decoded.which wastes a lot Figures 6.7 and 8 shows the execution time,CRCP and of time.Therefore,the optimal values of a and d are =4 ECRCP have better performances,and ECRCP achieves the and d=7,which are not relevant to the length of category highest time efficiency.In Fig.6,when the length of Ci is ID,the number of rules or tags,and are used in the following 64 bits,the execution time of ECRCP is about 0.5 seconds, experiments. which is 3 of the time required by ARCP.The perfor- mance of ECRCP is unrelated to the length of category ID. 8.3 Threshold value o* In Fig.7.when the number of tags is 5000.the execution time of ECRCP is about 0.5 seconds,which is 1.8 of the Figure 5 shows the execution time of each protocol while time required by ARCP.The number of tags has no effect varying the ratio of p.ECRCP-No estimation means that on ECRCP,because ECRCP focuses on the categories in the ECRCP does not estimate the tag size and works in the way rules instead of those in the scanning area.In Fig.8,when described in Section 7.2.2.In Fig.5,when p<0.25,ARCP the number of rules is 500,the execution time of ECRCP is achieves the best performance.This is because the tag size about 0.8 seconds,which is 3.1 of the time required by in the scanning area is much smaller than the number of BRCP.This is because ECRCP not only resolves the col- related categories.In regard to p,it is not relevant to the lision slot but also leverages the rule's logical feature.In length of category ID,the number of rules or tags.It only Fig.9.when the number of rules is 50.ECRCP only verifies concentrates on the ratio of tag size to the number of related 21 of the related category IDs. categories.Therefore,we set the threshold of the ratio to Figures 10 and 11 provide some fine-grained analysis p*=0.25.In our proposed ECRCP,if p<p*,ECRCP about the efficiency of the protocols.Figure 10 illustrates works as ARCP to check the rules.Otherwise,it works in the utilization ratio of the responses slots.CRCP and the way described in Section 7.2.2. ECRCP have higher utilization ratio of responses slots than the baseline ones,because CRCP and ECRCP resolve the 20 30 ·ARCP 16 PRCP 一ARCP 25 14 BRCP -PRCP ·CRCP 12 BRCP 20 一ECRCP CRCP ⊙一ECRCP aw!L 15 10 ǚ 16 2432404856 64 50100150200250300350400450500 Length of Categoy ID(bits) Number of rules Fig.6 Execution time under different length of C Fig.8 Execution time under different number of rules SpringerFig. 5 Execution time under different value of ratios 2, 3), the kinds of category IDs in the slot are small and collision slots cannot be fully utilized. If α¯ is too large, the collision slots are difficult to be decoded, which wastes a lot of time. Therefore, the optimal values of α¯ and d are α¯ = 4 and d = 7, which are not relevant to the length of category ID, the number of rules or tags, and are used in the following experiments. 8.3 Threshold value ρ∗ Figure 5 shows the execution time of each protocol while varying the ratio of ρ. ECRCP-No estimation means that ECRCP does not estimate the tag size and works in the way described in Section 7.2.2. In Fig. 5, when ρ ≤ 0.25, ARCP achieves the best performance. This is because the tag size in the scanning area is much smaller than the number of related categories. In regard to ρ, it is not relevant to the length of category ID, the number of rules or tags. It only concentrates on the ratio of tag size to the number of related categories. Therefore, we set the threshold of the ratio to ρ∗ = 0.25. In our proposed ECRCP, if ρ ≤ ρ∗, ECRCP works as ARCP to check the rules. Otherwise, it works in the way described in Section 7.2.2. Fig. 6 Execution time under different length of Cj Fig. 7 Execution time under different number of tags in scanning area 8.4 Execution time comparison Figures 6, 7 and 8 shows the execution time, CRCP and ECRCP have better performances, and ECRCP achieves the highest time efficiency. In Fig. 6, when the length of Cj is 64 bits, the execution time of ECRCP is about 0.5 seconds, which is 3 % of the time required by ARCP. The perfor￾mance of ECRCP is unrelated to the length of category ID. In Fig. 7, when the number of tags is 5000, the execution time of ECRCP is about 0.5 seconds, which is 1.8 % of the time required by ARCP. The number of tags has no effect on ECRCP, because ECRCP focuses on the categories in the rules instead of those in the scanning area. In Fig. 8, when the number of rules is 500, the execution time of ECRCP is about 0.8 seconds, which is 3.1 % of the time required by BRCP. This is because ECRCP not only resolves the col￾lision slot but also leverages the rule’s logical feature. In Fig. 9, when the number of rules is 50, ECRCP only verifies 21 % of the related category IDs. Figures 10 and 11 provide some fine-grained analysis about the efficiency of the protocols. Figure 10 illustrates the utilization ratio of the responses slots. CRCP and ECRCP have higher utilization ratio of responses slots than the baseline ones, because CRCP and ECRCP resolve the Fig. 8 Execution time under different number of rules Mobile Netw Appl (2014) 19:524–533 531