【正文】 N was applied to the 300 mm 300 mm 50 mm slab specimens. The wheel passes 42 times per minute at the center of specimen. The wheel tracking tests were conducted at the 60 LC of temperature to evaluate the permanent deformation characteristics of asphalt mixtures. Details on the wheel tracking test results are presented in Table 3. Fig. 4 shows the parison of rut depths with number of load cycles for the conventional mix and the HMAM. The conventional mix has a maximum rut depth of about 8 mm at 20,000 cycles. On the other hand, the maximum rut depth of the HMAM is about mm and there is no increase in the rut depth after 5000 of load applications. It can be concluded from the parison that the HMAM has a great potential to reduce the permanent deformation. Since the aggregate gradation of the both mixes is the same, the increase in the rutting resistance of the HMAM is mainly due to the HMAB used in the mixture. Table 2 Results of moisture susceptibility testMixtureDryWetTSR (%)AirStrengthAirStrengthvoid (%)(kPa)void (%)(kPa)ConventionalHMAMTable 3Results of wheel tracking testConventionalHMAMRut depth (mm)Number of load application (cycles)20,00020,000Load level (N)1372 1372Dynamic stability (N/mm)2747 7168Fig. 4. Cumulative rut depths with number of load applications for the conventional mix and HMAM.. Fatigue testFatigue testing was performed using a servohydraulic closed loop testing equipment manufactured by MTS. All the tests were conducted in the indirect tensile mode, and the loading and measuring systems proposed by LTPP [6] were used in this test. A haversine wave with a loading time with and without rest periods was used in the fatigue tests. Both fatigue tests with and without rest periods were conducted to evaluate the healing potential of the mixes. The number of load repetitions at which the current sti ness decreases to 50% of the initial value is defined as the fatigue life of specimen. The fatigue coe cients a and b shown in Eq. (1) were estimated through a regression analysis of the test data obtained from the both mixes, and presented in Table 4Nf 188。 a240。e0222。b。 240。1222。where Nf is the fatigue life of asphalt mixture, e0 is the initial tensile strain, a and b are fatigue coe cients.Table 4 Fatigue coe cients for tested mixtures with and without rest period Fig. 5 shows the fatigue test results with and without rest periods. Since the same aggregate gradation was used for the both mixes, the fatigue results shown in this figure directly represent the e ects of asphalt binders used in the mixes. It is observed from this figure that the fatigue lives of the both mixes increased with the application of rest period. The increase in fatigue lives is could be due to microdamage healing occurred during the rest period. The more increase in fatigue lives of the HMAM pared to the conventional mix is observed with rest period representing a better healing potential of the HMAM. The SBS polymer mixed in the HMAB could be the main source of this better healing potential.For the fatigue test results with rest period, the resistance to fatigue cracking of the HMAM seems to be better than that of conventional mix where the tensile strain value is lower than 150 microstrain. For typically thick asphalt pavements, the tensile strain values at the bottom of asphalt layer are lower than 150 microstrain. Thus, it can be said that the fatigue resistance of the HMAM is better than that of the conventional mix for the thick asphalt pavements. On the other hand, the fatigue resistance of the HMAM is worse than that of the conventional mix for thin asphalt pavements because of the increased sti ness of the HMAM pared to the conventional mix. A similar result is observed from the fatigue test results without rest period.4. Full scale performance testing. Accelerated pavement tester (APT)The Hanyang University Accelerated Pavement Tester (HAPT) used in this study makes it possible to simulate the large amount of loading in a short period and it is a vital tool for pavement performance evaluation. The HAPT can be driven backwards and forwards over m length of pavement. It can apply the 11 ton maximum wheel load at speed of up to 17 km/h to accelerate the pavement distress. It is also possible to simulate the lateral wandering of real tra c. The heating system installed in the HAPT can control the pavement temperature. The general specification of HAPT is given in Table 5. The test pit can acmodate three linear test tracks. Each test track is m long and m wide. The instrumentation and control system is intended to control the APT operation systematically, collect all the information relating to operation (speed, location, load level), and manage the data such as temperature, deformation, stress, and strain gathered from the di erent types of sensor.Fig. 5. Results of fatigue for the conventional mixture and HMAMTable 5 General Specification of HAPT. Construction and instrumentationThe pavement sections for full scale accelerated pavement testing (APT) are shown in Fig. 6. In order to reduce the testing time and monitor the pavement distresses directly, the surface course has been excluded from test sections. Since the test sections A and B were designed to investigate the fatigue cracking potential of the pavements, the target asphalt layer thickness was set to be 7 cm. However, the cored thickness of the sections A and B were found to be and cm, respectively. Since the test sections were relatively narrow and short, it was di cult to construct the asphalt thicknesses as designed. Test sections C and D were constructed to pare the structural cap