Fiberglass prepared from broken waste glass can be used in epoxy asphalt mixtures for performance enhancement and a toughening effect. There is no systematic study on the influence mechanism of the size and the amount of glass fiber on the properties of epoxy asphalt mixtures. The effects of fiberglass on the properties of epoxy asphalt concrete were evaluated using a tensile test, three-point bending test, four-point bending fatigue test and an SEM scanning test. The results verify that the tensile strength of epoxy asphalt mastic with a 6 mm length and 2% content increased the most. Compared with the nondoped glass fiber, it increased by 69.2%. Under the influence of the internal composition of the asphalt mixture, the optimal ratio scheme is different from that of epoxy asphalt mastic. A microscopic analysis showed that uniformly dispersed fiberglass in the epoxy asphalt mixture forms a spatial network structure, leading to reinforcement and the restraint of microcrack expansion. The addition of fiberglass with a length of 9 mm and at a concentration of 5% to the epoxy asphalt mixture resulted in the maximum road performance. The Marshall stability increased by 43.5%, and the flexural and tensile strength increased by 33.7%. The fiberglass length is the most important factor limiting the strength and toughening effects of epoxy asphalt mixtures.

Aiming to evaluate the applicability of Fiber Bragg Grating (FBG) strain sensors for semirigid pavement monitoring, beam and cylinder specimens with three types of FBG strain sensors embedded in two kinds of classic semirigid pavement materials, asphalt mixture (AC‐25) and cement‐stabilized crushed stones (CSCS), were prepared in the laboratory. Four‐point bending tests and uniaxial‐compression tests were carried out under different loading frequencies and temperatures to evaluate the working properties of these sensors and then obtain the corresponding real sensitivity coefficients (SCs). The experimental results showed that the synchronism, repeatability, and linearity of all these sensors were prominent. However, the real SC results were significantly different from the recommended and dependent on many factors including temperature, the loading frequencies, the stress state, and the type of embedded material to different degrees. The SCs remained stable when the moduli of the embedded materials were high enough; otherwise, the SCs varied. Two SC prediction models that used the modulus of the embedded material as the only independent variable were developed to deal with the problem of instability. The modulus difference level between the sensors and the embedded material could integrate the factors roughly, except for the stress state. It is recommended that the factors above should be considered when using FBG strain sensors in practice, and it is still necessary to perform laboratory calibration in advance.