The most traditional method for the determination of total thickness of a flexible pavement is based on the elastic deformations computed using the Burmister’s layer theory (1958) or by the California Bearing Ratio (CBR) method. But the current criteria relies on the quality control of the pavement subgrade and is based on the in-place dry density and moisture content of the subgrade during construction according to MORTH standards (2013). It is assumed that the pavement subgrade will perform to the anticipated mark throughout the design life if the target dry density or relative compaction of 98% is achieved in the field (MORTH standards, 2013). However, the subgrade material’s behaviour is more influenced by the modulus rather than the dry density and moisture content (Livneh and Goldberg, 2001). The empirical design procedure for the flexible pavement relies on the static-strength parameters such as California Bearing Ratio (CBR), triaxial test results and Soil Support Value (SSV), which does not represent the dynamic loading offered by the traffic. Resilient modulus which could be obtained from the repeated load testing is the better indication of the dynamic behaviour of a geomaterial (AASHTO, 1993). The Mechanical-Empirical pavement design procedure considers resilient modulus as the primary design factor for the pavement materials (Ali and Tayabji, 1999).

The resilient modulus is defined as the ratio of deviator stress to recoverable elastic strain under a transient dynamic pulse load. Although resilient modulus can be determined from laboratory repeated load triaxial test or can be back calculated from non destructive tests like Falling Weight Deflectometer (FWD), road rater and dynaflect etc., they have their own advantages and disadvantages. Resilient modulus values obtained from FWD is the most reliable among the available tests, but it is expensive and cumbersome (Chen et al., 2001). Laboratory repeated load triaxial test needs trained personnel and cumbersome as well (Rahim and George, 2002). Among the available technologies, Seismic Pavement Analyser (SPA) is the quickest and easiest method to determine the characteristics of the constructed layers. But the major disadvantage is that the modulus values estimated by the DSPA seem to vary over wide range (Meshkani et al., 2003). With the advent of the new technologies like geogauge, DCP and LWFD, estimation of the stiffness of the engineered subgrade has become much easier. These devices have the capability to estimate the level of compaction and degree of uniformity of the geomaterial being tested (Burnham, 1996; Siekmeir et al., 1999). But the stiffness represented by these devices cannot be representative of traffic, as the loading conditions and the stress-strain responses are different. Also, elastic modulus obtained from the conventional in-situ tests doesn’t account for stress conditions of the material being tested, whereas the resilient modulus of the material is actually stress-dependent (Tanyu et al., 2003). For these reasons, it is necessary to develop an equation to estimate the resilient modulus from the elastic modulus measured from various other available in situ devices. Earlier, a correlation chart was developed in order to determine the MR value from the CBR, Texas triaxial classification and internal friction R-value (Andrei, 1999). Relationships between MR, CBR, R-value and SSV were developed by several researchers (Til et al., 1972).

20 different relationships were developed by Livneh (1989) to correlate the results obtained from DCP to estimate the CBR. Moreover, the reliability of the DCP to estimate the CBR has been proved by several researchers (Smith and Pratt, 1983; Harison, 1987; Livneh and Ishai, 1988; Coonse, 1999; and George et al., 2009). Relationships between DCPI and CBR were formulated for a wide range of granular and cohesive materials by US Army Corps of Engineers (Webster et al., 1992). An equation was developed by Gabr et al. (2000) by conducting experiments on subgrades overlaid with aggregate base course. Relationships between the penetration rate (DCPI) and the subgrade elastic modulus were formulated by Chua and Lytton (1981) and Powell et al. (1984). A relationship between the back calculated resilient modulus from the falling weight deflectometer and the DCPI was developed by Chen et al. (1999). A correlation between DCPI and the back calculated elastic modulus from plate load tests was developed by Konrad and Lachance (2000). Following them, correlations connecting elastic modulus and DCPI were given by Chen et al. (2005).

Light weight falling deflectometer is a portable device developed as an alternative to the in situ plate load test. A non linear regression analysis and developed a correlation relating the back calculated resilient modulus from the FWD and the subgrade modulus determined from the LWFD Nazzal et al. (2007). An extensive study was conducted to correlate the moduli obtained from the 3 different types of LWFD (TFT, GDP and Prima 100) available in the market and the resilient modulus obtained from the FWD. It is observed that the modulus found out by the Prima 100 was more consistent with the FWD resilient modulus (MFWD = 1.031 ELWFD). However the correlations suggested for the moduli obtained from other LWFDs are MFWD = 1.05 to 2.22 EGDP and MFWD = 0.76 to 1.32 ETFT (Fleming et al., 2000). A linear relationship between the elastic modulus obtained from LWFD by Prima 100 and plate load test was given by Kamiura et al. (2000). A simple correlation between the back-calculated resilient modulus and the elastic modulus obtained from the soil stiffness gauge was developed by Sawangsuriya et al. (2005). A detailed study was carried out to establish a comparison of the soil stiffness gauge and the available quality control techniques by Wu et al. (1998). A linear correlation has been developed between the stiffness measured using the geogauge and the modulus measured by FWD, PSPA and Spectral Analysis of Surface Waves (SASW) and it is proved that the soil stiffness gauge has a great potential for use as an quality control device in the pavement construction (Sawangsuriya et al., 1849).

For the present study, data was collected from various regions consisting of different soil types by conducting SSG, LWFD and DCP tests. The data collected allowed the development of multiple regression equations to determine resilient modulus of different types of soil. The first objective of this study is to develop a correlation between the resilient modulus, plasticity index, moisture content and the dry density. The second objective is to examine the closeness between the experimentally calculated resilient modulus and the computed resilient modulus from the derived correlation. The third objective is to obtain a relationship between the resilient modulus and the elastic modulus obtained from different in-situ technologies.