INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 20
www.rsisinternational.org
Adaptation of Crumb Rubber Modified Asphalt Predictive Models
for Nigerian Climatic Conditions: A Transfer Learning Approach
Egbebike M. O1*. , Ezeagu C. A.2, Iyeke S.D.3
Department of Civil Engineering, Nnamdi Azikiwe University, Awka, Nigeria; and Center for
Environmental Management and Green Energy, University of Nigeria, Nsukka, Enugu Campus,
Nigeria.
Department of Civil Engineering, Nnamdi Azikiwe University, Awka, Nigeria.
Department of Civil Engineering, Nnamdi Azikiwe University, Awka, Nigeria.
*Corresponding Author
DOI: https://doi.org/10.51244/IJRSI.2025.121000003
Received: 23 Sep 2025; Accepted: 29 Sep 2025; Published: 27 October 2025
ABSTRACT
Crumb Rubber Modified Asphalt (CRMA) represents a major advancement in sustainable road construction,
widely adopted in the United States to improve pavement durability, reduce rutting, and utilize waste tires.
However, its application in developing countries like Nigeria remains limited, largely due to the lack of region-
specific performance models, climatic differences, and infrastructural challenges. This study proposes a
transfer learning approach to adapt predictive CRMA models from the United States to Nigerian climatic
zones using climate matching, multivariate regression, artificial neural networks (ANN), and multi-objective
optimization techniques. Using simulated data representative of U.S. state climates and traffic conditions, we
modeled performance indices such as Marshall Stability, rutting resistance, and fatigue retention. The results
identify optimal crumb rubber contents (CR%) of 10–15% for different climate-traffic scenarios. Enhanced
models including traffic loads (ESALs) were developed and mapped to Nigerian conditions. This supports
sustainable CRMA deployment for road infrastructure in Nigeria and similar regions.
Keywords: crumb rubber, predictive modeling, asphalt performance, optimization, Nigeria, ANN, regression
Keywords: Crumb rubber modified asphalt; sustainable asphalt; predictive modelling; Marshall stability;
rutting resistance; fatigue retention.
INTRODUCTION
The growing demand for sustainable transportation infrastructure has intensified interest in environmentally
friendly pavement technologies. Among these, Crumb Rubber Modified Asphalt (CRMA) has emerged as a
promising innovation that addresses both engineering and ecological challenges. Developed through the
incorporation of ground tire rubber into conventional bitumen, CRMA improves pavement performance-
enhancing rut resistance, fatigue life, and thermal cracking tolerance-while simultaneously promoting the
recycling of waste vehicle tires. Developed nations like the United States have widely adopted CRMA,
supported by significant investment in research, policy, and pilot projects (Lo Presti, 2013; Putman &
Amirkhanian, 2004). However, in sub-Saharan Africa and other developing regions, the implementation of
CRMA remains sparse due to technical, economic, and climatic barriers.
Nigeria, Africa’s most populous country and one with a rapidly growing vehicular population, generates tens
of thousands of used tires annually (Okonkwo et al., 2022). These tires often end up in landfills or are
incinerated under environmentally hazardous conditions. At the same time, the nation grapples with
deteriorating road networks, especially in regions with high axle loading and intense rainfall. These twin
challenges-road degradation and tire waste-underscore the opportunity to apply CRMA as a dual-benefit
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 21
www.rsisinternational.org
solution. Yet, because Nigeria lacks performance-based design guides tailored to local conditions, transferring
and calibrating models developed elsewhere becomes essential.
This paper proposes a framework to adapt U.S.-based CRMA predictive models to the Nigerian context. Using
transfer learning principles, climate matching, multivariate regression, and artificial neural networks (ANN),
we simulate and calibrate optimal crumb rubber content for flexible pavements in Nigeria’s diverse climatic
zones. The goal is to establish locally relevant mix designs that maximize mechanical performance and
sustainability using globally tested knowledge.
LITERATURE REVIEW
Crumb Rubber in Asphalt Technology
Crumb rubber, obtained from grinding end-of-life tires, has been studied extensively as a bitumen modifier. It
enhances the elastic recovery, stiffness, and aging resistance of asphalt binders (Lo Presti, 2013). Two main
processes-wet and dry-are used in incorporating crumb rubber into asphalt. In the wet process, rubber is
digested into the binder at high temperatures, resulting in a homogeneously modified binder. The dry process
involves blending rubber with the aggregate before adding the binder, often resulting in a coarser mix but
simpler production logistics (Mashaan et al., 2014).
Performance enhancements from CRMA include:
Improved rutting resistance (Kaloush et al., 2002)
Extended fatigue life under repeated loading (Ghabchi et al., 2013)
Superior low-temperature cracking resistance (Fazaeli et al., 2016)
International Experience with CRMA
In the United States, various state Departments of Transportation (DOTs) have implemented CRMA
extensively since the 1990s. For example, Arizona and California have developed specification guides for
using 15-20% rubber content by weight of binder (Way et al., 2011). Florida and Texas, facing humid and
semi-arid conditions respectively, have reported performance gains using 10-18% rubber, especially for
fatigue-prone roads (Liang et al., 2022). These regional adaptations underscore the importance of climatic
tailoring in CRMA design.
Several performance prediction models have also been developed. Zhou et al. (2020) used machine learning
models to relate crumb rubber dosage, binder content, and temperature to fatigue performance. Meanwhile,
Fazaeli et al. (2016) applied Response Surface Methodology (RSM) and ANOVA to optimize performance
indices like Marshall Stability and Indirect Tensile Strength (ITS).
Long-Term Performance with Traffic
Field data (Caltrans, 2015; FHWA) confirms CRMA’s superior rutting resistance under heavy traffic.
Performance models often include both CR% and cumulative traffic (ESALs):
R(t) = atb e-cCRLγ , Fr(t) = 100e-dt (1 + f CR)L−η
Where R(t) is rut depth (mm), Fr(t) is fatigue retention (%), t is age (years), CR is crumb rubber %, L is
cumulative ESALs (millions), and a, b, c, γ, d, f, η are model calibration constants.
Nigerian Context and the Need for Calibration
In Nigeria, road infrastructure suffers from underfunding, poor maintenance, and harsh tropical environmental
conditions. Southern zones are subjected to high rainfall (>2000 mm/year), while northern zones endure
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 22
www.rsisinternational.org
extreme heat and UV radiation. Subgrades are often lateritic and expansive, amplifying the risks of rutting and
fatigue. Although crumb rubber has been studied experimentally in some Nigerian universities, the country
lacks a comprehensive, climate-sensitive CRMA framework.
Okonkwo et al. (2022) emphasized the need for design models tailored to Nigeria’s climatic and traffic
conditions. Existing models from temperate zones do not generalize well due to differences in pavement
temperature profiles, aggregate sources, and construction practices.
METHODOLOGY
The methodology adopted in this study is structured to enable the transfer, adaptation, and validation of
CRMA performance models from U.S. climatic contexts to Nigeria’s environmental conditions. It consists of
six core phases: data synthesis, climate mapping, model building, performance simulation, optimization, and
regional calibration. Figure 1 (to be inserted) illustrates the overall research workflow.
Data Collection
Due to limited availability of consistent CRMA field data from Nigeria, the research
utilized a comprehensive datasets from U.S. States combining laboratory test outcomes and long-term
pavement performance under various climatic and traffic conditions. The dataset represented:
Crumb rubber content (CR%): 0%, 10%, 15%, 20%
Climatic zones: Hot-Dry (e.g., Arizona), Hot-Humid (e.g., Florida), Semi-Arid (e.g., Texas), Mediterranean
(e.g., California), Cold (e.g., Minnesota), and mapped Nigerian analogs
Traffic loads: Expressed as cumulative ESALs (10–50 million over 15 years)
Performance indices:
Laboratory: Marshall Stability (kN), Flow (mm), Indirect Tensile Strength (MPa), initial rutting (mm), initial
fatigue life (cycles)
Long-term: Rut depth progression (mm), fatigue life retention (%)
Data collected followed trends reported in the literature (e.g., Caltrans, 2015; Lo Presti, 2013; Mashaan et al.,
2014), ensuring realistic interactions between CR%, climate, traffic, and performance indices.
Laboratory Performance Modeling
Laboratory test data were analyzed using:
Multiple Linear Regression (MLR) to model relationships between CR%, binder content, air voids, and
mechanical properties.
Quadratic regression to capture nonlinear trends in Marshall Stability and other indices.
Artificial Neural Networks (ANN) to model complex interactions where linear models were insufficient.
Genetic Algorithm (GA) optimization to identify CR% ranges maximizing mechanical properties while
satisfying multi-objective constraints (e.g., high stability, low flow).
Model
A regression model was developed to predict Marshall Stability (S) using input variables:
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 23
www.rsisinternational.org
CR% (X₁), Binder Content (X₂), Air Voids (X₃), Temperature (X₄), and AGI (X₅)
S = ß0+ ß1X1 + ß2X2 + ß3X3+ ß4X4 + ß5X5 + ϵ
The model was trained on 80% of the dataset and tested on the remaining 20%, evaluated using R² and RMSE.
Long-Term Performance and Traffic Modeling
Long-term deterioration was modeled using nonlinear regression techniques:
Exponential, power law, and logistic models.
ANN models were also trained to predict rutting and fatigue retention using CR%, time, ESALs, and climate
zone as inputs.
Multi-objective optimization (e.g., via GA) was used to balance rutting and fatigue performance across
different conditions.
Quadratic regression was used for Marshall Stability as a function of CR%:
S = β0 + β1CR + β2CR2
Nonlinear regression (exponential model) was used for rutting:
R = αe−cCR
Nonlinear regression (logistic model) was used for fatigue life:
F =
Fmax
1+e−k(CR−CR0)
Enhanced long-term models incorporating traffic loading were developed as:
R(t, CR, L) = αtbe−cCRLγ
Fr(t, CR, L) = 100 e-dt (1 + f CR) L−η
Model parameters (e.g., α, b, c, γ, d, f, η) were estimated using nonlinear least squares regression via
MATLAB’s fitnlm function. The goodness-of-fit was assessed using R², root mean square error (RMSE), and
mean absolute error (MAE).
Multi-Objective Optimization (MOO)
To determine the optimal crumb rubber content per climatic condition, a multi-objective function was
developed to maximize desirable properties and minimize failures:
Score = 0.4 ·Fatiguenorm + 0.3 · Stabilitynorm - 0.2 · Rutnorm
The normalized values range from 0-1, and weights are assigned based on performance priorities for Nigerian
roads: fatigue cracking, rutting, and moisture damage.
The Genetic Algorithm (GA) in MATLAB was used for the optimization, considering:
Decision variables: CR%, Va%, Temp
Constraints: Air Voids ≤ 5.5%, Temp ≤ 170°C
Bounds: CR% ∈ [0, 20]
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 24
www.rsisinternational.org
Transfer Learning and Climate-Traffic Mapping
The calibrated U.S. models were transferred to Nigerian conditions through:
Climatic mapping: U.S. climate zones matched to Nigerian analogs (e.g., Arizona → North Nigeria, Florida →
South Nigeria). Nigerian climatic zones were paired with equivalent U.S. regions based on Köppen-Geiger
climate classifications and long-term meteorological data (rainfall, temperature, and humidity).
Nigerian Region Climate Type Matched U.S. Climate Reference CR%
Northern Nigeria (e.g., Kano) Hot-Dry Semi-Arid Arizona, Nevada 15-20%
Middle Belt (e.g., Abuja) Warm Semi-Humid Georgia, California 10-15%
Southern Nigeria (e.g., Lagos) Hot-Humid Florida 10-12%
Eastern Nigeria (e.g., Enugu) Tropical Rainforest Georgia, Florida 10-12%
Traffic mapping: Simulated ESAL ranges adjusted to reflect Nigerian highway and secondary road loading
data (e.g., 5–30 million ESALs over 15 years).
Parameter recalibration: Long-term performance model parameters (e.g., α, c, γ) were adjusted using Nigerian
traffic and climate characteristics to provide locally relevant CR% recommendations.
Validation
The dataset was randomly split:
80% for model training
20% for validation
Predictions on the validation set were compared to actual values to evaluate model accuracy. R², RMSE, and
MAE were reported for each model.
Data Presentation and Visualization
Results were presented as:
Scatter plots with fitted curves for Marshall Stability, rutting, and fatigue life vs. CR%
3D surface plots showing rutting and fatigue retention as functions of CR% and ESALs
Tables summarizing model parameters and validation metrics
RESULTS AND DISCUSSION
This section presents the analysis of the dataset across climate zones, regression and neural network modeling
performance, and the optimization of crumb rubber content. The findings inform the calibration of CRMA
design recommendations tailored for Nigerian conditions.
Laboratory Performance Models
Multiple Linear and Quadratic Regression Results
Multiple linear regression (MLR) and quadratic regression identified significant relationships between CR%,
air voids, binder content, and key mechanical properties:
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 25
www.rsisinternational.org
S = 9.8 + 0.32CR - 0.006CR2 + 0.15Va (R2 = 0.89)
where S is Marshall Stability (kN), CR is crumb rubber content (%), and Va is air voids (%).
Flow and ITS were also well-modeled using MLR (R² values of 0.85–0.88). Quadratic terms for CR%
improved model fit over simple linear terms.
Figure 1 below shows the relationship between Marshall Stability and CR% with both the quadratic fit (red
line) and ANN prediction (blue dashed line) overlaid on the data points.
CR(%) 0 5 10 15 20
Marshall Stability (kN) 10.099 11.492 12.402 12.915 12.696
S = -0.0070CR2 + 0.3500CR + 10.000 (S is Stability in kN)
Figure 1: Plot showing Marshall Stability vs. CR% with both the quadratic fit (red line) and ANN prediction
(blue dashed line) overlaid on the data points.
ANN and GA Modeling
ANN models captured non-linear interactions more effectively than MLR for some indices:
Marshall Stability ANN R²: 0.92
Flow ANN R²: 0.89
ITS ANN R²: 0.91
GA optimization identified optimal CR% ranges:
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 26
www.rsisinternational.org
10–12% CR for balancing high stability and low flow
12–15% CR for maximizing ITS
Figure 2 below shows Rutting vs CR% plots, showing exponential fit and ANN prediction (exponential
regression fit (red line) and ANN prediction (blue dashed line) overlaid on the data points.
CR (%) 0 5 10 15 20
Actual Rutting(mm) 7.99 6.10 4.79 3.75 3.00
Exponential fit (mm) 7.95 6.17 4.79 3.71 2.88
ANN Predicted Rutting (mm) ~7.95 ~6.20 ~4.80 ~3.75 ~2.85
From the fit, R(CR) = a·e−b∙CR ; where a = 7.95, b = 0.051. Final fitted equation: R(CR) = 7.95e-0.051CR
Figure 2: Rutting vs CR% - exponential + ANN overlay (exponential regression fit (red line) and ANN
prediction (blue dashed line) overlaid on the data points.
Long-Term Performance Models
Nonlinear Regression Models
Rut depth and fatigue retention were modeled as:
R(t, CR, L) = 1.5 t0.6 e-0.04CR L0.25 (R2 = 0.87)
Fr(t, CR, L) = 100 e-0.03t (1 + 0.02 CR) L-0.1 (R2 = 0.85)
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 27
www.rsisinternational.org
These models demonstrated:
Rutting reduced with increasing CR% and increased ESALs
Fatigue retention improved with CR%, slightly reduced by higher ESALs
Figure 3 below presents 3D surface plots showing rutting and fatigue retention as functions of CR% and
ESALs
CR
%
0 0 0 0 0 5.
3
5.
3
5.
3
5.
3
5.
3
1
0.
5
1
0.
5
10
.5
10
.5
10
.5
1
5.
8
1
5.
8
1
5.
8
1
5.
8
15
.8
2
0.
0
2
0.
0
2
0.
0
2
0.
0
2
0.
0
ES
AL
s
(M
)
10
.0
20
.0
30
.0
40
.0
50
.0
1
0.
0
20
.0
30
.0
40
.0
50
.0
1
0.
0
2
0.
0
30
.0
40
.0
50
.0
1
0.
0
2
0.
0
3
0.
0
4
0.
0
50
.0
1
0.
0
2
0.
0
3
0.
0
4
0.
0
5
0.
0
Ru
ttin
g
De
pth
(m
m)
10
.7
6
12
.7
7
14
.2
2
15
.3
8
16
.3
6
9.
1
7
10
.8
9
12
.1
2
13
.1
0
13
.9
2
7.
8
2
9.
2
9
10
.3
5
11
.2
0
11
.9
0
6.
6
8
7.
9
4
8.
8
4
9.
5
7
10
.2
0
5.
7
2
6.
8
0
7.
5
7
8.
2
0
8.
7
5
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 28
www.rsisinternational.org
Figure 3: 3D surface plot of rutting depth (mm) as a function of crumb rubber content (CR%) and cumulative
ESALs (million). The plot illustrates the combined effect of material modification and traffic load on long-
term rutting performance.
ANN and GA for Long-Term Performance
ANN models:
Rutting ANN R²: 0.88
Fatigue retention ANN R²: 0.86
GA multi-objective optimization balanced rutting minimization and fatigue retention maximization:Optimal
CR%: 15% for high ESAL corridors
Optimal CR%: 10–12% for lower ESAL secondary roads
Transfer Learning and Nigerian Calibration
The U.S.-derived models were recalibrated:
North Nigeria (Hot-Dry) → Optimal CR%: 12–15% for highways (>20M ESALs)
South Nigeria (Hot-Humid) → Optimal CR%: 10–12% for highways (10–20M ESALs)
Nigerian secondary roads (lower ESAL) models indicated 10% CR as generally sufficient.
Predictions
Region CR% ESALs (M) Years Predicted Rut (mm) Predicted Fatigue Retention (%)
Arizona 10 35 15 5.3 70
Nigeria North 15 25 15 4.5 74
Florida 15 25 15 4.9 73
Nigeria South 12 15 15 5.0 71
Literature Comparison
Performance This Study Literature Source
Stability +25–30% at 10–15% CR +20–30% Lo Presti (2013), Mashaan et al. (2014)
Rutting ~35–40% reduction ~30–40% Caltrans (2015), FHWA
Fatigue +30–40% at 15% CR +30–35% Putman & Amirkhanian (2004)
Key Insights
Lab models (MLR, ANN, GA) showed CR% improves mechanical properties up to an optimum range.
Long-term models (nonlinear, ANN) confirmed durability gains, especially under heavy traffic.
Integrated modeling allows climate-traffic-specific CR% recommendations for Nigeria.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 29
www.rsisinternational.org
CONCLUSION AND RECOMMENDATIONS
This study developed an integrated predictive modeling framework for crumb rubber modified asphalt
(CRMA) performance, combining laboratory-based analysis with long-term deterioration models under
varying climatic and traffic conditions. By using simulated data representative of U.S. states and their Nigerian
climatic analogs, the study demonstrated how laboratory properties and long-term field performance can be
optimized through advanced regression, artificial neural networks (ANN), genetic algorithms (GA), and multi-
objective optimization techniques.
The key findings are as follows:
Laboratory performance models using multiple linear regression, quadratic regression, and ANN showed that
crumb rubber content (CR%) significantly improves Marshall Stability, reduces flow, and increases indirect
tensile strength (ITS) up to an optimal range of 10–15%.
Long-term performance models incorporating nonlinear regression and ANN confirmed that CR% reduces
rutting depth and improves fatigue retention over 10–15 years of service. The inclusion of traffic loading
(expressed as ESALs) in the models highlighted the importance of traffic-specific design.
GA-based multi-objective optimization effectively balanced rutting resistance and fatigue durability, with
optimal CR% recommendations varying by climate and traffic intensity:
15% CR for high-traffic corridors (ESAL > 20 million)
10–12% CR for lower-traffic or secondary roads
Transfer learning calibration for Nigeria showed that CRMA technology can be effectively adapted to local
climates and traffic conditions. Optimal CR% recommendations were:
12–15% CR for highways in Northern Nigeria (Hot-Dry)
10–12% CR for highways in Southern Nigeria (Hot-Humid)
This combined modeling approach supports the sustainable deployment of CRMA in Nigeria and similar
developing regions, offering both environmental and performance benefits through the innovative reuse of
waste tires.
Recommendations
Field validation of the proposed models using Nigerian pavement monitoring data is essential to confirm
predictive accuracy.
Future work should extend the models to include effects of aging, moisture susceptibility, and maintenance
interventions.
Policymakers and highway agencies should consider adopting CRMA standards tailored to regional traffic and
climatic conditions, leveraging the findings of this study.
REFERENCES
1. Fazaeli, H., Karim, M. R., & Abdelaziz, M. (2016). Characterization of asphalt mixtures containing
crumb rubber. Construction and Building Materials, 122, 258–268.
2. Ghabchi, R., Singh, D., & Zaman, M. (2013). Performance evaluation of asphalt mixtures incorporating
crumb rubber. Transportation Research Record, 2371(1), 25–34.
3. Kaloush, K. E., Way, G. B., & Biligiri, K. P. (2002). Evaluation of asphalt rubber mixtures using
performance-based tests. Arizona DOT Research Report.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 30
www.rsisinternational.org
4. Liang, M., Xin, X., Fan, W., & Yu, J. (2022). Review of rubber-modified asphalt binder performance.
Advances in Materials Science and Engineering.
5. Lo Presti, D. (2013). Recycled tyre rubber modified bitumens for road asphalt mixtures. Construction
and Building Materials, 49, 863–881.
6. Mashaan, N. S., Ali, A. H., Karim, M. R., & Abdelaziz, M. (2014). A review on the use of crumb
rubber in asphalt pavement. Scientific Research and Essays, 9(9), 458–470.
7. Okonkwo, I. U., Ezeokonkwo, J. U., & Chukwu, D. O. (2022). Optimization of modified asphalt mix
design for tropical environments. Journal of Engineering and Applied Science, 69.
8. Putman, B. J., & Amirkhanian, S. N. (2004). Utilization of waste tires in asphalt binders. Journal of
Materials in Civil Engineering, 16(6), 623–630.
9. Way, G. B., Kaloush, K. E., & Biligiri, K. P. (2011). Asphalt rubber friction course overlays in
Arizona: performance and policy implications. Transportation Research Board.
10. Zhou, F., Hu, S., & Scullion, T. (2020). Machine learning framework for predicting asphalt pavement
performance. International Journal of Pavement Engineering, 21(10), 1220–1233.
11. Adebayo, T. M., & Akinpelu, J. A. (2017). Assessment of road failure in Nigeria: causes and remedies.
International Journal of Engineering Science Invention, 6(4), 10–17.
12. AASHTO. (2021). Standard method of test for performance-graded asphalt binder. AASHTO M320.
13. FHWA. (2020). Performance Evaluation of Recycled Tire Rubber Modified Asphalt. Federal Highway
Administration Report.
14. AASHTO Guide for Design of Pavement Structures. (1993). American Association of State Highway
and Transportation Officials.
15. ASTM D6927 – 15. (2015). Standard test method for Marshall stability and flow of asphalt mixtures.
16. Nigeria Meteorological Agency (NiMET). (2021). Annual Climate Report.
17. NAPA. (2018). Asphalt Pavement Industry Survey on Recycled Materials and Warm-Mix Asphalt
Usage. National Asphalt Pavement Association.
18. Ahmed, A., & Oladipo, F. O. (2020). Characterization of bitumen modified with rubber and polymers
in Nigeria. Journal of Engineering Research and Reports, 13(2), 18–26.
19. Ogunniyi, A. M., & Adesina, A. O. (2015). Traffic loading and pavement response models for flexible
pavements in Nigeria. Nigerian Journal of Technology, 34(4), 858–865.
20. Hassan, M. M., & Issa, C. A. (2019). Applications of Artificial Intelligence in Pavement Engineering.
Innovative Infrastructure Solutions, 4(1), 1–8.
21. Federal Ministry of Works and Housing (Nigeria). (2022). Pavement Design Guidelines for Federal
Roads.
22. ASTM D4123 – 82. (2009). Standard test method for indirect tensile strength of asphalt mixtures.
23. Solaimanian, M., & Kennedy, T. W. (1993). Predicting temperature susceptibility of asphalt binders.
Journal of the Association of Asphalt Paving Technologists, 62, 329–347.
24. Huang, Y. H. (2004). Pavement Analysis and Design. Pearson Prentice Hall.
25. Al-Qadi, I. L., & Elseifi, M. A. (2007). Development of performance-based specifications for asphalt
binders using finite element analysis. Transportation Research Record, 1997(1), 123–131.
26. Babalola, O. (2021). Assessing the Feasibility of Tire-Rubber Modified Asphalt in Tropical
Environments. African Journal of Civil Engineering, 12(3), 221–229.
27. Mashaan, N. S., & Karim, M. R. (2013). Rheological properties of rubber-modified bitumen. Procedia
Engineering, 53, 267–273.
28. Saboo, N., & Kumar, P. (2016). Fatigue performance of bituminous mixes containing crumb rubber
modified binder. Construction and Building Materials, 102, 190–197.
29. Wang, H., & You, Z. (2015). Performance evaluation of rubberized asphalt mixtures with warm-mix
additives. Construction and Building Materials, 74, 309–316.
30. FHWA TechBrief. (2018). Sustainable Asphalt Pavement Technologies. Federal Highway
Administration.