Abstract Ambulances and firefighting cars for example, cannot fritter away time waiting for traffic signals to turn green. These vehicles require a system that allows them to safely and quickly cross traffic lights. The researcher proposes a vehicle identification model-based smart traffic lighting system (STLS). The technology analyzes photos from cameras, counts the number of vehicles in each lane, and then determines which lane should be given priority taking into consideration waiting period, and type of service to be delivered by motorists. This study will use the CRISP-DM research approach to develop an accurate vehicle identification model for the system's success. Several algorithms (YOLO, SSD, Mask R-CNN, SqueezeDet, MobileNet and YOLOR) shall be implemented, and the best algorithm to produce the best car object detection, with at least 95% accuracy score, recall score of 95%, F1-score of 95% and support score of 95%. An algorithm for the system will also be designed in such a way that emergency service delivery vehicles are given priority and congestion is minimized. 1.1. Background Road accidents, fires, earthquakes, floods, and other calamities claim the lives of many people each year, causing them to lose their homes and leave them with permanent injuries or physical disabilities. This is exacerbated, if necessary, aid from emergency response vehicles such as ambulances or fire engines is delayed. It makes a tremendous difference in saving people's lives when an emergency response vehicle is delayed by a few minutes. According to Almuraykhi & Akhlaq (2019), 70% of the deaths may have been prevented if rescue teams had acted sooner. A system for emergency response vehicles has been devised. The basic concept is that when an emergency vehicle arrives at a traffic light, it will find it open and unaffected by other signals. The mechanism operates nearly imperceptibly, that is, without attracting the attention of others. To detect an oncoming vehicle, present systems use short-range technology (IR, optical sensors, RFIDs, Radar, and so on). When a car approaches a traffic signal, these devices can open it. Due to the short time available to open the required signal, there may be a conflict. Our proposed system, on the other hand, employs cameras that can detect emergency service delivery vehicles from afar and make a timely judgment on how to handle traffic lights at intersections along the path without interfering with other signals. 1.2. Statement of the problem The lack of modern and dynamic smart traffic lights contributes to congestion and traffic accidents at traffic lights junctions. This is evidenced by Qi et al. (2016) who reported that 30% to 60% of accident injuries and up to one third of fatalities occur at intersections. Kanungo et al. (2014) concurred that congestion is a serious problem and Barba et al. (2012) emphasized the need for road safety. 1.3. Aim To design vehicle identification model for a dynamic smart traffic lights system which grant access to either motorists in congestion or according to urgency-based priority, waiting period, and congestion level. 1.4. Objectives 1. To develop a model that can identify vehicles approaching the traffic lights. 2. To build an algorithm that can count the number of vehicles approaching the traffic lights. 3. To design an algorithm that make decisions on traffic lights signal values basing on waiting period, traffic velocity at traffic intersections, traffic density, and vehicle service delivery type. 4. To enhance traffic management at traffic intersections through the use of smart traffic lightning system. 5. To notify motorists about congested ways along their routes. 6. To determine an optimal route for motorists especially emergency service delivery motorists. 1.5. Research questions 1. What parameters can be used to train the best model that can identify vehicles approaching the traffic lights? 2. How to build an algorithm that determine the number of vehicles approaching the traffic lights? 3. What effect does waiting period, traffic velocity at traffic intersections, traffic density, and vehicle service delivery type have on making decisions on traffic lights signal values? 4. What are the most effective techniques to enhance traffic management at traffic intersections considering traffic lightning systems? 5. What are the best methods of notifying motorists about congested ways along their routes? 6. What determines an optimal route for motorists especially emergency service delivery motorists? 6.5. Significance of the study The success of this research will be an advancement in technology while saving lives and efficiently making use of time resource. Besides helping the public, the output of this research will contribute to the body of knowledge as other scientists and researchers will make use of this research in their projects. Research dealing with mobile vehicle networks, advancement in vehicle technologies, making traffic lights more intelligent, smart cities may need to have a look at the output of this research. This research is of great use to all those who may require data to be used for traffic monitoring at traffic lights. The data to be gathered in this research maybe used in big data analytics for instance, the data can be used to predict the occurrence of an event which require urgent response such as ambulance services and/or firefighting. 6.6. Literature review Lee and Chiu (2020) emphasized the importance of developing intelligent systems as we move towards building smart cities to reduce traffic congestion, and improve transport efficiency. Lee and Chiu (2020) proposed a system which could allow vehicles to communicate in such a way that if an emergency vehicle is detected, all vehicles are notified of its direction so that respective motorists can react accordingly. However, this does not nullify the need for traffic lights and traffic lights would need to synchronize with the messages communicated between vehicles. This is because it was found that decentralized systems are either not reliable, very costly, complex and not flexible as concurred by Tchuitcheu et al. (2020). Tchuitcheu et al. (2020) proposed a system that makes use of cameras to understand traffic flow in real-time. It is of paramount importance to note that this approach is one of the nascent technologies required in modern and dynamic world for visual machines. However, the researchers feel that, it would have been better if the centralized system is combined and synchronized with a decentralized system. Adding waiting period as a factor in deciding on which vehicles to cross the intersection first would also be beneficial to all motorists. Oza et al. (2020) delved into security issues related to sensor and controller networks. Therefore, high-breed systems of centralized systems with some independent systems require high security enforcement for them to successfully achieve their main goals. Hartanti et al. (2019) noted that traffic lights require proper timing which represent the real-time situations at traffic intersections. Therefore, factors such as road width, number of vehicles and queue size needs to be considered when granting access to cross intersections. 6.7. Methodology This research will be based on the implementation of the Cross-Industry Standard Process for Data Mining (CRISP-DM) research approach, which is an accepted standard and a non-industry-specific process model that is utilized for many data mining projects (Schröer et al., 2021). Research techniques are methods that a researcher employs when conducting research. These techniques include data gathering techniques, statistical data analysis tools, and methods for assessing the accuracy of the researcher's findings (Heath, 2021). Lawrence (2020) agrees that research methods refer to the researcher's approach to data collection. Data collection methods, analysis methods, outcome evaluation and testing methods are all discussed in this chapter. 6.7.1. To count the number of vehicles approaching the traffic lights junction automatically. A digital camera will be used to visualize vehicles approaching traffic lights intersections. The cameras will be installed in such a way that they visualize all ways used by vehicles to cross the intersection. The camera will be taking images and the images will be processed in real-time to give instant feedback signal. A model is going to be built using supervised machine learning. The model will take all images taken from all four ways approaching the intersection. The images will be compared to check the side and way with more vehicles than the other. The side with more vehicles is then given priority to pass the intersection. The granting procedure will follow the normal traffic lights principles of Red, Amber and Green, however, the timing will be minimized. The algorithm will compare the two pairs of opposite sides and if the total number of vehicles in one pair is more than that of the other pair and the waiting time for the other pair is less than ten (10) seconds, the pair with more vehicles will be granted access to pass through the intersection. However, if the number of vehicles for one pair is more than that of the other pair and the waiting period is above twenty seconds for at least three vehicles, the three vehicles will be granted access to pass through the intersection. 6.7.2. To identify urgent vehicles such as ambulances and fire brigade motorists with siren. To achieve this objective, the combination of a microphone and a camera will be utilized to identify urgent vehicles such as ambulances and fire brigades. These vehicles are normally found with a siren and strobing lights. Therefore, the camera will be checking for strobing lights and the microphone will be listen to siren sound. If both microphone and camera identify the urgent vehicle, the direction in which the vehicle is coming from will be granted access or all traffic lights goes red to stop all vehicles and allow the urgent motorist. If only siren is heard without observing the vehicle on the camera, all traffic lights go red to stop vehicles and allow urgent motorist. If only the camera identifies the urgent vehicle without siren, the vehicle will be treated as normal vehicles because it will not be on an urgent trip. 6.7.3. To measure vehicle’s distance from traffic lights automatically. Ultrasonic sensors will be used to measure vehicle distance from traffic lights. Vehicle’s distance before intersection is important because the researchers are trying to avoid instant stoppage of vehicles approaching intersection. They must be shown Amber before Red signal to allow motorists to stop safely. 6.7.4. To decide on which cars to go first To decide on which pair of ways is allowed to pass the intersection, the system uses the distance at which the vehicles are from the intersection, waiting period, the number of vehicles and the availability of urgent motorists approaching the intersection. Therefore, this objective depends on outcomes of the above objectives. 6.8. Proposed Research Organization 6.8.1. Chapter 1 This chapter details the background of the study, problem statement, research objectives, and significance of the study. It also covers how the research is to be covered and some brief notes on literature review and methodology. It also outlines how the whole research is organized. 6.8.2. Chapter 2 Reviewing literature is a crucial research activity required in all researches. This is because it gives the researcher a broader eye on the problem, it helps the researcher with alternatives in terms of methods of solving the problem at hand, it avoids reinventing the wheel, and guides the researcher in their researches. Chapter 2 will focus on literature covering artificial intelligence methods of automatically counting vehicles at traffic lights, and other smart traffic lights systems. This chapter will review journal articles, books and reliable internet sources. The aim of the chapter is to find the best methods of designing smart traffic lights and clarify on research gaps found in literature. The conclusion of this chapter determines this research’s methodology and product as well. 6.8.3. Chapter 3 In this chapter, the research technique, as well as the research design and philosophy, are explored. This chapter also covers ethical considerations, sample sizes, and the target demographic. Heath (2021) describes methodology as a method through which a researcher addresses a scientific problem in a systematic manner by following a set of stages to arrive at a solution. The underpinning for the entire project is research methodology, which supports the research methodologies (Lawrence, 2020). According to (Lawrence, 2020), research methodology entails the examination of research methodologies, assumptions, and principles. This chapter will focus on the implementation of the Cross-Industry Standard Process for Data Mining (CRISPDM) research approach, which is an accepted standard and a non-industry-specific process model that is utilized for many data mining projects (Schröer et al., 2021). Research techniques are methods that a researcher employs when conducting research. These techniques include data gathering techniques, statistical data analysis tools, and methods for assessing the accuracy of the researcher's findings (Heath, 2021). Lawrence (2020) agrees that research methods refer to the researcher's approach to data collection. Data collection methods, analysis methods, outcome evaluation and testing methods are all discussed in this chapter. 6.8.4. Chapter 4 In order to answer the study questions given in the first chapter of this research, this chapter is guided by the prior literature review chapter and theoretical framework. This chapter focuses on the results of the study and provides additional insight into the data acquired. The interpretation of the research findings will be aided by this dialogue. The CRISP-DM method was used to collect images data from the internet, YouTube videos, and movies, which was then analyzed to enhance model construction and ensure that no underfitting or overfitting occurs. The findings are presented in the form of figures and tables to help the reader grasp the information while also allowing tables and graphs to be easily related to the research objectives. 6.8.5. Chapter 5 Chapter 5 gives a summary of the research study, discussing the achievement of research objectives and areas of future study. It will explain the contributions the research has made from a theoretical and practical view. 6.9. Work plan Table 1¨Work-plan Table 2 ¨Work-plan Table 3 ¨Work-plan Figure 1