The growth of traffic densities has increased the concentrations of air pollutants, especially in city centres. Although the natural immune system of the human body can withstand low levels of exposure to air pollutants, exposure to high levels of air pollutants can lead to adverse health effects and even death. These high exposure levels are often present at city centres, due to high traffic densities at these areas. Therefore, in this research is explored if exposure levels to air pollutants can be decreased by improving traffic distributions. In this research the benefits and challenges are explored of a traffic load balancing system that has the objective to improve the air quality locally, while bounding the length of the detours that are needed to achieve this. Traffic dispersion is achieved by influencing routes of individual drivers by a collective navigation system. A weighted combination of travel time and air pollutant emissions is minimized to achieve lower air pollution exposure levels to citizens, while drivers maintain short travel times. Nitrogen Oxide (NOx) is used to bound the vehicle emissions locally, while the emission of Carbon Dioxide (CO2) is minimized globally. For each car an emission trace is constructed using an emission model. The emission traces are combined to form a time-dependent emission map. This time-dependent emission map is used for the exploration of routes that have a lower impact on the health of people that live near these routes. The approach is evaluated by simulation on real-life maps with an estimation of the traffic demand. The weighted combination approach is configured and compared to a base case where the drivers would only optimize their travel time. Using the simulation, it is shown that the exceedance of local thresholds of the NOx concentrations can be reduced, with only a small increase in travel time. Furthermore is explored what the effect is if only a part of the drivers would participate in the constructed load balancing system, while the other drivers would optimize their own travel time. By simulation is shown that even if only a part of the drivers would participate in this traffic load balancing system, the air quality can still be improved with a marginal cost of additional travel time.