The number of cars on the road and the need for transportation has increased rapidly during the past decades, leading to an increase of congestion. Traffic congestion and delays lead to high societal costs. Congestion leads to a capacity drop, i.e. the capacity of the road is reduced when congestion sets in. Due to limited financial and physical resources, it is desired to develop control measures that are able to reduce congestion and delays, and thus improve freeway throughput using existing infrastructure. Because congestion often sets in at fixed infrastructural bottlenecks, i.e. a part of the road at a fixed location at which the capacity is lower than the capacity at the other parts of the road, this research is focused on this type of bottleneck. Several traffic flow control measures have been the focus of recent research, of which some successful. One of these flow control measures is the use of speed limits, but a practically applicable control approach with improvement of throughput using speed limits as a flow control measure to prevent congestion and thus increase traffic flow at fixed infrastructural bottlenecks does not yet exist. The objective of this research is the development and evaluation of a controller that uses speed limits as a control measure, with the goal to improve freeway throughput by preventing congestion at a fixed infrastructural bottleneck. In order to reach this objective, first a literature study has been performed. This study was focused on existing approaches that use speed limits as a control measure, on different controller types and on approaches that use other control measures to control traffic flow. It was found that controlling traffic by using a dynamic speed-limited area is most promising. A density-based feedback controller has been chosen as most suitable controller for this research. After the literature study, a theory has been developed that explains the control approaches in traffic engineering terms, followed by an explanation in control engineering terms and algorithms for the developed controllers. In the theory chapter, it is explained that a speed-limited area is created with a certain desired density, which creates an outflow of the speed-limited area that is lower than the bottleneck capacity. Two different feedback controllers have been proposed to create this speed-limited area. For both of these controllers, algorithms have been developed. The first controller, feedback I, uses measurement data of the speed-limited area to calculate the average density, and compares this with a desired density value. The adjustment of the area is based on the difference between the measured average density and the desired density. The second controller, feedback II, compares the actual density with the desired density as well, but uses measurements upstream of the SL-area as well to determine the control action. The developed controllers have been evaluated both in a quantitative way and in a qualitative way by means of simulation. The second-order macroscopic simulation environment METANET is used for this purpose. The results of the evaluation show that both controllers show the expected qualitative behaviour, i.e. the flow into a fixed infrastructural bottleneck is reduced when the bottleneck is close to becoming active. This is done by generating a dynamic SL-area to control the flow. It is also shown that both controllers show a reduced total time spent compared to a situation without control, and thus an improved throughput. The improvement is between 10.8% and 23.9%. The results of the second feedback II controller are slightly better than the results of the feedback I controller. The conclusion of this research is that throughput can be improved by a variable speed-limited area. Because congestion often sets in at fixed infrastructural bottlenecks, the approach that is developed in this research could be used for field implementation. It is recommended to improve the control approach before it is implemented in the field. Several recommendations for this have been given in Section 6.2.