Increased travel times are often observed on urban roads, with signalized intersections being the major bottlenecks. One of the major factors for such increased delays is the inability of the traffic signals to respond to the demand fluctuations. An Adaptive Traffic Control System (ATCS) that adapts to the real time variations can improve the intersection performance. However, such a control system requires collection and communication of extensive amount of traffic data in real time and is often achieved through installation of location based sensors. Overcoming the limitations of location based sensors in terms of installation, operation and maintenance, this study proposes a mathematical framework of intersection delay optimization based on sample travel time information. The study collects travel time information of probe vehicles from mobile traffic sensors such as GPS/Bluetooth/Wi-Fi sensors to develop an ATCS under uniform and varying traffic demand for homogenous and lane disciplined traffic conditions. The proposed signal system quantifies and optimizes total delay for the current cycle to obtain optimal phase durations and adjust the signal timings for the next cycle accordingly. The performance of the developed strategy is evaluated with respect to the signal design when entire population data is available. With respect to such a base case, the probe-based model showed a difference of only 3.94% when computed theoretically and 9.61% when implemented in VISSIM. The results revealed that the proposed design is capable of handling traffic flow fluctuations without requiring the entire traffic stream data. The system demonstrated that sample data from four probe vehicles per phase is adequate for real-time optimal signal design. On performance evaluation, the results exhibited a strong correlation between the change in signal timings and the delay values within cycle and across cycles, reflecting the adaptability of the model to real-time traffic flow variations. In addition, it is observed that the proposed model ensures queue dissipation for near saturation situations along with delay reduction, without the need for changing the cycle length. The performance of the developed model against the field implemented signal design (Webster’s design), showed a delay reduction of 63.5% when developed in VISSIM.

Therefore, it can be inferred that the proposed signal design yields a better level of service on urban roads relative to the existing design procedures without collecting the whole traffic stream data. The developed model is then extended to incorporate mixed traffic characteristics to be suitable for implementation in developing countries with heterogeneous traffic conditions.