Traffic monitoring is a critical aspect of urban infrastructure management. With the advancement of technologies, traditional surveillance methods based on fixed sensor network systems could be potentially replaced by adaptive and easily redeployable systems, such as those based on drones. This paper wishes to contribute to the development of drones-based traffic monitoring and management systems by describing and evaluating a simulated swarm of drones monitoring traffic and communicating traffic data to adaptive traffic lights which adapt their green light duration to the current volume of traffic using the SPSA optimisation algorithm. A cell transition model (CTM) is used to simulate the behaviour, flow, and interactions of vehicles within a road network larger than most of networks used in similar studies. Evaluation tests compare the effectiveness of adaptive traffic unit with data generated by drones with a system of fixed duration signal traffic lights, and with an adaptive traffic unit with data generated by fixed cameras. The results shows that the optimised traffic lights system with data generated by drones is more effective than both the fixed signalling duration and the optimised system with data generated by fixed cameras in resolving traffic congestion due to a high volume of cars entering the road network. Further post-evaluation tests illustrate the limits of the adaptive traffic unit system with data generated by drones under a progressively higher volume of traffic entering the road network. We conclude the paper by discussing the current limitations of our model and by pointing to the most interesting directions for future work.
Alahvirdi et al. (Thu,) studied this question.