Transport Data Accuracy – Sharpening with Intelligence

This study investigates the accuracy level of traffic survey methods as a benchmark of how technology and equipment have improved accuracy levels over time. Hence, the hypothesis that we are trying to answer is “Transport Data Accuracy consistently sharpen by the level of intelligence applied over the years”. Different methods from past to the present include traditional techniques such as manual counts and pneumatic tubes, to semi-automated ones such as inductive loops, radar detectors, and video image processing (VIP). Newer technologies like advanced video image processing (AVIP), detection sensors, analytics were included. The accuracy levels of each method can be categorised according to time – the lowest is the manual counts at 65%, to the highest being VIP and AVIP at 95-98%. This shows that the accuracy level of data collection does increase over time, with its technological input directly correlated as well. This concludes that the hypothesis is true. As we look into the future of data collection, it looks set to maintain or increase as more intelligent inputs are fed into AI, machines to aid in their deep learning of different images/vehicles available on the road. Future research should focus on improving the robustness of AI algorithms, addressing privacy and security concerns, and developing cost-effective infrastructure solutions. It is no doubt that traffic data accuracy will continue to always be a matter of contention as end-users strive to apply such data for advance traffic management solutions.