Network Efficiency and Velocity Decay in Asymmetric Urban Corridors in Buea, Cameroon: An Empirical Calibration Baseline Towards Future AI-Integrated Traffic Control

Abstract

Urban road networks in Sub-Saharan Africa diverge from the sensor-rich, lane-disciplined environments conventional models assume. This paper establishes an empirical baseline of network efficiency and velocity decay along the 8.0 km Mile 17 to Governor’s Roundabout corridor in Buea, Cameroon, marked by a 25 m/km inbound gradient, unsignalized intersections, and a taxi share of 59.55% inbound and 65.27% outbound. Fifteen-minute PCU counts were recorded at three points across three peak windows over three consecutive weekdays (1–3 April 2026), yielding 18 h of counts and 36 floating-car GPS runs. Volume-to-capacity ratios (v/c), Peak Hour Factors, and space-mean speeds were derived for both directions, with SD, coefficient of variation, and 95% confidence intervals computed to quantify sampling uncertainty. Results show near over-saturation (v/c: 0.98–1.06) at Bonduma, particularly during the afternoon outbound window. However, the 95% confidence intervals span v/c = 1.00, indicating significant uncertainty in Level of Service classifications. Mean Travel Time Ratios ranged from 1.10 to 2.04. At Bonduma in the evening, an elevated Informal Transport Disruption Factor (ITDF = 0.96) co-occurred with only moderate volume (v/c = 0.87), consistent with, though not proof of, informal-transport-related delay. Finally, the paper introduces two new indicators, the Directional Flow Asymmetry Index (DFAI) and Corridor Velocity Decay Rate (CVDR), to quantify directional imbalances that standard v/c analysis misses. Alongside the ITDF, these metrics serve as empirical calibration inputs for a future reinforcement learning traffic control framework (BACATC). Because no AI controller was built or evaluated in this study, the findings establish a quantified field baseline rather than a demonstration of AI-based control.