Performance Assessment Through Capacity Estimation of Urban Uncontrolled Intersections Using Heuristic Techniques

Capacity estimation at uncontrolled intersections is vastly performed using probabilistic gap acceptance procedure (GAP) according to which the maximum sustainable hourly flow rate of minor stream vehicles is determined based on un-realistic assumptions and stochastic distribution of only major stream accepted or rejected gaps. In this paper, potential movement capacity has been estimated using the gap acceptance probabilistic procedure (GAP), the empirical regression-based deterministic (EMCAP) and the numerical additive conflict flow (ACF) technique. Three uncontrolled intersections with four-lane divided major roads (2 nos. three-legged and 1 no. four-legged) from urban regions of Bhopal, a city in Madhya Pradesh were chosen for this purpose, thus, having a total of 24 movement/potential capacities (2 nos. three-legged = 2 × 6 = 12 and 1 no. four-legged = 1 × 12 = 12). EMCAP model was created in SPSS and then validated after training adjustment factors in adaptive neuro fuzzy inference (ANFIS) tool keeping GAP capacities as base values. The difference in normalized mean absolute normalized errors (MANE) in volume to capacity (v/c) ratios for all three intersections during comparison of ACF and EMCAP with respect to GAP indicates that ACF (i.e., MAPE_GAP-ACF = 3.13% < MAPE_GAP-EMCAP = 4.85%) is more appropriate in estimating capacity appropriately, whereas EMCAP is more relevant in predicting level of service (with v/c ratios being > 0.80 operating in LOS E) than GAP, which suggests that there is a need to modify the current Indian Highway Capacity Manual guidelines considering uncontrolled intersections.