Density-based mixed platoon dispersion modelling with truncated mixed Gaussian distribution of speed

In China, urban arterial traffic presents a mixed-flow feature because the percentage of bus flow is relatively high. This affects the applicability of traditional platoon dispersion models, which are generally only suitable for homogeneous traffic flow. Based on field observations, this paper proposes a mixed platoon dispersion model (MPDM) to macroscopically simulate the mixed platoon dispersion process along the road segment between two successive signalised intersections from the density view. In order to capture the heterogeneity in mixed platoon speeds, the truncated mixed Gaussian distribution is adopted here to fit the speed data collected in the field, and expectation-maximisation (EM) algorithm is employed to estimate the distribution parameters. Later, the piecewise platoon density function is developed to examine the platoon dispersion characteristics. By applying this density function, the formulations of the expected number of vehicles in the front of the platoon that have passed and the expected number of vehicles at the rear of the platoon that have not passed a downstream intersection, as well as the downstream arriving flow function, are derived. Furthermore, numerical calculation for signal coordination verifies the effectiveness of the proposed MPDM.