Lattice-Based Dynamic Programming Trajectory Planning in an Intersection Crossing Scenario with a Connected Autonomous Electric Bus and Multiple Human-Driven Vehicles

Journal article, 2026

Building upon authors’ two recent papers published in 2025 and 2026 to reduce energy consumption, improve passenger comfort, and ensure trajectory tracking in an intersection crossing scenario involving a Connected Autonomous Electric Bus (CAEB) and a set of Human-Driven Vehicles (HDVs) resulting in problem setting and dynamic models of the CAEB and HDVs, this paper proposes an optimization approach for solving the optimization problem specific to Trajectory Planning (TP) in this scenario. The discrete-time TP for the CAEB is carried out on a spatial lattice obtained by segmenting the trip into road segments and discretizing the lanes and speed levels. The optimization problem is formulated as a lattice-based dynamic programming (DP) problem in the spatial domain, where each node represents a discrete vehicle configuration in relation to a certain road segment, lane index, and speed level, and a Bellman recursion is employed in order to determine the optimal sequence of lane changes and speed decisions that can minimize a segment-wise objective function involving the CAEB energy use, travel time, and penalizing of frequent lane change increments, under traffic lights and safety constraints. The resulting optimal discrete trajectory provides three reference profiles for the CAEB, namely the longitudinal position, the lateral position and the speed, defined at the level of road segments. The open-loop simulation results for six scenarios are reported in order to illustrate the behavior of the CAEB and HDVs, the impact of the traffic light modeling, and the effect of the lattice-based DP TP on the lane change behavior of the CAEB and the evolution of its speed.