Publications

Accurate modeling of bicycles in microsimulation tools is challenging due to the limited availability of detailed data, complexity of cyclist decision-making, and heterogeneity in cycling behavior. This paper proposes an agent-based bicycle simulation method in which generative adversarial imitation learning (GAIL) is used to infer the uncertain intentions and heterogeneous preferences of cyclists from observational data. The model is tested on videoderived data of cyclists on a unidirectional path in Vancouver, Canada. In cross-validation, multivariate distributions of movement variables such as speed, direction, and spacing are similar between observed and simulated cyclist trajectories. The model also performs well in comparison to two other cyclist simulation models from the literature. The proposed approach to agentbased microsimulation is a significant advancement, with continuous, non-linear, and stochastic representation of cyclist states, decisions, and actions. The enhanced consideration of cyclist diversity is necessary for developing bicycle networks for all ages and abilities of riders.

Although unconstrained cyclists may seem to move arbitrarily, it is hypothesized that their movements can be modeled given variables such as acceleration, change in direction and deviation from path centerline. Unconstrained cyclist trajectories can be thought of as correlated random walk movements, in which the state of the present time step is dependent on the previous time step. Hidden Markov Models (HMM) can be used to infer the hidden states that cyclist encounter at each time step given observations about their movement. The purpose of this research is to formalize a generative model that can describe unconstrained cyclist movement and be used to generate cyclist trajectories that are similar to observed ones. This model can then be combined with other constrained bicycle models to generate realistic interactions for bicycle microsimulation.

Reward functions are a key component in developing agent-based microsimulation models. The objective of this research is to estimate reward function parameters for cyclists in following interactions with other cyclists on bicycle paths. Decisions of cyclists (acceleration and direction) in following interactions are modeled as a finite state Markov Decision Process, in which the reward function describing the desired state of the cyclist is unknown. Two algorithms of imitation learning using Inverse Reinforcement Learning (IRL) are evaluated to estimate reward function parameters: Feature Matching (FM) and Maximum Entropy (ME) IRL. The algorithms are trained on 1297 cyclist trajectories in following interactions extracted from video data using computer vision, and then validated using a separate set of 349 trajectories. The estimated reward function parameters indicate how cyclists weigh the five state features in the reward function: speed, speed difference from leading cyclist, lateral position in path, lateral distance from leading cyclist, and longitudinal distance from leading cyclist. Following cyclists tend to prefer intermediate values of longitudinal and lateral distance to leading cyclists. Cyclists also prefer high speeds, with low speed difference from the leading cyclist and low deviation from the center of the path. Implementation of the reward functions derived from the FM and ME algorithms correctly predicted 58% and 67%, respectively, of the observed cyclist decisions (acceleration and direction) in the validation data set. This research is a key step toward developing operational bicycle traffic microsimulation models with applications such as facility planning and bicycle safety modeling.

A better understanding of cyclist behavior during various interactions is needed to enhance bicycle microsimulation models. This study aims to characterize cyclist maneuvers in following and overtaking interactions using multivariate finite mixture model-based clustering. Several variables that potentially affect cyclist state and future decisions are extracted from video data using computer vision techniques, including the longitudinal distance, lateral distance and speed difference between interacting cyclists. Observations of cyclists in following interactions are clustered into constrained and unconstrained states. Observations of overtaking cyclists are clustered into initiation, merging and post-overtaking states. Multivariate distributions within each cluster are examined, along with state transitions for each type of interaction. These characterizations are a key step toward development of agent-based bicycle traffic microsimulation models, which can be used to enhance bicycle facility planning and design, safety modeling, and energy modeling.

Traffic simulation has proved to be a vital tool for planning and operating transportation systems. Traffic simulation models need to be calibrated by adjusting model parameters to ensure the model’s ability to reproduce local traffic conditions and serve as a reliable test-bed for evaluating modification scenarios. This research developed and calibrated a mesoscopic traffic simulation model for the exceptionally large traffic network of Greater Cairo Region (GCR). The scope of the study is limited to calibrating traffic stream parameters, while a typical user equilibrium traffic assignment model was adopted. Open source traffic simulation software “DynusT” was used as a modeling platform. A wide range of field data was consolidated from previous related studies. The calibration procedure involved two levels: theoretical-based, and simulation based calibration. In the theoretical-based calibration stage, traffic stream parameters of the modified Greenshield’s traffic flow model was estimated using non-linear regression approach. On the other hand, the simulation-based calibration involved the estimation of the Anisotropic Mesoscopic Model parameter using a genetic algorithm optimization approach. A sensitivity analysis on estimated parameters values was conducted to verify the appropriateness of chosen values. Testing results revealed the potential of the adopted calibration approach and the credibility of estimated traffic stream parameters values. Limited discrepancy was observed between simulation-based link traffic volumes and actual ones in most observed links, with a normalized root mean square error (NRMSE) of 10.6 %.

Accurate modeling of bicycles in traffic calls for taking into account individualistic actions and behavioural schemes in different scenarios. Bicycle traffic is characterized by possessing significant unobserved heterogeneity, with the differences between individual cyclists driving how cyclists behave as a group. A Generative Variational Autoencoder-based model is developed to serve for two purposes. First, the encoder part summarizes the individual differences between unconstrained cyclists to a more compact latent dimension layer. Second, a decoder part where a trajectory is reconstructed using only coordinates of that compact latent layer. Different groups representing different styles of cyclists are identified using Gaussian mixture model (GMM) clustering. The model proved high reconstruction accuracy with a root mean square error of about 0.32 meters. Latent variables were clustered into four classes representing different styles of cyclists. When reflecting the clustering results on motion variables, such as speed, acceleration, jerk and directions angles, were shown to produce clearly different distributions. The model and results could help elevating the accuracy of bicycle microsimulation models and gain deeper understanding on the heterogeneity in bicycle traffic behaviour.

Understanding and modeling cyclist movement patterns is an essential step in developing agent- based microsimulation models. The aim of this study is to infer how cyclists in following interactions weigh different state features, such as relative distances and speeds, when making guidance decisions. Cyclist guidance decisions are modeled as a continuous state and action Markov Decision Process (MPD). Two Inverse Reinforcement Learning (IRL) algorithms are evaluated to estimate the MPD reward function in a linear form based on Maximum Entropy (ME) and in a nonlinear form based on Gaussian Processes (GP). The algorithms are trained on 856 cyclist trajectories in following interactions extracted from video data using computer vision, and then validated using a separate set of 172 trajectories. The estimated reward functions imply cyclist preferences for low lateral distances, path deviations, speed differences, accelerations and direction angles, but high longitudinal distances from leading cyclists. The mean and variance of the reward function learned using GP can be applied to simulate heterogeneous cyclist preferences and behavior. Predicted trajectories based on Q-learning with the linear and non-linear reward functions are compared to the validation data. This research is a fundamental step toward developing operational bicycle traffic microsimulation models with applications such as facility planning and bicycle safety modeling. Key novel aspects are the investigation of continuous, non- linear, and stochastic reward functions for cyclist agents using real-world observational data.