A Driver Behavior Modeling Structure Based on Non-Parametric Bayesian Stochastic Hybrid Architecture

Abstract

Heterogeneous nature of the vehicular networks, which results from the co-existence of human-driven, semiautomated, and fully autonomous vehicles, is a challenging phenomenon toward the realization of the intelligent transportation systems with an acceptable level of safety, comfort, and efficiency. Safety applications highly suffer from communication resource limitations, specifically in dense and congested vehicular networks. The idea of model-based communication (MBC) has been recently proposed to address this issue. In this work, we propose Gaussian Process based Stochastic Hybrid System with Cumulative Relevant History (CRH-GP-SHS) framework, which is a hierarchical stochastic hybrid modeling structure, built upon a non-parametric Bayesian inference method, i.e. Gaussian processes. This framework is proposed in order to be employed within the MBC context to jointly model driver/vehicle behavior as a stochastic object. Nonparametric Bayesian methods relieve the limitations imposed by non-evolutionary model structures and enable the proposed framework to properly capture different stochastic behaviors. The performance of the proposed CRH-GP-SHS framework at the inter-mode level has been evaluated over a set of realistic lane change maneuvers from NGSIM-US101 dataset. The results show a noticeable performance improvement for GP in comparison to the baseline constant speed model, specifically in critical situations such as highly congested networks. Moreover, an augmented model has also been proposed which is a composition of GP and constant speed models and capable of capturing the driver behavior under various network reliability conditions.

Controlling Steering Angle for Cooperative Self-driving Vehicles utilizing CNN and LSTM-based Deep Networks

“…We then study the effect of changing the value of x on the performance of our model in terms of RMSE. We train our model at separate x values where x is set to 1,2,4,6,8,10,12,14,20 and computed the RMSE value for both the training and validation data respectively at each x value. The results were plotted in Fig.…”

V2X System Architecture Utilizing Hybrid Gaussian Process-based Model Structures

“…The regression problem here is equivalent to inferring the characteristics of the unknown target functions which have generated these timeseries through their available training sets, which are finite sets of known function output realizations. In this work, following our previous works in [16] [17], a non-parametric Bayesian inference framework is proposed to find an appropriate representation and abstraction of the driver behavior using his observed actions through the recorded time-series of the vehicle dynamics. In general, the main advantage of any non-parametric inference method is relaxing the function-specific characteristics during the learning process and letting the model complexity to be derived from and adapted to the available training set.…”

Representing Realistic Human Driver Behaviors using a Finite Size Gaussian Process Kernel Bank

“…This section is briefly describing our overall design and its core components, i.e., the error-driven MBC strategy and GP inference, which are essential for the rest of this work. Interested readers could refer to our previous works for further information [10], [11], [12], [13], [14], [15].…”

“…Recently, another paradigm, namely Model-Based Communications (MBC), has been proposed for the first time in [10] to address the scalability issue from a different point of view. This proposal, which has been more elaborated in [11], [12], [13], [14], [15], can be regarded as a PHY-and MACtechnology agnostic scheme, designed from the application layer perspective. The core concept of MBC proposes a new packet formation mechanism compared to the current Basic Safety Message (BSM) definition of the SAE J2735 standard.…”

“…Different modeling strategies have been explored in our previous works in order to find an appropriate scheme which could be integrated to our MBC architecture as its modeling sub-system. For instance, we have investigated Hierarchical Dirichlet Process-Hidden Markov Model (HDP-HMM) [12], and Gaussian Processes (GPs) [13], [14], [15], which both fall under the Bayesian non-parametric inference umbrella. Our main rationale for pursuing this field of inference mechanisms, i.e., non-parametric Bayesian, to solve our problem is its fascinating and important capability to model different behaviors without imposing any prior restrictions and requirements on the behavior characteristics.…”