Search Results

Dynamic programming (DP) has been used for optimal control of hybrid powertrain and vehicle speed optimization particularly in design phase for over a couple of decades. With the advent of autonomous and connected vehicle technologies, automotive industry is getting closer to implementing predictive optimal control strategies in real time applications. The biggest challenge in implementation of optimal controls is the limitation on hardware which includes processor speed, IO speed, and random access memory. Due to the use of autonomous features, modern vehicles are equipped with better onboard computational resources. In this paper we present a comparison between multiple hardware options for dynamic programming. The optimal control problem considered, is the optimization of travel time and fuel economy by tuning the torque split ratio and vehicle speed while maintaining charge sustaining operation.

Due to the infeasibility of exhaustive on-road testing of Automated Vehicles (AVs) and vehicles with Advanced Driver Assistance Systems (ADAS), virtual methods for verification and validation of such vehicles have gained prominence. In order to incorporate the variability in the characteristics of test scenarios such as surrounding traffic, weather, obstacles, road network, infrastructure features, etc., as well as provide the option of varying the fidelities of subsystem models, this study discusses a modular software block-set for virtual testing of AV/ADAS controllers based on open source tools. The core concept is to co-simulate the traffic, vehicle dynamics, sensors, and the 3D scenes required for perception. This is achieved using SUMO (Simulation of Urban MObility, a microscopic road-network-based traffic generation tool) and Unreal Engine (for 3D traffic flow generation).

Testing any new safety technology of Autonomous Vehicles (AV) and Advanced Driver Assistance Systems (ADAS) requires simulation-based validation and verification. The specific scenarios used for testing, outline incidences of accidents or near-miss events. In order to simulate these scenarios, specific values for all the above parameters are required including the ego vehicle model. The ‘criticality’ of a scenario is defined in terms of the difficulty level of the safety maneuver. A scenario could be over-critical, critical, or under-critical. In over-critical scenarios, it is impossible to avoid a crash whereas, for under-critical scenarios, no action may be required to avoid a crash. The criticality of the scenario depends on various parameters e.g. speeds, distances, road/tire parameters, etc. In this paper, we propose a definition of criticality metric and identify the parameters such that a scenario becomes critical.

This research leverages publicly available crash data to construct safety-critical scenarios focusing primarily on Level 3 Automated Driving Systems (ADS) safety assessment under highway driving conditions. NHTSA’s Crashworthiness Data System (CDS) has a rich dataset of representative crashes sampled from numerous Primary Sampling Units (PSUs) across the country. Each of these datasets includes the storyline, road geometry information, detailed description of actors involved in the crash, weather information, scene diagrams, crash images, and a myriad of other crash-specific details. The methodology adopted aims to generate critical scenarios from real-world driving to complement the existent regulatory tests for the validation of L3 ADS. For this work, a four-step approach was adopted to extract safety-critical scenarios from crash data.

For realistic traffic modeling, real-world traffic calibration data is needed. These data include a representative road network, road users count by type, traffic lights information, infrastructure, etc. In most cases, this data is not readily available due to cost, time, and confidentiality constraints. Some open-source data are accessible and provide this information for specific geographical locations, however, it is often insufficient for realistic calibration. Moreover, the publicly available data may have errors, for example, the Open Street Maps (OSM) does not always correlate with physical roads. The scarcity, incompleteness, and inaccuracies of the data pose challenges to the realistic calibration of traffic models. Hence, in this study, we propose an approach based on spatial interpolation for addressing sparsity in vehicle count data that can augment existing data to make traffic model calibrations more accurate.

Increasing complexity of automotive systems along with growing safety and performance requirements, is causing development cycle costs to swell. A common solution is to use a Model-Based Design (MBD) approach, particularly using x-In the Loop (xIL) simulation methods for Validation and Verification (V&V). MBD allows efficient workflow from offline control design using high-fidelity models to real time V&V using Hardware-in-the-Loop (HIL) simulations. It is very challenging to reduce the complex non-linear high-fidelity models to real-time capable models for HIL simulation. Current literature does not provide a standard approach for obtaining the HIL-capable reduced model for complex non-linear systems. In this paper we present an approach to perform model reduction in light of HIL-level requirements. The approach is presented using an example of a 10-speed automatic transmission. The system constitutes three subsystems - the hydraulic network, mechanical gearbox, and torque converter.

Global regulatory targets and customer demand are driving the automotive industry to improve vehicle fuel efficiency. Methods for achieving increased efficiency include improvements in the internal combustion engine and an accelerating shift toward electrification. A key enabler to maximizing the benefit from these new powertrain technologies is proper systems integration work - including developing optimized controls for the propulsion system as a whole. The next step in the evolution of improving the propulsion management system is to make use of available information not typically associated with the powertrain. Advanced driver assistance systems, vehicle connectivity systems and cloud applications can provide information to the propulsion management system that allows a shift from instantaneous optimization of fuel consumption, to optimization over a route. In the current paper, we present initial work from a project being done as part of the DOE ARPA-E NEXTCAR program.