Search Results

Research interests in autonomous driving have increased significantly in recent years. Several methods are being suggested for performance optimization of autonomous vehicles. However, weather conditions such as rain, snow, and fog may hinder the performance of autonomous algorithms. It is therefore of great importance to study how the performance/efficiency of the underlying scene understanding algorithms vary with such adverse scenarios. Semantic segmentation is one of the most widely used scene-understanding techniques applied to autonomous driving. In this work, we study the performance degradation of several semantic segmentation algorithms caused by rain for off-road driving scenes. Given the limited availability of datasets for real-world off-road driving scenarios that include rain, we utilize two types of synthetic datasets.

While convolutional neural networks (CNNs) have revolutionized ground-vehicle autonomy in the last decade, this class of algorithms requires large, truth-labeled data sets to be trained. The process of collecting and labeling training data is tedious, time-consuming, expensive, and error-prone. In order to automate this process, an automated method for training CNNs with simulated data was developed. This method utilizes physics-based simulation of sensors, along with automated truth labeling, to improve the speed and accuracy of training data acquisition for both camera and LIDAR sensors. This framework is enabled by the MSU Autonomous Vehicle Simulator (MAVS), a physics-based sensor simulator for ground vehicle robotics that includes high-fidelity simulations of LIDAR, cameras, and other sensors.

A dynamic modeling framework was established to predict status (position, displacement, velocity, acceleration, and shape) of a towed vehicle system with different driver inputs. This framework consists of three components: (1) a state space model to decide position and velocity for the vehicle system based on Newton’s second law; (2) an angular acceleration transferring model, which leads to a hypothesis that the each towed unit follows the same path as the towing vehicle; and (3) a polygon model to draw instantaneous polygons to envelop the entire system at any time point. Input parameters of this model include initial conditions of the system, real-time locations of a reference point (e.g. front center of the towing vehicle) that can be determined from a beacon and radar system, and instantaneous accelerations of this system, which come from driver maneuvers (accelerating, braking, steering, etc.) can be read from a data acquisition system installed on the towing vehicle.

The Powertrain Analysis and Computational Environment (PACE) is a powertrain simulation tool that provides an advanced behavioral modeling capability for the powertrain subsystems of conventional or hybrid-electric vehicles. Due to its origins in Argonne National Lab’s Autonomie, PACE benefits from the reputation of Autonomie as a validated modeling tool capable of simulating the advanced hardware and control features of modern vehicle powertrains. However, unlike Autonomie that is developed and executed in Mathwork’s MATLAB/Simulink environment, PACE is developed in C++ and is targeted for High-Performance Computing (HPC) platforms. Indeed, PACE is used as one of several actors within a comprehensive ground vehicle co-simulation system (CRES-GV MERCURY): during a single MERCURY run, thousands of concurrent PACE instances interact with other high-performance, distributed MERCURY components.

The Powertrain Analysis and Computational Environment (PACE) is a forward-looking powertrain simulation tool that is ready for a High-Performance Computing (HPC) environment. The code, written in C++, is one actor in a comprehensive ground vehicle co-simulation architecture being developed by the CREATE-GV program. PACE provides an advanced behavioral modeling capability for the powertrain subsystem of a conventional or hybrid-electric vehicle that exploits the idea of reusable vehicle modeling that underpins the Autonomie modeling environment developed by the Argonne National Laboratory. PACE permits the user to define a powertrain in Autonomie, which requires a single desktop license for MATLAB/Simulink, and port it to a cluster computer where PACE runs with an open-source BSD-3 license so that it can be distributed to as many nodes as needed.

EcoCAR 2: Plugging Into the Future is a three-year collegiate design competition challenging student teams to redesign a stock 2013 Chevrolet Malibu as a hybrid to improve its fuel economy and emissions. Mississippi State University, an eight-year veteran of AVTC competitions, has chosen to design a series-parallel plug-in hybrid electric vehicle. During the Year Two phase of the competition, the team has been implementing their design from Year One into the stock vehicle.

EcoCAR 2: Plugging In to the Future is a three-year design competition co-sponsored by General Motors and the Department of Energy. Mississippi State University has designed a plug-in hybrid powertrain for a 2013 Chevrolet Malibu vehicle platform. This vehicle will be capable of 57 miles all-electric range and utility-factor corrected fuel economy of greater than 80 miles per gallon gasoline equivalent (mpgge). All modifications are designed without sacrificing any of the vehicle's utility or performance. Advanced modeling, simulation, and Hardware-in-the-Loop (HIL) simulation capabilities are being used for rapid control prototyping and vehicle design to ensure success in the following years of the competition.