Search Results

The rise of Software-Defined Vehicles (SDV) has rapidly advanced the development of Advanced Driver Assistance Systems (ADAS), Autonomous Vehicle (AV), and Battery Electric Vehicle (BEV) technology. While AVs need power to compute data from perception to controls, BEVs need the efficiency to optimize their electric driving range and stand out compared to traditional Internal Combustion Engine (ICE) vehicles. AVs possess certain shortcomings in the current world, but SAE Level 2+ (L2+) Automated Vehicles are the focus of all major Original Equipment Manufacturers (OEMs). The most common form of an SDV today is the amalgamation of AV and BEV technology on the same platform which is prominently available in most OEM’s lineups. As the compute and sensing architectures for L2+ automated vehicles lean towards a computationally expensive centralized design, it may hamper the most important purchasing factor of a BEV, the electric driving range.

Driver Assistance and Autonomous Driving features are becoming nearly ubiquitous in new vehicles. The intent of the Driver Assistant features is to assist the driver in making safer decisions. The intent of Autonomous Driving features is to execute vehicle maneuvers, without human intervention, in a safe manner. The overall goal of Driver Assistance and Autonomous Driving features is to reduce accidents, injuries, and deaths with a comforting driving experience. However, different drivers can react differently to advanced automated driving technology. It is therefore important to consider and improve the adaptability of these advances based on driver behavior. In this paper, a human-centric approach is adopted to provide an enriching driving experience. We perform data analysis of the naturalistic behavior of drivers when performing lane change maneuvers by extracting features from extensive Second Strategic Highway Research Program (SHRP2) data of over 5,400,000 data files.

Rapid adoption of battery electric vehicles means improving the energy consumption and energy efficiency of these new vehicles is a top priority. One method of accomplishing this is regenerative braking, which converts kinetic energy to electrical energy stored in the battery pack while the vehicle is decelerating. Coasting is an alternative strategy that minimizes energy consumption by decelerating the vehicle using only road load. A battery electric vehicle model is refined to assess regenerative braking, coasting, and other deceleration strategies. A road load model based on public test data calculates tractive effort requirements based on speed and acceleration. Bidirectional Willans lines are the basis of a powertrain model simulating battery energy consumption. Vehicle tractive and powertrain power are modeled backward from prescribed linear velocity curves, and the coasting trajectory is forward modeled given zero tractive power.

Structural Health Monitoring (SHM), especially in the field of rotary machinery diagnosis, plays a crucial role in determining the defect category as well as its intensity in a machine element. This paper proposes a new framework for real-time classification of structural defects in a roller bearing test rig using time domain-based classification algorithms. Along with the bearing defects, the effect of eccentric shaft loading has also been analyzed. The entire system comprises of three modules: sensor module – using accelerometers for data collection, data processing module – using time-domain based signal processing algorithms for feature extraction, and classification module – comprising of deep learning algorithms for classifying between different structural defects occurring within the inner and outer race of the bearing.

The significance and the number of vehicle safety features enabled via connectivity continue to increase. OnStar, with its automatic airbag notification, was one of the first vehicle safety features that demonstrate the enhanced safety benefits of connectivity. Vehicle connectivity benefits have grown to include remote software updates, data analytics to aid with preventative maintenance and even to theft prevention and recovery. All of these services require available and reliable connectivity. However, except for the airbag notification, none have strict latency requirements. For example, software updates can generally be postponed till reliable connectivity is available. Data required for prognostic use cases can be stored and transmitted at a later time. A new set of use cases are emerging that do demand continuous, reliable and low latency connectivity. For example, remote control of autonomous vehicles may be required in unique situations.

The latest edition of the US Department of Energy's (DOE) Advanced Vehicle Technology Competition (AVTC) series is the EcoCAR Mobility Challenge (EMC). In the third year of the EMC, the Mississippi State University (MSU) team developed and tested a perception system and a longitudinal controller to achieve SAE level 2 autonomy. Our team leveraged the model-based design approach to iterate between developing software components and executing tests in multiple environments in the loop (XIL) to verify that design requirements are met. This workflow allowed us to detect and resolve issues early in the development process. The perception system is composed of a sensor fusion and tracking algorithm. It relies on detections from a front facing camera and radar to generate tracks for a leading vehicle. The tracks from the perception system are used by a model predictive controller (MPC) to maintain a safe distance to the leading vehicle.

The EcoCAR Mobility Challenge (EMC) is the latest edition of the Advanced Vehicle Technology Competition (AVTC) series sponsored by the US Department of Energy. This competition challenges 11 North American universities to redesign a stock 2019 Chevrolet Blazer into an energy-efficient, SAE level 2-autonomous mild hybrid electric vehicle (mHEV) for use in the Mobility as a Service (MaaS) market. The Mississippi State University (MSU) team designed a P4 electric powertrain with an 85kW (113.99 HP) permanent magnet synchronous machine (PMSM) powered by a custom 5.4 kWh lithium-ion energy storage system. To maximize energy efficiency, Model Based Design concepts were leveraged to optimize the overall gear ratio for the P4 system. To accommodate this optimized ratio in the stock vehicle, a custom offset gearbox was designed that links the PMSM to the rear drive module.

A new and unique electric vehicle powertrain model based on bidirectional power flow for propel and regenerative brake power capture is developed and applied to production battery electric vehicles. The model is based on a Willans line model to relate power input from the battery and power output to tractive effort, with one set of parameters (marginal efficiency and an offset loss) for the bidirectional power flow through the powertrain. An electric accessory load is included for the propel, brake and idle phases of vehicle operation. In addition, regenerative brake energy capture is limited with a regen fraction (where the balance goes to friction braking), a power limit, and a low-speed cutoff limit. The purpose of the model is to predict energy consumption and range using only tractive effort based on EPA published road load and test mass (test car list data) and vehicle powertrain parameters derived from EPA reported unadjusted UDDS and HWFET energy consumption.

Four crash modes are overrepresented in traffic fatalities: run-off-road crashes, non-tracking run-off-road crashes, head-on crashes, and pedestrian crashes. Two advanced driver assist systems developed to help prevent tracking run-off-road crashes and head-on crashes are lane departure warning (LDW) and lane keeping assist (LKA). LDW acts to warn the driver when they are encroaching the lane boundary, whereas LKA performs automatic steering to prevent the vehicle from departing the lane. The objective of this research was to use real-world crash data to estimate current LDW and LKA system effectiveness in reducing run-off-road crashes and cross-centerline head-on crashes. All passenger vehicles that experienced a lane departure from 2017 to 2019 in the Crash Investigation Sampling System (CISS) were analyzed.

The autonomous collision avoidance problem for multi-trailer vehicle maneuvering is investigated in this paper. Different from conventional vehicle systems that contain one single moving part or multi-parts that can be considered as one rigid body, the interconnection between the tractor and each trailer, and interactions between trailers in the multi-trailer system introduce a high dimensional and highly complex dynamic system for the controller design. The external disturbance and parametric uncertainties further increase the difficulty in system identification and state space formulation. To implement a real time control system for various scenarios where the locations and states of the obstacles are not known beforehand, a supervisory algorithm is designed to convert the control problem to a discrete event system. The model predictive control (MPC) using limited lookahead policy is employed in the proposed algorithm.

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.

Recent developments in the area of autonomous vehicle navigation have emphasized algorithm development for the characterization of LiDAR 3D point-cloud data. The LiDAR sensor data provides a detailed understanding of the environment surrounding the vehicle for safe navigation. However, LiDAR point cloud datasets need point-level labels which require a significant amount of annotation effort. We present a framework which generates simulated labeled point cloud data. The simulated LiDAR data was generated by a physics-based platform, the Mississippi State University Autonomous Vehicle Simulator (MAVS). In this work, we use the simulation framework and labeled LiDAR data to develop and test algorithms for autonomous ground vehicle off-road navigation. The MAVS framework generates 3D point clouds for off-road environments that include trails and trees.

Vehicle headlamps and roadway lighting are the major sources of illumination at night. These sources affect contrast - defined as the luminance difference of an object from its background - which drives visibility at night. However, the combined effect of vehicle headlamps and intersection lighting on object contrast has not been reported previously. In this study, the interactive effects of vehicle headlamps and overhead lighting on object contrast were explored based on earlier work that examined drivers’ visibility under three intersection lighting designs (illuminated approach, illuminated box, and illuminated approach + box). The goals of this study were to: 1) quantify object luminance and contrast as a function of a vehicle’s headlamps and its distance to an intersection using the three lighting designs; and, 2) to assess whether contrast influences visual performance and perceived visibility in a highly dynamic intersection environment.

Intersection collisions currently account for approximately one-fifth of all crashes and one-sixth of all fatal crashes in the United States. One promising method of mitigating these crashes and fatalities is to develop and install Intersection Advanced Driver Assistance Systems (I-ADAS) on vehicles. When an intersection crash is imminent, the I-ADAS system can either warn the driver or apply automated braking. The potential safety benefit of I-ADAS has been previously examined based on real-world cases drawn from the National Motor Vehicle Crash Causation Survey (NMVCCS). However, these studies made the idealized assumption of full installation in all vehicles of a future fleet. The objective of this work was to predict the reduction in Straight Crossing Path (SCP) crashes due to I-ADAS systems in the United States over time. The proportion of new vehicles with I-ADAS was modeled using Highway Loss Data Institute (HLDI) fleet penetration predictions.

This paper focuses on the analysis of an EcoRouting system with minimum and maximum number of conditional stops. The effect on energy consumption with the presence and absence of road-grade information along a route is also studied. An EcoRouting system has been developed that takes in map information and converts it to a graph of nodes containing route information such as speed limits, stop lights, stop signs and road grade. A variable acceleration rate synthesis methodology is also introduced in this paper that takes into consideration distance, acceleration, cruise speed and jerk rate as inputs to simulate driver behavior on a given route. A simulation study is conducted in the town of Blacksburg, Virginia, USA to analyze the effects of EcoRouting in different driving conditions and to examine the effects of road grade and stop lights on energy consumption.

One of the principal goals of modern vehicle control systems is to ensure passenger safety during dangerous maneuvers. Their effectiveness relies on providing appropriate parameter inputs. Tire-road contact forces are among the most important because they provide helpful information that could be used to mitigate vehicle instabilities. Unfortunately, measuring these forces requires expensive instrumentation and is not suitable for commercial vehicles. Thus, accurately estimating them is a crucial task. In this work, two estimation approaches are compared, an observer method and a neural network learning technique. Both predict the lateral and longitudinal tire-road contact forces. The observer approach takes into account system nonlinearities and estimates the stochastic states by using an extended Kalman filter technique to perform data fusion based on the popular bicycle model.

Tire-pavement interaction noise (TPIN) is a dominant source for passenger cars and trucks above 40 km/h and 70 km/h, respectively. TPIN is mainly generated from the interaction between the tire and the pavement. In this paper, twenty-two passenger car radial (PCR) tires of the same size (16 in. radius) but with different tread patterns were tested on a non-porous asphalt pavement. For each tire, the noise data were collected using an on-board sound intensity (OBSI) system at five speeds in the range from 45 to 65 mph (from 72 to 105 km/h). The OBSI system used an optical sensor to record a once-per-revolution signal to monitor the vehicle speed. This signal was also used to perform order tracking analysis to break down the total tire noise into two components: tread pattern-related noise and non-tread pattern-related noise.

This paper presents a system designed to develop a high-resolution map of public roads by capturing high-resolution surface images. Unlike conventional system, the proposed system applies a field programmable gate array (FPGA) to synchronize camera, Inertial Measurement Unit (IMU), and Global Positioning System (GPS) by using FPGA’s high clock frequency and flexibility to multiple devices. The proposed system, which can be mounted on a regular vehicle, contains a Complementary Metal–Oxide–Semiconductor (CMOS) camera which can achieve 0.006 ms shutter speed and 150 fps frame rate. This camera’s high shutter speed and high frame rate can help capturing images with overlapping region at fast driving speed so that no area is missing from road surface image capturing.

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.

This paper presents a novel infrastructural traffic monitoring approach that estimates traffic information by combining two sensing techniques. The traffic information can be obtained from the presented approach includes passing vehicle counts, corresponding speed estimation and vehicle classification based on size. This approach uses measurement from an array of Lidars and video frames from a camera and derives traffic information using two techniques. The first technique detects passing vehicles by using Lidars to constantly measure the distance from laser transmitter to the target road surface. When a vehicle or other objects pass by, the measurement of the distance to road surface reduces in each targeting spot, and triggers detection event. The second technique utilizes video frames from camera and performs background subtraction algorithm in each selected Region of Interest (ROI), which also triggers detection when vehicle travels through each ROI.