Imagery classification for autonomous ground vehicle mobility in cold weather environments

Autonomous ground vehicle (AGV) research for military applications is important for developing ways to remove soldiers from harm’s way. Current AGV research tends toward operations in warm climates and this leaves the vehicle at risk of failing in cold climates. To ensure AGVs can fulfill a military vehicle’s role of being able to operate on- or off-road in all conditions, consideration needs to be given to terrain of all types to inform the on-board machine learning algorithms. This research aims to correlate real-time vehicle performance data with snow and ice surfaces derived from multispectral imagery with the goal of aiding in the development of a truly all-terrain AGV. Using the image data that correlated most closely to vehicle performance the images were classified into terrain units of most interest to mobility. The best image classification results were obtained when using Short Wave InfraRed (SWIR) band values and a supervised classification scheme, resulting in over 95% accuracy.

Joint Chilean and US mobility testing in extreme environments

Vehicle mobility in cold and challenging terrains is of interest to both the US and Chilean Armies. Mobility in winter conditions is highly vehicle dependent with autonomous vehicles experiencing additional challenges over manned vehicles. They lack the ability to make informed decisions based on what they are “seeing” and instead need to rely on input from sensors on the vehicle, or from Unmanned Aerial Systems (UAS) or satellite data collections. This work focuses on onboard vehicle Controller Area Network (CAN) Bus sensors, driver input sensors, and some externally mounted sensors to assist with terrain identification and overall vehicle mobility. Analysis of winter vehicle/sensor data collected in collaboration with the Chilean Army in Lonquimay, Chile during July and August 2019 will be discussed in this report.

A Deep Reinforcement Learning Approach for Ramp Metering Based on Traffic Video Data

Ramp metering that uses traffic signals to regulate vehicle flows from the on-ramps has been widely implemented to improve vehicle mobility of the freeway. Previous studies generally update signal timings in real-time based on predefined traffic measurements collected by point detectors, such as traffic volumes and occupancies. Comparing with point detectors, traffic cameras—which have been increasingly deployed on road networks—could cover larger areas and provide more detailed traffic information. In this work, we propose a deep reinforcement learning (DRL) method to explore the potential of traffic video data in improving the efficiency of ramp metering. Vehicle locations are extracted from the traffic video frames and are reformed as position matrices. The proposed method takes the preprocessed video data as inputs and learns the optimal control strategies directly from the high-dimensional inputs. A series of simulation experiments based on real-world traffic data are conducted to evaluate the proposed approach. The results demonstrate that, in comparison with a state-of-the-practice method, the proposed DRL method results in (1) lower travel times in the mainline, (2) shorter vehicle queues at the on-ramp, and (3) higher traffic flows downstream of the merging area. The results suggest that the proposed method is able to extract useful information from the video data for better ramp metering controls.

Characterizing snow surface properties using airborne hyperspectral imagery for autonomous winter mobility

With changing conditions in northern climates it is crucial for the United States to have assured mobility in these high-latitude regions. Winter terrain conditions adversely affect vehicle mobility and, as such, they must be accurately characterized to ensure mission success. Previous studies have attempted to remotely characterize snow properties using varied sensors. However, these studies have primarily used satellite-based products that provide coarse spatial and temporal resolution, which is unsuitable for autonomous mobility. Our work employs the use of an Unmanned Aeriel Vehicle (UAV) mounted hyperspectral camera in tandem with machine learning frameworks to predict snow surface properties at finer scales. Several machine learning models were trained using hyperspectral imagery in tandem with in-situ snow measurements. The results indicate that random forest and k-nearest neighbors models had the lowest Mean Absolute Error for all surface snow properties. A pearson correlation matrix showed that density, grain size, and moisture content all had a significant positive correlation to one another. Mechanically, density and grain size had a slightly positive correlation to compressive strength, while moisture had a much weaker negative correlation. This work provides preliminary insight into the efficacy of using hyperspectral imagery for characterizing snow properties for autonomous vehicle mobility.

Mitigating Broadcasting Storm Using Multihead Nomination Clustering in Vehicular Content Centric Networks

Vehicles are highly mobile nodes; therefore, they frequently change their topology. To maintain a stable connection with the server in high-speed vehicular networks, the handover process is restarted again to satisfy the content requests. To satisfy the requested content, a vehicular-content-centric network (VCCN) is proposed. The proposed scheme adopts in-network caching instead of destination-based routing to satisfy the requests. In this regard, various routing protocols have been proposed to increase the communication efficiency of VCCN. Despite disruptive communication links due to head vehicle mobility, the vehicles create a broadcasting storm that increases communication delay and packet drop fraction. To address the issues mentioned above in the VCCN, we proposed a multihead nomination clustering scheme. It extends the hello packet header to get the vehicle information from the cluster vehicles. The novel cluster information table (CIT) has been proposed to maintain several nominated head vehicles of a cluster on roadside units (RSUs). In disruptive communication links due to the head vehicle’s mobility, the RSU nominates the new head vehicle using CIT entries, resulting in the elimination of the broadcasting storm effect on disruptive communication links. Finally, the proposed scheme increases the successful communication rate, decreases the communication delay, and ensures a high cache success ratio on an increasing number of vehicles.

A Practical Method of Optimal RSUs Deployment for Smart Internet of Vehicle

The roadside units (RSUs) are absolutely indispensable elements in sparse highway senario for smart Internet of Vehicle (IoV). Most recent RSU deployment methods consider vehicle mobility, warning message reception probability and time delay respectively, do not mentioned the situation once the accident vehicle cant send the warning message. In this paper, we present a comprehensive analysis that integrates all these factors above. In particular, we model three kinds of common highway scenarios and give the closed-form expression of RSUs deployment number along the highway. The proposed method has been validated by extensive simulations using Matlab and NS2, its performance has been compared with TAPC method. Results reveal that our proposed method has better performance under the condition of high warning message probability.

True Decommissioning: Removing Internal Combustion Engine Light-Duty Vehicles Permanently, Quickly, and Equitably

In this paper, we define True Decommissioning as the removal of internal combustion engine light-duty vehicles from the road permanently, quickly, and equitably. We discuss each interlinked component of True Decommissioning. We then outline the next steps for implementation, including engaging stakeholders, evaluating economic costs and benefits, and understanding policies and programs. Finally, we present a table of unanswered research questions in this area, including those our research group is working on. We welcome discussions on how we can achieve True Decommissioning and work together to facilitate an equitable transition to clean light-duty vehicle mobility for all.