WHITEPAPERS
Cooperative Driving Automation: Is this the Missing Link for Widespread AV Deployment?
The rush to deploy fully automated driving functions is faced with increased scrutiny. Recent events have led to pushback from communities and agencies that fear robotaxis are not ready for prime time. Most AV deployments rely exclusively on environmental sensors and lack situational awareness with other road users including maintenance and emergency vehicles, law enforcement, vulnerable road users, and the general population at large.
SAE J3216, also known as Cooperative Driving Automation (CDA), is a relatively new approach to V2X that has the potential to enable AV deployment at scale.
Connected Car Roadmap: Technologies & Applications
The connected car roadmap started over 25 years ago with OnStar’s first deployment in 1996. Today, nearly all vehicles sold in the U.S. are connected. The early telematics vehicles were primitive compared to current technology, and there is much more technology and functionality on the way in the next decade. This whitepaper will explore what is emerging and what other future innovations may happen.
Why Lane Markings Matter: Where ADAS and Infrastructure Meet
Lane markings are the ultimate ground truth for automated driving systems. Any compromise in the contrast and/or placement of markings directly compromises the safety and performance of the automated driving system. VSI has been studying this for years having logged thousands of miles in the pursuit of understanding the limitations of automated driving systems. This paper examines common causes of poor lane detection that can compromise the safety of ADAS and/or automated driving systems. This is largely the detection of lane markings but often within the context of other factors such as lighting, weather, geometry, merges, exits, interchanges, and more.
Ecosystem Analysis: The Building Blocks of Automated Vehicles
Developing automated vehicle systems is a complex endeavor for anyone trying to compete in this space. OEMs and traditional automotive suppliers have been very active through tie-ups, investments, and acquisitions designed to improve their strategic position. Large tech companies are also very active in developing complete platform strategies as well as aggressive investments through their venture funds. Beyond traditional auto and big tech companies, there are literally hundreds of other companies vying for a piece of the AV ecosystem.
The purpose of this report is to decompose the ever-changing AV ecosystem by looking at the latest version of VSI’s infographic, which reflects the major players within the value chain for autonomy. The report provides a high-level analysis of the global AV landscape by explaining each domain of the AV ecosystem.
Can Automated Driving Handle Minnesota Winters?
Active safety and automated driving systems rely on a variety of sensory inputs to make decisions about the environment the vehicle operates within. For ADAS, this generally includes one or more visible cameras plus radar. Together, they may enable Adaptive Cruise Control (ACC) and lane keep assist (LKA).
ACC and LKA are the basis of SAE Level 2 automated driving, what we consider to be the new ADAS. In fact, more than half of the cars sold today have a forward-facing camera and/or radar. Though applications like ACC and LKA are advanced systems, they are still highly dependent on environmental factors, one of which is inclement weather.
Crystal Group’s Autonomous Vehicle Compute System Powers Through Challenging VSI Labs ADAS Testing Cases
At the heart of all ADAS and AD development systems is a domain controller, also known as the central computer. The central computer in a development vehicle like VSI’s is no ordinary computer. For one thing, the computer must have a great deal of compute headroom to support network accelerators, GPUs, and storage. In addition, the onboard computer’s ability to operate in extreme conditions involving heat, shock, and vibration is also vital. The durability of the onboard computer is pushed further by extreme duty cycles associated with ADAS testing. Over the past year, VSI has been using computers from Crystal Group to handle the extreme conditions associated with ADAS/AD test and development. Some 25,000 miles later, VSI Labs has not experienced a single failure in its computer systems.
Destination ACM: Demonstrating Sensors Making the Unseen Visible
VSI Labs visits the American Center for Mobility (ACM) once a quarter to showcase the ADAS/AD innovations from our sponsors. This is a unique opportunity for sponsors to demonstrate their innovation to industry and media and do so in an environment that is designed to do just that. On June 25 of this year, VSI hosted its second open house at ACM. What made this event special was some of the newer sensor technologies being shown to the industry for the first time. This included near-infrared imaging and long-range lidar. Both new sensor technologies bring capabilities to ADAS, and AD applications not seen before. Essentially, NIR can illuminate a scene better than anything else when faced with challenging environmental conditions.
The Automated Drive South: Highlights & Observations
The purpose of this report is to shed some light on a few of the key findings. Specifically, we would like to make some observations about automated lane-keeping using the latest technology stack from Trimble and Ushr. We also discovered some interesting sensor performance issues while driving in heavy rain.
AVs vs. Infrastructure: How Infrastructure Can Impede Performance
Active safety and automated driving systems are as good as the infrastructure around them. For proof of this try engaging lane keep assist on a dirt road! Pavement surface quality, lane geometry, lane markings, blend areas and intersections have a large impact on the performance of camera-based active safety and automated driving features.
Examination of Pedestrian Detection Systems: Benchmarking Thermal Camera Performance
In July of 2020 FLIR Systems, Inc. engaged VSI to test the world’s first fused automatic emergency braking (AEB) sensor suite, employing a thermal longwave infrared (LWIR) camera, a radar, a visible camera, and a convolutional neural network (CNN). VSI tested the fused AEB system along with the AEB systems deployed in four 2019 vehicles. All tests were conducted at the American Center for Mobility (ACM) in Michigan and were based on Euro New Car Assessment Programme (NCAP) positive detection tests. The VSI-designed tests included driving toward a soft pedestrian target (SPT) heated to mimic a human at 25 MPH in a variety of common driving conditions, which are not currently tested as part of NCAP. This article discusses the results of these tests.
The Role of Big Tech in AV Development: What Amazon, Apple, Google, and the Like are Doing to Advance the Industry
In the summer of 2020, Amazon made headlines after buying AV company, Zoox, for $1.2 billion. Upon first thought, one might think companies like Amazon have nothing to do with the automotive industry, let alone self-driving technology. In reality, driverless technology (and the greater business of mobility services) is simply too important to ignore for these big tech companies. Tech giants are well-positioned to participate in the business of self-driving cars; they have access to massive amounts of capital, deep software expertise, and all have an anchor position in the greater technology ecosystem. This article will decompose Amazon, Apple, Google, Microsoft, and Facebook's activities in the self-driving car space.
Calibrating Camera Sensors - What You Need to Know
Camera calibration is a process that will obtain parameters to describe a camera internally (intrinsically) and externally (extrinsically). Determining the exact location of cameras on a vehicle is vital to achieving stereo vision, enabling methods such as rectification and depth perception. As it becomes common for vehicles to come equipped with multiple cameras, camera calibration with be an important factor to consider during the automotive manufacturing and repair processes to ensure an accurate calibration for ADAS and automated vehicle functionality. This whitepaper provides foundational knowledge of camera calibration fundamentals, describes some commonly used calibration targets, and walks through the calibration process for mono or stereo cameras. The article also includes various tools that can be used for camera calibration and tips for a successful calibration based on VSI Labs’ extensive experience with camera calibration.
Understanding Euro NCAP - Setting the Standard for AEB Performance Validation
The pathway to automated feature enablement in series production vehicles begins with ADAS, specifically active safety systems that support lateral and/or longitudinal control. These safety systems form the basis of L1 automation that rely on sensors and sophisticated algorithms, one of which is automatic emergency braking (AEB) for pedestrians and bicyclists. Safety-critical systems like AEB are vital to test and validate. Euro NCAP has created testing standards to evaluate the safety and functionality of AEB systems from all different vehicle manufacturers. This report summarizes test conditions, procedures, and scenarios of Euro NCAP’s AEB testing. It also discusses the challenges and the outlook of Euro NCAP testing for more advanced active safety systems.