The autonomous vehicle (AV) industry continues to evolve across many aspects: technologies, use cases, safety standards, safety legislation, and many others. This column explores what key factors are driving and shaping the AV industry.
The figure below helps contextualize the current AV landscape by summarizing some of the key SAE definitions, hardware and software technologies, legislations, and regulations currently shaping the AV industry.
SAE AV terminology
SAE terminology is key to understanding AVs because it describes the capabilities each vehicle must have to be considered autonomous (either partially or fully).
- Operational design domain (ODD): Defines where AVs can operate and specifies use cases, including many types of passenger and goods transportation. ODDs will have different specifications for each use case and driving segment. ODDs require a standards effort (explained further below).
- Object and event detection and response (OEDR): Monitors and responds to the driving environment via vision and software driver platforms.
- Dynamic driving tasks (DDT): Driving tasks done by a software driver platform. This includes steering, speed control, and braking. It also includes the OEDR functions and the most important function — crash avoidance.
- Automated driving system (ADS): The hardware and software platforms that perform the entire DDT on a sustained basis for a specific ODD. The ADS consists of the AV computer system, including the AV sensor system and the software driver platform.
AV use cases
There are two segments — passenger transportation and goods transportation — each with multiple categories.
In the passenger transportation category, AVs have to replace or augment trips by individual vehicles and/or trips by mass–transit systems.
AVs for ride–hailing or robotaxis are currently the leading segment and opportunity for individual vehicle trips. Eventually, AV software technology will advance enough to allow personal AV deployment.
AVs for fixed routes are the main deployment option for mass-transit systems. The ISO 22737 Low–Speed Autonomous Driving (LSAD) standard is especially important for fixed–route AVs. Many mass-transit operators are exploring the introduction of vans and small buses for fixed–route AVs for existing bus routes.
In goods transportation, AVs have to replace or augment goods deliveries in three categories: first–, middle–, and last–mile transport.
First–mile delivery refers to the first stage of goods transportation. For a manufacturing company, this is from the factory to a customer warehouse. For a retailer, the first mile is from a large warehouse to a smaller local warehouse or store. A key characteristic is that first–mile delivery is primarily interstate or highway travel. Large autonomous trucks are the main AVs for this segment.
The middle mile is usually from a small warehouse or distribution center to a store or fulfillment center such as a customer pickup location. There is less highway or interstate driving for middle–mile deliveries, resulting in lower speeds and more complex traffic patterns. There can be some overlap in how first–mile and middle–mile are used. Small autonomous trucks or vans are the main AVs for this segment.
The last mile is delivery from a retail location or fulfillment center to the end customer’s home location. The route is primarily low–speed and may have complex traffic patterns in large cities. Suburban deliveries are where most tests and trials are deployed today. Much innovation has been developed for last–mile deliveries, including sidewalk AVs and on–road goods–only AVs.
The vision software platform completes event detection and object detection/recognition based on sensor data.
The software driver platform is the most crucial element because it has to complete all DDTs and OEDR functions. Ideally, the performance would be flawless; however, that is not yet feasible.
The sensor portfolio varies by use case, with robotaxis having as many as 30 cameras, 20 radar sensors, 16 light detection and ranging (LiDAR) sensors, and far–infrared sensors.
The sensor system cost for robotaxis is significant, primarily due to LiDAR prices. In 2020, an AV with $55K worth of hardware and a 30–sensor system would have as much as 80 percent of the total hardware cost invested in sensor tech. By 2025, while the same hardware will cost just $10,000, the sensors will remain 70 percent of the total hardware cost.
AV hardware is defined by the computer system and will require constant performance advances to meet the increasing demands of the software driver platform. Redundant computer system architecture is a must to prevent system failures.
AV computers must follow emerging technology and safety standards described below. The computers must also meet necessary cybersecurity and software legislation and compliance regulations.
Legislation and standards
As AV usage increases, new legislation and regulations are necessary to ensure the safety of those using and in proximity to AVs.
- AV road–use legislation: One of the most difficult legislation items is defining AV road use and authorization — essentially defining who can use AVs and where they are allowed to use them. One of the key questions is how much better AVs must be compared with human drivers both at deployment and during the AV’s lifetime. The solutions include difficult and controversial societal and political decisions on initial AV safety levels versus historical human–driver safety.
- Software legislation: Automotive software legislation is already emerging for software–defined vehicles but will be even more important for AVs. Cybersecurity and OTA software update legislation has already passed. Privacy legislations are also on the books with more expected due to detailed data availability thanks to extensive tracking to improve AV software driver platforms.
- AV road–use authorization: New laws, infrastructure, and AV safety–tracking systems are needed. Once AV Road legislation is set, infrastructure to authorize which AVs pass safety and are allowed on the road will become necessary. Then all AV safety must be tracked over their lifetime and upgraded via OTA software updates.
- Tech and safety standards: ISO 26262 is the core functional safety standard that is heavily used in all safety–critical systems.
ISO 21448 extends functional safety to AVs. UL4600 extends safety to AVs with no human intervention. IEEE P2851 is a formal, mathematical model that applies a technology–neutral approach for avoiding crashes, based on Mobileye technology.
The German OpenODD project is a future standard for describing ODDs and defining where AVs can drive. The goal is to create a machine–readable format to represent ODD specifications. An ODD should be valid through the operating life of an AV.
The ODD is used for the functional specification of AVs. It specifies what static and dynamic environmental parameters an AV must manage. They include all types of traffic participants, the weather conditions, the infrastructure, the location, the time of day, and everything else that can have an impact on driving situations.
AV development remains a leading focus for the automotive and transportation industries. This relatively fledgling vertical still requires significant technological advances that are on the way, even if they seem to move more slowly than predicted.
As expected, the legislation and regulations required to govern autonomous vehicles and their usage move more slowly than new technology is created. The AV industry is pushing for legislative actions in many countries with some recent success stories. There are indications in several countries around the world that legislative and safety regulations will happen soon. As they do, expect to see continued growth in the AV sector.
>> This article was originally published on our sister site, EE Times.
|Egil Juliussen has over 35 years’ experience in the high-tech and automotive industries. Most recently he was director of research at the automotive technology group of IHS Markit. His latest research was focused on autonomous vehicles and mobility-as-a-service. He was co-founder of Telematics Research Group, which was acquired by iSuppli (IHS acquired iSuppli in 2010); before that he co-founded Future Computing and Computer Industry Almanac. Previously, Dr. Juliussen was with Texas Instruments where he was a strategic and product planner for microprocessors and PCs. He is the author of over 700 papers, reports and conference presentations. He received B.S., M.S., and Ph.D. degrees in electrical engineering from Purdue University, and is a member of SAE and IEEE.|
- Multiple tech issues stall AV deployment
- The role of artificial intelligence in autonomous vehicles
- Artificial intelligence algorithms and challenges for autonomous vehicles
- What HD mapping brings to autonomous vehicles
- AV developers build on ROS framework
For more Embedded, subscribe to Embedded’s weekly email newsletter.