Publicada el 23 de Octubre de 2020
Several years back I visited the Kennedy Space Center in Florida. Upon seeing the SpaceX launching pad, I could hardly envision a manned mission for the International Space Station (ISS) being launched from this very pad in the following years. Nevertheless, their success is not what caught my attention, but the fact that SpaceX has reduced the cost of the Space Shuttle to launch a kg into outer space by a factor of 10 (Moore’s law at its best!). Then, why shouldn’t we expect to have functionally reliable Level 5 autonomous vehicles offered at a reasonable price within this decade? The investment is huge, and the day to day evolution of the industry is significant. Some initiatives worth noting:
- Lidar: Despite their obvious advantages, they are too large, complex and expensive for mass market use. Additionally, they have a high susceptibility to vibration and shock, and feature limited resolution and range. This is about to change. Solid state lidar (Hartman, Continental HFL110, or Waymo Laser bear) fixes some of these constraints, at the time that provide both range and angular resolution, and they are close to mass production. Several of them can be mounted in a vehicle to provide the appropriate geometry to serve autonomous vehicles needs well. Particularly appealed by the integration of Luminar’s Lidar as part of the new Volvo’s Scalable Product Architecture (SPA2).
- Radar: Although still targeting levels 2 and 3, Automotive-grade high-resolution radar chipsets can receive data from multiple antennas and improved algorithms to handle interference are entering the marketplace.
- Location: On top of more conventional Global Navigation Satellite System/Inertial Measurement Unit (GNSS/IMU), Qualcomm’s visual odometry is promising trajectory drift below 1%, and 3rd Generation Partnership Project (3GPP) Rel 17 features location accuracy.
- HD Maps: The deployment of low orbit constellations of satellites offering global coverage of HD images, up to 10-inch accuracy, and 24h refreshment ratios are an intriguing initiative. We are already seeing key players emerging in this industry. Geely is planning to launch 500 satellites a year, Toyota working with Maxar, and Musk’s SpaceX is launching 12,000 as part of their Starlink constellation (although this last initiative seems to focus on satellite internet, it seems difficult to believe it will not have an automotive dimension considering the sponsor).
- Teleoperation: Worth to note here UMTRI’s research to combine on board Artificial Intelligence and Machine Learning (AI/ML) capable of predicting the likelihood of a disengagement in the coming 10 to 30 seconds, and a remote center able to take control if necessary.
These (and many other) initiatives would seem to draw a future in which uncountable autonomous entities will be able to share in an orderly manner structured and unstructured urban spaces with conventional cars, bicycles, pedestrians and so forth, with (at most) their own supervisory and teleoperation service, end of the story, right? Not so fast. This is not all.
Digital Infrastructure and 5G
Early this year, an article was published on LinkedIn with the intention to “deconstruct the myth of AV and 5G”. The title speaks for itself. The content was loud and clear, and the comments of the audience were overwhelming. I do not want to contribute to this debate, although there are some obvious biases (e.g. presenting Cellular Vehicle to Everything (C-V2X) as a subscription-based system needed for SIM card). The point I am trying to make here is about autonomous vehicles and digital infrastructure, of which 5G is one component.
What do I mean by digital infrastructure (in this context)? I mean a set of applications running on a combination of cloud and edge architectures (and supported by a variety of comms including satellite, Wi-fi, G4 LTE, G5 and the whole myriad of Vehicle to Everything -V2X) designed and operated not only to support the on board blocks of the autonomous driving platform (recognition, prediction, planning, situational awareness, and control), but also the needs of a mixed traffic flow which includes connected and non-connected vehicles and autonomous vehicles with different levels of automation. This infrastructure needs to offer Hardware and Software integrity, data security (Security Credential Management System -SCMS- will become a key asset), universal coverage and wide interoperability. It must be flexible enough as to be adapted to urban and interurban use cases, congestion and different traffic composition (with different levels of penetration of connected vehicles and autonomous vehicles mixed with platooning and cooperative driving). I am a little more skeptical of the need to move large volumes of data (infotainment aside), or to offer sub millisecond latencies (love the way Traffic Technology Services -TTS- is disseminating Signal Phase and Time -SPaT- traffic lights phases to Audi).
I know that I am treading on thin ice here, but it would seem to me that all this flurry of new and exciting technologies, comes to the transportation arena along with a whole packet of challenges, but this will be a topic for another blog.
3GPP: 3rd Generation Partnership Project
ADAS: Advanced Driver Assistant Systems
AI/ML: Artificial Intelligence/Machine Learning
AV: Autonomous Vehicle
BSM: Basic Safety Message
C-V2X: Cellular Vehicle to Everything
DMS: Dynamic Message Sign
FCC: Federal Communications Commission
GNSS/IMU: Global Navigation Satellite System/Inertial Measurement Unit
HD: High Definition
ISO: International Standard Organization
MVDS: Microwave Vehicle Detection System
NLOS: Non-Line of Sight
NPRM: Notice of Proposed Rulemaking
ODD: Operational Design Domain
OEM: Original Equipment Manufacturer
OTA: Over the Air Updates
RSS: Responsibility-Sensitivity Safety
SCMS: Security Credential Management System
SPA2: Scalable Product Architecture
SPaT: Signal Phase and Time
SWaP-C: Size, Weight, Power and Cost
TTS: Traffic Technology Services
UL: Underwriter Laboratories
UMTRI: University of Michigan Transportation Research
VOD: Visual Odometry
WtP: Willingness to pay