A Look at the Latest AV Advancements

Article By : Egil Juliussen

Autonomous vehicles are still considered a long way from primetime use, but there have been a lot of recent significant news announcements.

Autonomous vehicles (AVs) are still considered a long way from primetime use, but there have been a lot of recent significant news announcements. Advancements in robotaxis and robotrucks are plentiful, and recent developments from Argo, Waymo, and Nvidia will be interesting to keep an eye on. Slowly but surely, AVs are evolving.


Robotaxis continue to have the most activities and there is a lot of news from the U.S. and China.

Lyft and Argo.ai are now offering public robotaxi service in Austin, Texas. This is the second city for Lyft and Argo. They started operating a robotaxi service in Miami in December 2021 after several years of Argo and Ford testing.

Argo is testing its AVs in four more U.S. cities: Detroit, Pittsburgh, Palo Alto (California), and Washington. Argo is testing with Volkswagen in Munich and Hamburg, Germany.

Lyft is also operating a robotaxi service in Las Vegas with Motional. The service started in August 2022 with a safety driver and uses Hyundai Ioniq BEVs. Driverless robotaxi service is planned for 2023. Lyft and Motional are planning to launch the service in other cities in 2023. Lyft and Motional started testing AVs in Las Vegas in 2018 and have AV services for over 100,000 Lyft customers.

Cruise plans to started testing robotaxis in Austin and Phoenix by the end of 2022. Cruise has been testing goods deliveries in the Phoenix area with Walmart since 2020. Its main robotaxi effort is in San Francisco, where it has graduated to running a driverless commercial robotaxi service in parts of San Francisco. A previous column has data on Cruise’s AV testing in San Francisco.

Cruise has also started mapping Dubai, UAE, for a planned robotaxi service in 2023 using the Chevrolet Bolt AV. Dubai plans to convert 25% of its transportation journeys to self-driving trips by 2030 and use 4,000 robotaxis.

Mobileye started testing in Detroit in September 2022 using Nio ES8 BEVs. It is the first L4 using Mobileye Drive in the U.S. The system is based on Mobileye’s True Redundancy sensing (camera system independent from the LiDAR-radar system), Road Experience Management (REM) crowdsourced mapping, and Responsibility-Sensitive Safety (RSS) driving policy.

Mobileye Drive uses a suite of 26 sensors, including 11 cameras, six radars, three long-range LiDARs, and six short-range LiDARs.

Mobileye is a leader in AV development and has tested AVs in 20 cities in 10 countries. It plans to test robotaxi services with partners in Germany and Israel in a few months.

Baidu’s robotaxi service platform, Apollo Go, started driverless tests in Shanghai Driverless Demo Zone on Sept. 27. Apollo also received permission to start driverless, commercial robotaxi operation in two cities in August 2022. In Chongqing, China, Apollo is using five robotaxis in a 30-square-kilometer area. In Wuhan, China, Apollo is using five robotaxis in a 13-square-kilometer area.

Apollo is the Chinese leader in AV development, with over 20 million miles of testing and deployment. It has offered over 1 million public robotaxi rides in a dozen major Chinese cities as of July 2022. Baidu Apollo currently operates 500 robotaxis in China and plans to have 3,000 robotaxis in 30 cities by the end of 2023.


Kodiak Robotics started AV testing with Werner, a leading long-haul truck company. The robotruck tests were completed between Dallas and Lake City, Florida, to pilot 24/7 operation. Of the miles driven, 94% were done in autonomous mode. Autonomous trucking lanes were used to showcase how efficiently autonomous trucks can be used with a transfer hub model at truck ports. Werner is in the top 15 list of logistics companies.

Plus and Iveco announced completion of the initial phase of their autonomous trucking pilot in Italy. The next phase of the pilot is to conduct public road testing in multiple countries in Europe, including Austria, Germany, Italy, and Switzerland.

Gatik has signed a contract with Pitney Bowes for goods delivery in the Dallas–Fort Worth area. Gatik will make multiple daily deliveries starting in 2023. The goal is to speed deliveries and lower transportation costs. Gatik is currently operating box truck deliveries to over 30 Sam’s Clubs in the Dallas area.

Gatik also did the first driverless delivery to a Loblaw store in Canada. Since January 2020, Loblaw and Gatik have transported over 150,000 autonomous deliveries with a safety driver, with a 100% safety record.

Delivery AVs

Nuro is the leader in goods-only AVs that focus on deliveries. Nuro received a 10-year deal with Uber to deliver meals and other goods to Uber Eats customers. The deliveries will start in Houston and Mountain View, California, and expand to other areas. Nuro has other important delivery customers, such as Walmart and Kroger.

Starship Technologies is the leader in sidewalk AV deliveries. It is especially strong in delivering fast-food meals on university campuses. Starship recently surpassed 3.5 million deliveries and has driven over 4 million miles worldwide.

Argo AV platform

In mid-September, Argo.ai introduced its suite of products and services for AVs to enhance ride-hailing or goods-delivery operations. The products offer businesses in a range of industries to deploy self-driving vehicles. There are multiple products:

  • Argo Concert: The gateway to Argo’s product ecosystem. It uses a suite of application programming interfaces to integrate Argo-powered AVs into customers’ existing operations.
  • Argo Autonomy Platform: It consists of the software, hardware, high-definition maps, and backend support implemented as four products:
    • Argo Drive: the self-driving software and hardware that manages all aspects of vehicle operation
    • Argo Lidar: Argo’s proprietary long-range LiDAR with a range of over 400 meters with camera-like resolution
    • Argo Map: a high-definition map network with street-level knowledge of roadways, traffic infrastructure, speed limits, and more
    • Argo Hub: a suite of cloud tools and infrastructure to support AVs in the field
  • Argo Autonomy Solutions: Running commercial autonomous services requires new operational and fleet management tools:
    • Argo Operations: a set of services to help customers deploy and operate AVs
    • Argo Fleet: products to maximize fleet uptime and meet the service and operational needs of AVs
  • Argo Autonomy Data and Analytics: Driving data collection and analysis to provide future improvements of Argo’s products. The data also provides real-time traffic and road insights, which will be useful to city governments, logistics companies, and other stakeholders.

The Argo Autonomy platform is trained with data from extensive testing in eight cities across the U.S. and Germany.

Waymo Driver safety simulation

Waymo created a reference model for a non-impaired, always attentive driver. The human driver reference model was used to simulate crash avoidance and for comparisons with the Waymo Driver. All of these simulated events were reconstructed fatal crashes that occurred during a 10-year period in Chandler, Arizona, which is the operational design domain where Waymo ADS is operating.

Waymo has also developed a framework for measuring and modeling response times in traffic conflicts applicable to automated driving systems and other traffic safety domains.

Two papers describing Waymo’s research efforts are available as downloadable PDF files. The papers are called:

  1. Collision Avoidance Effectiveness of an Automated Driving System Using a Human Driver Behavior Reference Model in Reconstructed Fatal Collisions (27 pages)
  2. Modeling Road User Response Timing in Naturalistic Traffic Conflicts: A surprise-based framework (20 pages)

Waymo ran simulation using these models to assess how the Waymo Driver performs compared with a non-impaired, alert driver. The Waymo Driver did much better and avoided 75% of the reconstructed fatal crashes and reduced serious injury risks by over 90%. The first technical paper has much more details and explanation.

Chips for AVs

Nvidia announced its next-generation processor for AVs on Sept. 20. Thor, the automotive-grade SoC, is built on the Arm Grace CPU and Nvidia Hopper GPU with performance of 2,000 teraflops. Thor replaces the Atlan SoC, which was announced in April 2021, with performance of 1,000 teraflops.

Thor is expected to be available in MY 2025 vehicles—the same year as Atlan was promised. Nvidia believes Thor can be used for both AV and infotainment functionality and can be used for multi-domain ECU consolidation.

With Thor as the top end of its Drive product family, Nvidia will retain its lead in the automotive SoC performance race. There are questions of why such high performance is needed. In my long experience in the IT and automotive industry, those questions are always answered by new software and applications that find multiple ways of leveraging extra performance.


AV progress has been disappointing due to over-optimistic expectations in the last five years. AVs have followed the Gartner hype curve, as many new technologies do. But where on the hype curve is AV technology now? I think AVs are well past the Peak of Inflated Expectations and recently went through the Trough of Disillusionment. Wikipedia explains the hype curve here.

I think all of the activities in the above table show progress, and this is just a listing of key events in the last month. The AV industry still has to go through the Slope of Enlightenment, and that will take five or more years. To reach the Plateau of Productivity, AV technology probably has another decade to go. What do you think?


This article was originally published on 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.


Subscribe to Newsletter

Leave a comment