January 17, 2022
By Keith Shaw
As companies continue to grow, innovate and deploy their robotics and autonomous technologies, certain milestones are met. This week, two companies made announcements regarding achieving new goals – sidewalk delivery firm Serve Robotics said it was the first autonomous vehicle company to complete commercial deliveries at Level 4 autonomy; and WeRide announced it reached 10 million kilometers (6.2 million+ miles) of autonomous driving on public roads, with 2.5 million km in fully driverless mode.
Generally, I tend to ignore announcements like this, because the milestones are typically generated by the companies themselves once they’ve achieved the goal, and are done to promote further innovations or just give companies proof of their general awesomeness.
Sometimes, milestone announcements are done as a competitive one-upmanship showcase. A few years ago, companies developing autonomous mobile robots (AMRs) for the warehouse and fulfillment space made announcements around how many miles their robots traveled in warehouses during the holiday shopping season, or the number of orders filled. The time period, however, was never established, so one company proclaimed victory on Black Friday, while another company planted their flag on delivering the most during “Cyber Week” (Thanksgiving through Cyber Monday).
It’s all rather silly, of course, but sometimes there is value in these announcements. Stepping back from a 50,000-foot view of the announcement, you can get a sense of where the overall industry is with these achievements. With WeRide, for example, the 10 million km seems less impressive when you consider that Waymo announced in July 2021 that its vehicles have autonomously driven more than 20 million miles on public roads, as well as 20 billion miles in simulations.
But at least we know where WeRide stands in comparison to Waymo, even if the comparisons are not completely apples to apples. WeRide also claims 8 billion km in simulations, which it compares to the equivalent of the mileage that 300 autonomous vehicles would drive on public roads for 50 years.
Also, Level 4 autonomy now seems to be the Holy Grail for companies, rather than Level 5. The differences are slightly different – high automation versus full automation – and for most robotaxi services, it seems like Level 4 would be good enough for some regions. Even the SAE levels are open to interpretation and criticism, so relying on this for absolute truth is probably not recommended either.
In the Serve Robotics announcement, the company said one of its robots was able to navigate “fully autonomously in designated areas,” without giving more details on when, where or, most importantly, what was ordered.
The company did provide a video, however, in which you can watch to see the robot make its delivery, the view of a “supervisor station” that showed there was not a monitor/human there to do the driving, and the view from the robot’s camera.
I have some quick thoughts on the video:
- Wow, some sidewalks are really bad, so I’m somewhat impressed that the robot could navigate the bumps in the sidewalk. But clearly the order was not soup or a souffle.
- The delivery was done in a straight line in terms of the sidewalk. While the robot did cross a street, it never made a turn on the sidewalk. But I was happy to see that the robot waited for the walk signal in the crosswalk.
- But wait! The camera feed was on so the supervisor could monitor the crosswalk crossing. Was this truly “autonomous”?
- The delivery speed was slow enough that Serve needed to speed up the delivery to 4x speed in order to get a two-minute video. So that delivery probably took about seven minutes to happen.
- What the heck is that car doing there on the sidewalk? Was that planted?
All kidding aside, kudos to Serve for at least proving its claim with video. I hope more companies will provide some visual proof in the form of videos when they achieve their next milestone, like being able to autonomously deliver a pizza to my house in the dead of winter with all the snow on the ground and avoiding New England drivers. Then I’ll really know that the technology is here to stay.
Keith Shaw is the managing editor of Robotics-world.com.