How the Urban Robotics Foundation Is Creating a Technical Standard for Delivery Robots
Bern Grush talks ISO 4448 for public mobile robots
As the delivery robotics space continues to evolve, we’re apt to see certain form factors and functionalities win out over others. Concurrently, governments are also starting to pay closer attention to the industry - and with that comes increased regulatory scrutiny. One way to address both these intersecting issues is by creating a unified standard that could apply to the wide array of robots that are showing up on public streets and sidewalks. I sat down with Bern Grush, Executive Director of the Urban Robotics Foundation, to better understand how his organization is working to lead that charge.
Jonah Bliss: Can you start by giving us a bit of background on your work in the mobility industry, as well as the history of the Urban Robotics Foundation?
Bern Grush: I started out in the late 70s in computer vision, writing software to interpret images from the Landsat and SPOT satellite systems. Had it been 40 years later, I would have been writing the vision systems for automated vehicles and mobile robots instead — but I was born too soon! Fast forward to the early 2000s, I switched into the mobility industry to develop autonomous road-pricing and parking payment systems using GPS. After 15 years of that I found myself writing a book about deploying automated vehicles within a public transportation context and that led to an interest in goods delivery and naturally to the last-mile problem — hence delivery robots.
Because I had been repeatedly involved in standards drafting in those early machine vision, road pricing and autonomous parking days, I had become a “standards nerd.” So, when I saw the potential of mobile robots within an urban landscape, I immediately recognized that the lack of deployment standards would prevent the scale up of these devices in public spaces.
In 2019, I applied to the ISO to be project leader for the draft global standard now called TS4448 Intelligent transport systems - Public-area mobile robots (PMRs) and automated pathway devices. This initially started as a side-gig in my engineering consulting firm, and not long after I understood that ISO work needed to be moved to a dedicated environment. It had to be in a nonprofit in order to be funded – in our case by membership subscriptions. The Urban Robotics Foundation was established in 2021 to bring together stakeholder members to provide feedback and real-life experience to the standards development process. We are now about half way to completing the draft ISO 4448 suite to inform safe PMR integration into cities that want them.
JB: What is ISO 4448, what problem does it intend to solve, and what is the URF’s role in the process?
BG: The 4448 PMR standard directly addresses deployment, orchestration, behavioral, and governance issues. It focuses specifically on bystander safety, comfort and acceptance. It is only interested in device operation, configuration and safety insofar as it pertains to bystander safety. It is a deployment and operations standard, and a device capability standard, but not a manufacturing standard. There may be design and innovation requirements buried within deployment or behavioral clauses, but we try to stay as far away as possible from innovation and in fact we are required by the ISO to stay away from commercial IP. 4448 is focussed on what is workable for cities and for the active transportation users within those cities. Of course, we have to understand enough about what is technically feasible so that what the standard requires is achievable. That is one of the reasons we feel fortunate to have a number of technology members — they ensure that what we are asking on behalf of cities and active transportation users is doable.
The Urban Robotics Foundation (URF) was founded in 2021 to fund and strengthen project leadership. Our role is to bring stakeholders together to discuss, debate and explore the complexities involved in deploying PMRs in urban environments and to propose and document the standards that will help cities decide whether to reject or permit PMRs, in what areas and for what purposes. As subject matter experts, we also offer training workshops, webinars and will be publishing guidebooks to help with knowledge transfer.
JB: Which organizations are members of the URF? And are there any non-members that have had a hand in drafting the specification as well?
BG: We have members that are municipal governments, urban planners, traffic planners, technology companies, fleet operators, urbanist groups, accessibility advocates, researchers and, of course, the ISO working group that hosts this draft. In addition to the US and Canada, we have members and collaborators from over a dozen countries in Europe and Asia and we continue to build our membership through webinars, workshops and other outreach activities. In the early days of my work on ISO 4448, I solicited input from my global network but soon realized that a membership model was really the most appropriate way to move forward.
“…standards will have little impact on pilots or trials, but they matter very much at scale.”
JB: What’s the current timeline for the standard being finalized? And after it’s finalized, what does adoption and deployment look like?
BG: We think we are about halfway through, at this point. In March of this year, the ISO working group reviewed the draft and rearranged its parts into an order that would be funneled through the acceptance process of the ISO technical committee 204. We are expecting to see the 15-17 parts of 4448 enter that funnel beginning later this year and continuing over the following year and a half — 2024-2025.
It is harder to predict how adoption would work, but we believe it will mostly be the case that some municipalities, when asked to permit a fleet to operate at a scale beyond a pilot, will ask questions that will point to the need for a common body of standards for the municipality to judge whether it should permit — or ban – such fleets. That’s where we think standards adoption will begin. In other words, standards will have little impact on pilots or trials, but they matter very much at scale. Pilots and trials often focus on innovation or exploration. But scale means business and that means investment and risk. Since investment often matters more than technology, scale demands standards.
This means initial deployments might pay only lip-service to standards, but as soon as large organizations within the fast food, express delivery and security industries wish to scale, adoption of standards will be rapid.
So, I guess I’m saying this will be a collaboration between cities protective of their social, pedestrian spaces and the companies that will bring scale — such as Uber Eats or DHL.
As with a lot of innovation, the first half of the coming S-curve will be hailed as a “hockey stick.” My guess is that we will be deep into that hockey stick well before 2030.
JB: ISO 4448 addresses “public-area mobile robots” - whereas others in the industry, or outside observers, may be more familiar with terms like PDDs, delivery bots, or robotaxis. How do you define the differences between those, and which are incorporated into the standard you’re developing?
BG: Let me first read the draft standard’s definition of PMRs: “Public-area mobile robots are devices that operate under software or teleoperator control in pedestrianized spaces among non-involved humans (bystanders).” So, that simple definition includes PDDs, delivery bots, sidewalk robots, sidewalk drones, etc. But it also includes surveillance robots, floor-mopping robots in airports, robots spraying brine on icy sidewalks, mobile “social robots” guiding tourists, etc.
Looked at this way, the standard is describing the behavior, orchestration or capabilities of a robot in a public space, So, as far as 4448 is concerned, it doesn’t matter what type of robot it is or what its task is. 4448 is concerned with behavior, competency, and fleet governance within those environments.
JB: What are some of the biggest challenges facing the public-area mobile robot industry as a whole, and how does that impact companies in the sector, the cities they operate in, and the public at large?
BG: I think there are two challenges. First, as soon as consumers stop being enthralled by robotics, they will start demanding far more physical agility and capability than we are seeing now. Having a box on wheels bring your pizza may currently be novel and something to write home about, but that wears off quickly. Cities that have permitted delivery robots are full of pedestrians that no longer even notice them as they walk along looking at their phones.
At maturity many ordinary, every-day robots will be legged, they will be nimble, capable, competent, and will start exhibiting mobility capabilities beyond human. That’s the technical innovation challenge. That will cost billions. This level of mechatronics ability and intelligence will be the mobility correlate to Artificial General Intelligence.
Second, there is the regulatory challenge. In many ways this will be harder and take longer. Pedestrian spaces are highly unstructured, no two walkways are the same. The edge cases in pedestrian spaces including inside malls and hospitals are more numerous than those on the road. The saving grace is that in these pedestrianized spaces speeds are 3 - 4 mph. Partially automated passenger vehicles have already been involved in numerous fatal crashes. Partially automated delivery robots have caused only the smallest of scrapes. Still the deployment of multiple fleets of a high number of PMRs, each engaged in independent tasks, under the independent oversight of multiple, independent, for-profit, fleet operators, could mean all sorts of negative outcomes. The challenge of safely orchestrating a managed number of these robots is easy enough to understand technically — it’s similar to an air traffic control problem — but I think that the regulatory demand for this solution is at least as difficult as the technical demand for the mechatronics.
I hope I am wrong.
JB: You just painted a troubling picture of the regulatory environment for public area mobile robots. What can the industry do to stay in the good graces of regulators and politicians?
BG: Well, when we get to the point of a reliable, and well-accepted standard that sees a reasonable level of behavioral compliance— whether in PMR code or in PMR teleoperation — that would help reduce the risk regulators face when making decisions to accept a new technology or system. They would have something to test against. They would have vocabulary, definitions and benchmarks. They could better understand liability.
I think without that most regulators will do what the city of Toronto did. In December 2021, Toronto City Council said “no thank you, until we see some reasonable guidance.” And I happened to agree with that. I blogged at the time that had I been a city counselor, on that day, hearing those arguments, I would’ve had to vote for the temporary ban. I don’t believe there was any other responsible, immediate action that this council could have taken.
But the other and longer view of this is that robotics can bring us enormous urban benefits — more safety, cleaner environments, and, surprisingly, new job benefits. When we start to see these kinds of outcomes, and city councils that see safe, acceptable ways to deploy, they will be hard-pressed to continue saying “no.” So my answer is if you want to stay in the good graces of regulators, get behind a competent standard, make sure that standard works both for the cities and for the industry. Then build systems that comply.
Without standards, industry cannot scale. That has been true for hundreds of years — and will continue for hundreds more.
JB: Finally, do you have any general insights into the state of the industry you’d like to share? Any thoughts on where you see the space evolving in the next few years?
BG: The expression “the industry” is complicated. If you are mostly thinking about last-mile delivery — which may continue as the largest single sector — then this industry is barely starting. The impact so far has been tiny, while the potential remains enormous. Probably the greatest thing holding back explosive growth is regulatory.
But that is just scratching the surface because we may be waiting for infrastructure updates in some places, or the evolution from wheels and self-serve to legs and graspers to navigate steps, doors and elevators, or end-to-end systems that begin with self-extraction at a shippers warehouse all the way through to automated drop off at the receiver’s locker.
Expecting a person to reliably stand by to extract a delivery from the carrier’s bin is a very limited business model. It is well-suited to hungry students, but not much else.
But even this is only a tiny fraction of an answer to your question. “The industry” includes surveillance, maintenance, and many other robot types and tasks. Each of those have their own issues, infrastructure questions, mechatronic innovations, and so on.
Throughout all of this – especially outside on streets and sidewalks we have matters of urban planning, traffic management, IoT, enforcement systems, and more. These are industries, too.
How can PMR technology improve the prospects for complete streets, 15-minute cities, or car-free living? I am sure it can, but there is no roadmap, yet, to get from exciting mechatronic innovations to substantially improved livability. That’s my wider-industry view. It’s massive and touches everything. And that’s where I believe this space is evolving over the next decade and beyond.
And that’s the driving purpose behind the mission here at the Urban Robotics Foundation.