top of page
drone delivery 1.PNG

Ground Based
Automated Mobility


13.1 Context

13.1.1 Automated vehicles: ”It’s life Jim, but not as you know it”


Over recent years there has been much focus on so called “autonomous vehicles”, or to give them their more accurate description “automated vehicles”, in popular parlance, “driverless cars”, and of the possibilities and capabilities of “connected vehicles”.

Connected and automated vehicles are vehicles that interact with each other and their operating environment, notably the road infrastructure and its agents. See Section 9.


All of our cars are already slowly becoming ‘connected vehicles’. But, despite the hype regarding automated vehicles, the much-vaunted driverless car, and robo-taxi, promised by 2020, will be a lot longer coming. Of all the automated vehicle scenarios that our cities are most likely to face in the foreseeable future, the one that is likely to come soonest and to have the most unanticipated impacts, is the introduction of small robotic vehicles on the sidewalk.

Forget the hype about driverless cars, and forget the hype about aerial drone parcel deliveries, these ground-based robotic machines will engage in package and food delivery, sweeping, snow removal, surveillance, measurement, monitoring, repositioning dockless scooters to where they can be re-charged, and potentially many other tasks.

Ground-based, delivery robots will arrive sooner than will automated cars or robo-taxis. Delivery robots are essentially containers on wheels that can ply sidewalks, intersections and roads over modest distances — without a human attendant on hand — to carry food, packages, and documents. The promise of their widespread deployment in driverless fleet operations is much closer to reality than it is for automated passenger vehicles.


Dozens of companies including Amazon, FedEx, Starship, and Uber, together with many start-ups, are building and piloting small, electric sidewalk delivery robots with the goal of reducing the costs of delivering packages, parcels and food and parcels over their last mile. At the same time, cities are interested in reducing congestion and emissions from the use of trucks, vans, and other motor vehicles for deliveries — which have more than tripled in the past decade.











Figure 13.1      Some current robotic sidewalk vehicle form factors

Illustration: Samika Prupas


The four robots caricatured above will likely be among the top 10 models initially used. These robots or their design successors are expected to frequent sidewalks/pavements over the next few years. As they become more capable, their adoption will become more pervasive.


The reasons for this are quite easy to explain. The rapid developments in ICT and AI have enabled the ‘dream’ of automated servant vehicles to be realisable in respect of their mechanical instantiation, but the barriers to delivery of automated passenger vehicles in a high volume/high speed multi-agent road system lie as much or more in the traffic management and operator behaviour than in the mechanical manipulation of the vehicle; and the speed and impact forces at 130 kph, or even 50 kph, of a mass exceeding a ton, has usually life threatening consequences that have to be safely and reliably controlled and managed before driverless cars become a common reality.


While there are challenges to be overcome before ground-based, kerbside robot deployment can be safely achieved, there are far fewer and lower challenges than those facing the driverless car or robo-taxi.


Additionally, the accelerators driving development of delivery robots are more accessible to innovators, investors, and other participants.


13.1.2  The evolutionary role of the sidewalk   History


The ‘curb’ (North American English), or kerb, is the edge where a raised sidewalk (pavement in British and Singaporean English; pavement or footpath in Australian English) or road median/central reservation, meets a street or other roadway.


Pavements/sidewalks are a relatively modern development that has evolved as cities have developed and been populated by an ever-expanding mixture of pedestrians and faster-moving, powered vehicles.





Figure 13.2 – Roads before sidewalks

(Images courtesy CSi Library)


As cities grew larger, with larger populations walking the streets, and ever more and ever faster vehicle movements, sidewalks (pavements) have become the norm, are built into every planning application, and in the wide variety of widths and forms that we know today.


Some parts of sidewalks/kerbsides may be reserved for loading zones and the occasional bus stop, or to house a bus shelter. Street furniture (signs, lampposts etc) are more normally set into the sidewalk than the roadway. Nearer to buildings, the remaining space tends to be the area used by pedestrians, and, to the widespread annoyance of pedestrians, routes atop the sidewalk are increasingly reserved for cycle paths (although in some places these more properly intrude into the roadway space). The remaining sidewalk space is interrupted by driveways, fire hydrants, bus shelters, and sometimes a tree, a place to sit, or a post to lock and park bicycles. In many cities, cars also, legally and illegally, park on sidewalk space.


This 20th century description of the curb (kerb) and sidewalk (pavement), while still dominant in our communities, is slowly giving way to much greater variety and more organised/managed usage of sidewalks such as ride-hail pick-up and drop-off, segregated cycle/micromobility lanes, ecommerce delivery, micro-transportation docks, and al fresco dining.


There are calls for wider sidewalks and more cycling lanes in many places, especially Europe. Widened sidewalks often result in fewer traffic lanes and on-street parking spaces, in some places leading to the removal of road traffic within so-called ‘pedestrian’ zones, although these zones usually also allow bicycles and micromobility passage.


The removal of road traffic from ever larger town centre zones will further increase the demand for ground-based, delivery robots for last mile deliveries.



One reason that delivery robots will become endemic before driverless vehicles is that the safety barrier for delivery robots is much lower.


Delivery robots come in a variety of sizes and configurations. Smaller units for single deliveries are the size of a filing box and weigh less than 50 kilograms fully loaded.  The top speed of these smaller robots is usually constrained to about six kilometres an hour, a hurried walk. Small and slow, they can stop quickly. Even if the system fails to avoid a collision with a stationary object or pedestrian, the damage is usually slight as the robot devices are well protected., and the impact speed and thrust of the device is minimal.


Larger delivery robots — half the size and weight of a passenger sedan and perhaps travelling at 40 km/h — present greater safety challenges and may therefore largely be confined to cycle/micromobility routes and roadways except for very low speed operation in the final metres to delivery.


The sheer variety of delivery robots presents challenges for protecting pedestrians and cyclists. Any regulations will need to account for a wide range of safety considerations. For example, smaller robots might best be kept off the roadway except when crossing at intersections, and the larger robots (as described above) may need to be banned from sidewalks.


One initial catchpoint is that there are already a number of somewhat contradictory regulations that are currently in place that were written in an age when no-one could have seriously imagined ground-based delivery robots. For instance, an analogous, but close, example is the position of e-scooters, indeed, any scooters, in UK.


The Highway Act 1835 Section 72 (England and Wales) states “…..If any person shall wilfully ride upon any footpath or causeway by the side of any road made or set apart for the use or accommodation of foot passengers; or shall wilfully lead or drive any horse, ass, sheep, mule, swine, or cattle or carriage of any description, or any truck or sledge, upon any such footpath or causeway; or shall tether any horse, ass, mule, swine, or cattle, on any highway, so as to suffer or permit the tethered animal to be thereon……….”


Peculiarly, it is this section, which concerns animal herding, that renders bicycles illegal on the pavement (bicycles were classified as ‘carriages’ in 1888) and also renders cars illegal on the pavement (cars were classed as ‘carriages’ in 1903). Because of this catch-all definition of “carriage”, it is presumed that kick-scooters are probably illegal on pavements unless a local by-law permits them, but they have never officially been classed as ‘carriages’. As they, to date, have more or less exclusively been ridden by small children, this has never been tested in court.


The painting of bicycle lanes on pavements is of course legal, but the riding of bicycles on those bicycle lanes is technically illegal unless a local authority bylaw has specifically re-designated that bicycle lane to be something other than a “footpath or causeway by the side of any road made or set apart for the use or accommodation of foot passengers”. If the local authority fails to do this, users are transgressing the law.


Highway Code rules 37 and 38 apply to wheelchairs or “invalid carriages” as it is legal for wheelchairs to share the road and pavement and state:


“Rule 37: When you are on the road you should obey the guidance and rules for other vehicles; when on the pavement you should follow the guidelines and rules for pedestrians.


Rule 38: Pavements are safer than roads and should be used when available. You should give pedestrians priority and show consideration for other pavement users, particularly those with a hearing or visual impairment who may not be aware that you are there.”


But in respect of powered scooters, they clearly fit under the law as a “carriage of any description”. So, at the moment, they are classified as Personal Light Electric Vehicles (PLEVs), so are treated as motor vehicles and are subject to all the same legal requirements - MOT, tax, licensing and specific construction, and as motorised vehicles are banned from use on the pavement (sidewalk). But because they don't always have visible rear red lights, number plates or signalling ability, they can't be used legally on the roads either.


It is hard to see how robotic motorised delivery vehicles, or robotic vehicles used for any of the purposes described above will be classified differently, and so a change of law will be required before they can be used. Similar, but different, laws exist in most countries. At least UK law permits anything unless it is prohibited or limited by law, so it is simply a matter to modify the restriction imposed by Section 72 of the 1835 Act. Many other countries will require completely new regulation(s) to allow and permit these devices, otherwise there is an assumption that it is a prohibited activity.



In respect of the subject of this subsection, safety, one advantage that delivery robots have over e-scooters and bicycles is that they are designed not to hit anything (e-scooters and bicycles rely on the judgement of their riders, and bicycle riders around the world are noted for their frequent behaviour to ignore or disobey regulations), and delivery robots are generally well protected. If one of the smaller robots were to hit an adult human, it is unlikely the collision would be anything more than light bruising, and certainly not an injury that was fatal or even life-threatening. (However, an exception to this is that a robot could precipitate behaviour that cause an action that results in a fatality in the same way that a pet running into the street might cause a vehicle driver to swerve and crash.)  Low fear factor


The fear barrier for ground-based delivery robots is much lower than it is for robo-taxis or driverless vehicles.


Many people express fear of using an automated vehicle, or being in the presence of automated vehicles sharing the same road-space, or being harmed by a robotaxi or afraid of riding in one. This fear causes both manufacturers and regulators sensibly to be conservative about removing the vehicle’s safety driver. Note that in all of the thousands of videos where a driverless-taxi safety driver is absent, the weather is always especially clear, the roads are in excellent repair, and traffic is notably light.


Consumers may accept that the company using a ground-based delivery robot to deliver their sandwich or parcel from Amazon might wait until a downpour lets up. They might not accept that from a passenger vehicle when they are late for an appointment.


Pedestrian feedback in trials show that they affectionately compare the threat presented by ground-based delivery drones more to that presented by well-trained small pet dogs than risks presented by vehicles, and considerably lower risk than bicycles.




Figure 13.3 – Starship Technologies ground-based robot in Milton Keynes UK

(Image courtesy CSi Library)


Concerns about job losses for those impacted by delivery robots carry less political weight than similar fears about robo-taxis or automated trucks.


Employment for truck, transit, and some taxi drivers is frequently permanent, full-time, and unionized. This means contracts and employee benefit packages. Setting aside projections of driver shortages and arguments promoting “career retraining” — which are often not accepted by the people so employed — many workers and their families feel threatened by automation. In many cases, unions and associations can create effective, if limited, barriers to the employment of larger, automated vehicles for passengers and goods.


Short-haul delivery — especially in the fast food, e-commerce and delivery van sectors — generally utilises temporary, part-time or second jobs, and jobs for youths and gig workers.


There are fewer coherent voices to speak out against automation of these jobs, implying that the union, social, and employment-equity barriers to the diffusion of sidewalk robots should be much lower than that for robo-taxis.



The cost barrier to developing and deploying ground-based delivery robots is much lower than that for automated vehicles because the investment required to build and prove driverless passenger vehicles is far greater than that required to build and prove unmanned delivery robots. These devices are generally relatively low cost, like short life drones with processing ability. Consider for example how the cost of recreational aerial drones has fallen dramatically where their capabilities have significantly increased. Flying, self-homing, camera and wi-fi-equipped devices are available for a little over a hundred Euros, and prices continue to fall. Ground-based delivery robots will cost a small fraction of an automated vehicle.


While complex shared traffic and infrastructure issues are common to both automated vehicles and ground based delivery drones, even for ground-based delivery drones, Infrastructure is a complex barrier to consider.


Ground based delivery robots will have to run a gauntlet of human legs, barking dogs, baby strollers, planter boxes and uneven pavement, not to mention negotiating kerb steps and crosswalks — a much more disorderly environment than the highly regulated city streets where robo-taxis will operate.


Ground based delivery robots are expected to operate on existing infrastructure, but there is a critical difference in that the rules governing the configuration, condition and certification of sidewalks, and the systems to manage and broadcast information about construction and configurations in those spaces, are neither as well-formed nor as frequently complied-with as they are for roadways. Cities have many more undigitized and non-conforming sidewalks than roadways. This constitutes a relative barrier for operating delivery robots.


Fewer people would be offended if a cycling lane is shared with robotic delivery vehicles than would be offended if such delivery vehicles were added, without constraint, to pedestrian clearways. But a key expectation of this technology is to bring deliveries to doorways, and that implies at least crossing the pedestrian clearway. Furthermore, it is highly likely that the current “picnic-cooler-on-wheels” design will be a limited, even short-lived, solution to robotic last-block deliveries. This is because many aspects of navigating the sidewalk and curb would be better handled by robots that can walk.


In other words, if this technology is to be truly successful, sustainable and widely adopted, eventually many, if not most, last-block robots will be walking (ambulatory) rather than rolling. If this sounds fanciful see the demonstration video at the link below:


Demonstration video:


Among all the ground-based drones/mobile robots, the quadrupedal robot is a legged robot, and is superior to wheeled and tracked robots because of its abilities to explore in all the terrain in the same manner as the human and animal, making traversing kerbs, or mounting stairs achievable, and better equipped to manage being impacted, kicked or shoved. The demonstration video above shows, ‘Spot’ and its colleagues, (Boston Dynamics {now owned by Google}) which is only one of a number of research projects in this area.

13 kerb fig 2 charicatures.PNG
13 kerb fig 1 Roads before sidewalks.PNG
13 kerb fig 3 starship robot in milton k
Kerb 13.1 Context
13 kerb fig 4 ambulatory robots.PNG

Figure 13.3 – Ambulatory ground-based robots

(Images courtesy Boston Dynamics)            The challenges



As discussed above, the curb and sidewalk space in many towns and cities are under increasing pressure for access from a growing variety of users, innovations, devices, businesses, and services.


Over the past decade, digitalization of mobility and commerce has brought rapid growth in new forms of taxi-class operations loading and unloading passengers at city curbs as well as a dramatic rise in goods delivery from e-commerce systems, exacerbated by the effects of the pandemic.


In many areas of some cities, this change has already reached unsustainable conditions, and some of these tend to be addressed on a local and urgent short-term basis — often without a long-term framework for future change, growth, or innovation.


In addition, the rise in active transportation has added cycling, scooter, and board lanes at the curb in many cities, as well as scooter and bicycle storage spaces on the sidewalk.


The onset of the COVID-19 pandemic created rapid and unexpected demands for these sidewalk and curb spaces to accommodate social distancing, an uptake in the use of micromobility vehicles such as scooters and e-bikes, and increased demand for al fresco dining space.


Wider sidewalk areas are being created to accommodate these new demands. These areas sometimes extended temporarily beyond the curb and into cycling and parking lanes.


Additional width invites more variety and creates an even greater need for management as social distancing continues, micromobility grows, and demand for walkability increases along with a growing need for cleaning, maintenance, and snow removal for these expanded and complex spaces.


The near future will see growing demand for the delivery of passengers and goods to the curb — soon using driverless vehicles and the final-block delivery of goods via sidewalk robots. Indeed, such systems are already in trials and pilots.


This will not only lead to increasing traffic volume requiring highly digitalized management, but also a change in the nature of the interaction of these vehicles and their mobility systems with each other, with the curb, with payment systems, with active human mobility, and with our existing manual vehicles and devices.


It is logical that these systems will use a communication and security system compatible with that used for C-ITS, indeed there is strong logic to use a similar architecture and communications means because parking often interacts with kerbside, and delivery, robots will need to cross roads and therefore at least be aware of approaching vehicles, and probably interact with them.



The traffic and parking rules that cities have relied on prior to 2020 represent governance that is already under duress — their inadequacy and shortcomings made evident by the pandemic.



Neither current rules nor their temporarily-modified versions will support the new, automated systems that are anticipated. Cities will need new operating guidelines as automated taxis and delivery robots arrive on our sidewalks and curbs and stop, park, wait, load and unload under sensor, effector, and software control.


Often unaccompanied by human passengers or attendants, these machines will need to be prioritized, scheduled, queued, bumped, and placed in holding patterns regardless of nearby human oversight, and all without blocking crosswalks, bicycle lanes, micromobility users, no-stopping areas, or transit stops.


This must be achieved safely, mixed with human-operated vehicles, without inconveniencing active transportation or pedestrian traffic, and with regard for human accessibility challenges.


For those walking to a destination or to make a delivery, the sidewalk is a path.


For those who may be window shopping, sitting on a bench, paying for parking, meeting someone, sleeping, sipping coffee, begging, or walking their dog, the sidewalk is a place.


This fundamental conflict between path and place is mediated by social behaviours and low speeds.


The coming use of delivery robots on the sidewalk implies a purely path-oriented use, except for departure and arrival terminus points. Functionally and navigationally, we can compare this to a pedestrian in a wheelchair using the sidewalk as a pure travel path.


Ground based robotic devices, whilst utilising the sidewalk as a path, must recognise and accommodate those using it as a place.


Ground based wheeled robots (such as the sidewalk delivery robot) have some characteristics similar to a wheelchair:


They can easily travel faster or slower than the average human pedestrian, and must confront issues of climbing over uneven, damaged, steep, sloped, or potholed pavement or ramps to sidewalks. They have difficulty managing a single step of any size and cannot negotiate multiple steps. They cannot “step aside” as an ambulatory human, or ambulatory ground-based robot normally can, and cannot streamline their width by turning sideways while walking as an abled pedestrian can.


Basically, ground based wheeled robots exhibit many of the constraints and properties of a wheelchair. Depending on wheel diameter, number of wheels and their suspension system, a non-ambulatory robot may have fewer or more constraints than does a wheelchair. Indeed, several models of these robots already in pilots exhibit these variations.


By comparison, an ambulatory robot has none of these disabilities, but is an inherently more complex, and therefore more expensive, device.


As a machine, it might be expected that the delivery robot (wheeled or ambulatory) might be regulated to have fewer social rights, or diminished rights of way compared to a pedestrian.


Conversely, as a working machine, it may be playing an important economic role, or it may be delivering something critical to someone who has protected social rights. Perhaps some specially-marked robots might inherit those protected rights in the way that a guide-dog inherits some social rights-of-way from the human it is helping.


A wheeled sidewalk robot will be unable to cross certain barriers or obstacles that an able-bodied human can traverse; it may be subject to vandalism or mischief in ways that are different or more frequent than those confronting a wheelchair user; and it might have a very much lower height profile compared to a wheelchair user, making it less visible to pedestrians unless specially marked or equipped in some way with flags, lights, or motion alarms.


Although it is expected that this will be a ‘connected’ vehicle, it will be frequently required to operate as an autonomous machine, as the delivery robot has no onboard or accompanying human to provide or receive social signals, but must rely on information from other connected ITS-stations, and its own calculations. It may be programmed to send and receive social or directional signals and to exhibit more patience than the average human.


Semi-autonomous robots might be teleoperated, but the ability of a teleoperator to engage in social signalling might be quite limited. (An example of this might be teleoperated micro-mobility devices such as three-wheeled scooters being guided to a docking station.)


13.2   Operating principles for robots on sidewalks:


To provide a proper grounding for operations in a shared, human social environment, operating standards need to be able to  rely on a list of general rules for robots on the sidewalk mixed with pedestrians of all abilities.


These pedestrians may have pets, carry packages, push, drag or ride in wheeled objects, containers, chairs, scooters and more. Some of the proposed rules suggested are:


  • Robots shall grant rights-of-way to humans in close proximity; but rules of engagement shall consider how to prevent a robot from being immobilized for an extended period in a crowded circumstance.


  • Robots shall respect the typical distances normally observed by humans walking or standing in a public place.


  • Robots shall not harm or alarm humans or animals on the sidewalk.


  • Robots shall be visible and/or audible to all humans on the sidewalk (flags, lights, sounds). This is not only to accommodate people who may have visual or auditory challenges but to avoid mishaps with distracted pedestrians.


  • Robots shall signal their presence, priority, and properties to other machines. This enables rights-of-way decisions and can help differentiate autonomous mobility devices from human operated devices, humans, and non-mobility entities.


  • Robots shall not diminish the privacy of people on the sidewalks; (this will constrain recordings and retention of recorded data).


  • Robots shall not diminish the security of humans or other machines on the sidewalks.


  • Robots might be guided by localized infrastructure, high-resolution mapping, and so on, but any additional infrastructure cannot negatively affect (make more cluttered, riskier, more confusing, or less accessible) the use of this shared space by humans.




13.3      Use Cases


13.3.1      Ground-based robotic sidewalk delivery robots


Tom orders some goods online from Amazon. The Amazon delivery van has several deliveries in the area so parks in the town centre car park, programs the addresses into a number of ground-based delivery drones and sends them on their way. An automated phone call is sent to Tom and the other recipients to notify them of the delivery, and again when the delivery arrives. The drone then homes back to the delivery van in the town centre car park.



Jane is sitting with her friend Jill on a bench it the park. They have been exercising their dogs and really fancy a coffee while they chat. Jane calls cup-a-chino, the local coffee bar, who get her location from her phone, which she also uses to pay, and for a small service charge send a ‘bot to the park bench to deliver a latte and a cappuccino.


13.3.2        Ground-based robotic street sweepers


Utopia Borough Council use a fleet of street sweeper ‘bots to systematically clean the high street and central park overnight.


A specially adapted ‘bot is used by Utopia Borough Council to systematically clear the streets of rubbish, and remove dog and bird faeces, blasting the exact site with antiseptic spray once cleared.



13.3.3       Ground-based robotic services inspection


See the following link and its video for an already in-the-market-today example



13.3.4      Ground-based robotic snowploughs


A bad snowstorm falls overnight. Utopia Borough Council re equips its street sweepers with snow ploughs and clear snow from the high street roads, pavement and car park before the morning rush.



13.3.5     Ground-based robotic garbage removal


It is bin day. Utopia borough Council send out its fetch ‘bots ahead of the garbage truck ‘bots to collect wheelie bins left outside by householders to a central point where the garbage truck empties them. The ‘bots then return the emptied bins to their address (identified by a bar code on the bin).




13.3.6     Ground-based robotic al fresco food service delivery


Utopia city centre and cathedral is a magnet for tourists, and restaurants have sprawled al fresco dining into the city park nearby making an open-air food court with some fifteen different catering bars serving every type of food, and the council have set out many tables and benches to accommodate the tourists. When ordering, guests are given a radio tag to take back to their chosen table. When the food is ready it is loaded into a ground-based delivery drone who homes in to the radio signal. When given the radio tag by the guest, it opens its warmed food compartment and the guests unload their food and drinks before the ‘bot returns to its base.




13.3.7    Ground-based robotic washing services


Washing shop windows is a necessary chore. Utopia borough Council offers a subscriber service to its high street shops, where the shopfront windows of subscribing shops are washed and wiped every morning at 6am by a council owned ‘bot adapted for this purpose.




13.3.8    Ground-based robotic construction site materials delivery

The ACME building company is redeveloping a row of three shops in the high street. But space is limited. The company has hired some space in an adjoining street to store building materials such as bricks, sand cement etc. ground based drone deliver the materials from the materials staging area to the building site on a just-in-time basis, called when about to be needed.


13.3.9   Ground-based robotic Valet service

Bob lives in Utopia and has his own ground-based drone. He orders a take-away snack from the shop in the high street, and sends his valet ‘bot to collect it. The local laundry has washed some shirts and dry cleaned his suit, so he puts his clothes hangar rail on his valet ‘bot and sends it to the shop to collect them, while he cooks the evening meal. He realises that he needs unsalted butter for the dessert he is cooking, but there is none in the fridge. So he calls the local grocery shop, and, once he has unloaded the cleaning and removed the hangar rail from his valet ‘bot, sends the valet ‘bot to the shop to collect the butter. That done, and the dinner cooked, he instructs the valet ‘bot to sweep the garden paths while he and his family eat their meal.





13.4      Actors

Cyclist/micromobility user:

rider of a bicycle, e-scooter, e-board, Segway, etc. and therefore likely to be sharing the sidewalk space with the ground-based drone.


Ground based drone/robotic device:

intelligent robotic device, wheeled or ambulatory, under the overall control of a remote operator, capable of limited autonomous decision making within its programmed limitations, used to provide the service.



person who is walking or using a helper animal or device (wheelchair, assistive scooter), especially in a town or city, rather than travelling in a vehicle, and therefore likely to be sharing the sidewalk space with the ground based drone.


Regulatory Authority:

the regulator that determines the regulations and constraints within which the ground-based drone is allowed to operate.


Road user:

vehicles and other carriages or persons using the road space adjacent to the sidewalk that may be encountered by the ground-based drone when crossing a roadway from one sidewalk blockface to another.


Service provider:

With the exception of the valet ‘bot use case, the service provider is the party who normally owns/leases and deploys the ground based delivery drone.


Service recipient:

The beneficiary/recipient of the service provided by the ground-based drone.


Vulnerable sidewalk user:

user of a mobility assistance device, or challenged person under the supervision of another who may be sharing the sidewalk space with the ground-based drone.




13.5      Links with C-ITS

While the ground based robotic drone with make autonomous decisions to manage its immediate movements/actions, it will also need to communicate with others in at least the following situations:

  a)  To receive its instructions (for destination, routing, actions required contingency behaviour etc.)

  b)  To interact with vehicles, and potentially the infrastructure (road crossing lights etc.) when it is en-route

  c)  Overall communication with its operator (progress requests, instruction updates and potentially overrides etc

  d)  To interact with other equipped users on the sidewalk (for example, equipped vulnerable road users)

  e)  To communicate with the recipient of the service (on my way, arrived, etc.)

  f)   To  communicate with emergency service signals relevant to stopping and standing aside

  g)  To communicate or comply enforcement signal  for detour or arrest instruction.


As stated above, during its trips, it cannot be assumed that every party that it encounters will be well intentioned. Indeed, there may be some with malevolent intent.

For interaction with parties b), d), f), and g) it will need to use communications means that are used by those parties.


For obvious reasons vehicles, road carriages, equipped vulnerable road users communicate via secure means, in a ‘bounded secure managed domain’ as specified in ISO 21217 ‘Intelligent transport systems — Station and communication architecture’. This standard does not determine the particular communications means, but determines that the communications means is routed via an ‘ITS-Station’, and that ITS-Station is a cyber secure device protected in accordance with ISO 21177 ‘Intelligent transport systems - ITS station security services for secure session establishment and authentication between trusted devices’.


Further, it may use communications profiles as determined in ISO 21185 ‘Intelligent transport systems — Communication profiles for secure connections between trusted devices’.


Consistent use of media, and commonly use of common media, is required. In a situation where the objective is to avoid collisions, a low latency communications means is required. This eliminates internet routed communications.

Vehicle technology for so called C-ITS (cooperative intelligent transport systems) will either use IEEE 802.11p (WiFi adapted for C-ITS) or will use 5G, as its communications means, or more likely both. At this stage it is recommended that ground-based robots are equipped with an ITS-station of the type supported by vehicles in order to interact with vehicles, the infrastructure, and VRUs, and to maintain an adequate level of cybersecurity.

Communications a) and c) (with the device operator) may securely use the same interface,

Communication e) (with the service recipient) will almost certainly need to be via a mobile phone and will therefore likely use 3GPP specifications for 4G or 5G mobile communications. As these can cover a much longer range than so called G5 C-ITS communications, the same medium used for communication e) (with the service recipient), may be desirable for communication with the operator of the drone as well.

However, communications a) and c) (with the device operator) require the security offered by the bounded, secure, managed domain offered by ISO 21217 , so need to conduct those communications via ITS-stations.

The bounded secure managed domains of ISO 21217 /ISO 21177 operate within the Governance offered by ISO 5616 (at the time of writing, still under development and finalisation).


See Section 8 for more detail of ITS communications.

13.6      Standards


13.6.1     ISO 4448 Sidewalk and kerb operations for automated vehicles


In early 2020, the International Standards Organization launched an Intelligent transport systems project called “Sidewalk and kerb operations for automated vehicles.”

ISO 4448 will be a series of at least seventeen Parts (TR and TS) described in the subsections below.


ISO 4448, when completed, will comprise a set of terminology, guidelines, and real-time procedures for coordination of operations at the curb and on the sidewalk.



The purpose of ISO 4448 is to define the data and communication systems needed to organize and expedite the flow of vehicular ground traffic in cities, specifically with regard to the loading and unloading of goods and passengers at the kerb, and the allocation and movement of service vehicles for garbage removal, sweeping, washing, snow removal, repair, food trucks, construction, and any other service conducted on sidewalks (pavements) or crosswalks.


The data and communications standards being defined in this technical standard are intended to enable carefully defined (mapped) and growing areas of cities to manage any number of vehicles and vehicle varieties operated by any number of operators (public, commercial, and private) for these various activities.


A preview of ISO/4448


ISO/4448 “Intelligent transport systems – Sidewalk and kerb operations for automated vehicles” is

expected to be published in seventeen parts.             ISO 4448-1: Intelligent transport systems - Ground based automated mobility  -  Part 1: Overview of Paradigm


 (Under development)

An overview of the ground based robot paradigm

This document- Part 1 of a multi-part standard for the description and management of Sidewalk and Kerb Behaviour for Automated Vehicles: Arriving, Stopping, Parking, Waiting, and Loading.


The Scope of this deliverable (Part 1) is to provide an overview of the ground based robot paradigm, which covers such kerbs and sidewalks as are suitable for co-temporal, collaborative use by various classifications and combinations of automated and non-automated, wheeled or ambulatory, as well as motorized and non-motorized, mobility-related vehicles and devices as well as for various levels of automated operation of such vehicles. This includes vehicles and devices that move people as well as goods.

A general (non-prescriptive) overview of likely architectures is included in this Part (and will be further detailed in Part 3).

NOTE: This work is descriptive and therefore a Technical Report.

Scheduled  2022.             ISO 4448 Part 2 Data definitions for ground based robotic systems/drones 


(Under development)


4448-2 will provide data definitions and structure for use-procedures and protocols for the sector, and in particular for the other parts of the ISO 4448 series.

Data definitions for parts 4448-4 to 4448-16, inclusive. Scheduled  2023. (but it is likely that there will be successive iterations as the parts of ISO 4448 develop.             ISO TS 4448-3: Intelligent transport systems  -  Ground based automated mobility   -  Part 3: Communications and cybersecurity for ITS ground based robotic systems/drones

(Under development)


Secure interface between devices (vehicles and infrastructure). Ensure access to or request of data that each device has the appropriate credentials to access. Scheduled  2023.

This part of ISO 4448 (Part 3) provides the following for cooperative telematics applications for Ground based robotic systems directly communicating via a ‘Secure Interface’:

a)          A framework for the provision of cooperative telematics application services for Ground based robotic systems;

b)         A description of the concept of operation, regulatory aspects and options and the role models;

c)     A conceptual architecture using an on-board platform and wireless communications to a regulator or his agent;

d)          References for the key documents on which the architecture is based;

e)          Specification of the architecture) of the facilities layer;

f)           A taxonomy of the organisation of generic procedures;             ISO TS 4448-4: Intelligent transport systems  -  Ground based automated mobility - Part 4: Kerb/Kerb loading / unloading for ground based robotic devices


(Under development)



Procedures and protocol for goods and passenger vehicles to reserve, queue, access, loading/unloading spaces at the kerb . (including queue matching).

4448-4 will provide the procedures and protocols to find, prioritize, reserve, schedule, accept, queue, decline, bump, and release vehicles plus numerous other aspects of managing a ground control system for loading and unloading in allocated areas in urban ground environments.


While designed for unmanned vehicles, 4448- 4 can also apply to non-autonomous taxi and ride-hail vehicles.


Activities covered by this section are limited to matters of stopping in order to load and unload people and goods, as well as provisions to accommodate service vehicles that would provide maintenance services such as snow removal or street sweeping.


The data descriptive of stopping a vehicle to load or unload is very similar to that needed for vehicle parking. Hence, ISO 4448 uses as many data definitions as possible from existing standards for parking, in particular, ISO 5206. However, a key difference is that the focus of parking standards relates to short-term storage of vehicles while not being used, whereas 4448:4 relates entirely to a temporary pause for loading and unloading. Data systems that deal with both parking and loading will need to reference multiple standards.


The following is a simple scenario to illustrate:


The loading spot that was assigned for Alice’s taxi to drop her off is withdrawn just prior to the taxi’s arrival at the spot. This might have happened because the spot was claimed for a higher priority vehicle, or a previous vehicle was unable to evacuate the spot, a scofflaw parker, or some other unforeseen circumstance. While there would likely be a procedure for responding to each of those reasons, a solution in this case would be agnostic to the specific reason.             ISO TS 4448-5: Intelligent transport systems - Ground based automated mobility  -  Part 5: Footway, bikeway, roadway procedures and protocols for ground based robotic devices


(Under development)


Procedures and protocol for robotic devices to reserve, access, and queue at/on public places/surfaces. Including procedures and protocols for automated devices on the sidewalk and in crosswalks.

Scheduled  2023/4.

4448:5 will provide similar procedures and protocols for robotic devices providing delivery and other services in order to request, prioritize, reserve, schedule, accept, queue, decline, bump, or release access permission to a blockface. It also provides a number of rules to inform its micro-navigation behaviours along that blockface. Here is an example scenario:


A sidewalk robot from LunchBots is delivering several lunches to a building 400 meters away. It approaches a passage on the pedestrian clearway that is too narrow to traverse while any other pedestrian or robot is within that passage. The narrow passage in question was not included in the robot’s internal map of the sidewalk features because the passage was made narrow just this morning by the placement of a sandwich board to advertise a sale:


  • What must this robot determine before it can proceed to move through this narrow passage?

  • How quickly should it be permitted (or constrained) to execute that passage?

  • What sound or signal should the robot display (if any)?

  • How far beyond the narrow passage must the robot assess its ability to proceed without forcing a pedestrian to wait or step aside at the other end? In other words, what steps is the robot expected to take in order not to inconvenience or interrupt a pedestrian’s progress?

  • If the robot has entered such a narrow passage and a pedestrian subsequently enters from the other end, how must the robot respond?             ISO TS 4448-6: Intelligent transport systems - Ground based automated mobility -  Part 6: Integration of kerb and pavement/sidewalk deployment


(Under development)


Procedures and protocol for the special case of vehicles constrained to the kerbside to load/unload smaller robotic vehicles to move goods & passengers along public pathways that do not admit larger vehicles (e.g., move from truckload to delivery robot).Procedures and protocol for using a (possibly automated) delivery vehicle stationed at a kerb to release one or more sidewalk robots for delivery on that sidewalk.


Scheduled  2023/4.

Part 6, integrates the procedures and protocols from Parts 3 and 4 in order to coordinate the expected logistics systems needed to allow a delivery van carrying multiple packages to park and deploy one or more onboard sidewalk delivery robots. This will require sufficient space at the curb as well as reservations at both curb and sidewalk and, potentially, space for staging or waiting until all its mobile components are re-united.


The reason this integration needs particular consideration is that each delivery van and its robots are elements of a whole subsystem. Planning and positioning of all these elements must be coordinated among themselves, and also among other systems of automated and non-automated vehicles. Hence, the nature of requests, priorities, reservations, etc. are more complex than that for individual vans or individual sidewalk robots.  ISO TS 4448-7: Intelligent transport systems - Ground based automated mobility  - Device behaviour


“Rules of the road” for service robots in public places/spaces.


(Under development) Scheduled 2022. ISO TS 4448-8 Intelligent transport systems - Ground based automated mobility  - Social communication (common signals) 


Sound, light and gestural/motional displays to indicate social communications from robot to proximate humans; including those with sight and hearing challenges.


Scheduled  2023/4. ISO TS 4448-9 Intelligent transport systems - Ground based automated mobility  - Kerb infrastructure “ready for AV use”             


Methods/metrics to determine whether a kerb on a block-face is suitable to a particular level/intensity of use of automated vehicles/devices.


Scheduled  2023/4.  ISO TS 4448-10  Intelligent transport systems - Ground based automated mobility  - Active infrastructure “ready for robot use” 


Methods/metrics to determine whether a footway, bikeway, road shoulder segment is suitable to a particular level/intensity of use of service robots.


Scheduled  2023.  ISO TS 4448-11         Intelligent transport systems - Ground based automated mobility  - Weather-worthiness        


Methods/metrics to determine extreme weather conditions a service robot can safely operate.

Scheduled  2024.  ISO TS 4448-12         Intelligent transport systems - Ground based automated mobility  - Crash Procedures


Crash management. Description, cleanup post crash.


Scheduled  2023. ISO TS 4448-13         Intelligent transport systems - Ground based automated mobility  - Mapping Procedures        


Mapping parameters: resolution, update frequency, error tolerances.


Scheduled  2024.  ISO TS 4448-14         Intelligent transport systems - Ground based automated mobility  - Personal assist — Goods  

e-tethered robots. (shopping trolley that follows you etc)


Scheduled  2024. ISO TS 4448-15         Intelligent transport systems - Ground based automated mobility  - Personal assist — Passenger


Passenger robots, such as wheelchairs etc.


Scheduled  2024.
  ISO TS 4448-16         Intelligent transport systems - Ground based automated mobility  - Safety (incl certification)  


(Under priority development)

The scope of 4448-16 includes the capabilities, characteristics and certification of automated mobile robots (AMRs) required for safe operation in public spaces. This is focused on safety aspects of service robots operating in pedestrianized spaces and is exclusive of road-scaled motor vehicles for passenger and goods transport. It also describes the procedures that shall be performed in emergency situations. 

Robots must be sufficiently reliable and certifiably equipped, programmed, guided, etc. to cause no harm (including alarm or confusion) to proximate people, pets or property. Details about machine design (including motion control) that are within the machine control envelope that do not affect its surroundings are not in scope — i.e., any machine design aspect that has no safety impact on external participants is the involved space is not included.


Scheduled  2022.  ISO TS 4448 17         Intelligent transport systems - Ground based automated mobility  - Privacy


Privacy aspects.

Scheduled  2023.

13.6.2  ISO 21217 Intelligent transport systems — Station and communication architecture


ISO 21217 describes the communications reference architecture of nodes called “ITS station units” designed for deployment in intelligent transport systems (ITS) communication networks. The ITS station reference architecture is described in an abstract manner. While this document describes a number of ITS station elements, whether or not a particular element is implemented in an ITS station unit depends on the specific communication requirements of the implementation.


This document also describes the various communication modes for peer-to-peer communications over various networks between ITS communication nodes. These nodes can be ITS station units as described in this document or any other reachable nodes.


This document specifies the minimum set of normative requirements for a physical instantiation of the ITS station based on the principles of a bounded secured managed domain.



13.6.3  ISO 21177 Intelligent transport systems — ITS station security services for secure session establishment and authentication between trusted devices



ISO 21177  contains specifications for a set of ITS station security services required to ensure the authenticity of the source and integrity of information exchanged between trusted entities:


— devices operated as bounded secured managed entities, i.e., "ITS Station Communication Units" (ITS-SCU) and "ITS station units" (ITS-SU) specified in ISO 21217 , and

— between ITS-SUs (composed of one or several ITS-SCUs) and external trusted entities such as sensor and control networks.


These services include authentication and secure session establishment which are required to exchange information in a trusted and secure manner.


These services are essential for many ITS applications and services including time-critical safety applications, automated driving, remote management of ITS stations (ISO 24102-2), and roadside/infrastructure related services.



13.6.4  ISO 21185 Intelligent transport systems — Communication profiles for secure connections between trusted devices


ISO 21185 specifies a methodology to define ITS-S communication profiles (ITS-SCPs) based on standardized communication protocols to interconnect trusted devices. These profiles enable secure information exchange between such trusted devices, including secure low-latency information exchange, in different configurations. This document also normatively specifies some ITS-SCPs based on the methodology, yet without the intent of covering all possible cases, in order to exemplify the methodology.


Configurations of trusted devices for which this document defines ITS-SCPs include:


a) ITS station communication units (ITS-SCU) of the same ITS station unit (ITS-SU), i.e., station-internal communications;

b) an ITS-SU and an external entity such as a sensor and control network (SCN), or a service on the Internet;

c) ITS-SUs.


Other ITS-SCPs can be specified at a later stage.


The specifications given in this document can also be applied to unsecured communications and can be applied to groupcast communications as well.




13.7     Regulations and Legislation

There undoubtedly will be regulation in this area, but there are none in Europe at the time of writing, Some US States are introducing legislation regarding size, speed and insurance It is intended to get the ISO 4448 series of Standards in place for regulators in time for them to be able to use the standards as references in their legislation.


13.8      Terms and abbreviations

Section 13 Kerbside Terms




































Section 13 Kerbside   Abbreviations






13.9      Bibliography


[ 1]        The-Last-Block-whitepaper 

[ 2}        Boston Dynamics 1) About Boston Dynamics

                                          2) Boston Dynamics

s13 kerb terms.PNG
Kerb 13.2 op principles
Kerb 13.3 Use cases
Kerb 13.4 Actors
Kerb 13.5 link with C-ITS
kerb 13.6 Standards
Kerb 13.7 Regs
Kerb 13.8 Terms & Abrs
Kerb 13.9 Biblio
s13 kerb abbrs.PNG
bottom of page