Augmented Reality, Autonomous Vehicles, Artificial Intelligence, and Activity Pub for Access and Mobility
How Emerging Technologies may shape Transport Accessibility and Mobility
Advances in technology have driven the evolution of transport accessibility and mobility. As faithful readers know, transportists like alliteration.1 The recent rise of Augmented Reality (AR), Autonomous Vehicles (AV), Artificial Intelligence (AI), particularly Large Language Models (LLMs) and Ensemble Models, along with Activity Pub, gives us a new set of A’s. These are significantly transforming the transport landscape, promising greater accessibility and enhanced mobility.
Augmented Reality, also known as Mixed Reality, overlays digital information onto our physical world, offering opportunities to enhance both accessibility and mobility in transport. AR applications can provide real-time navigation assistance, improving mobility by reducing the risk of getting lost or missing a connection. Constant heads-up displays on glasses or windshields will change how we perceive the world. Furthermore, AR technology can radically enhance accessibility for individuals with disabilities, by providing them with additional layers of contextual information, which could help them navigate transit systems more efficiently and safely.
Autonomous Vehicles embody advanced AI, and are likely to radically reshape the future of traffic. AVs can improve accessibility by providing a reliable mode of transport for those who may have difficulty driving due to age or disability. This would allow these individuals to access facilities and services that might have been challenging to reach before.
The role of Artificial Intelligence, especially Large Language Models and Ensemble Models, in enhancing transport accessibility and mobility is multi-faceted. LLMs, with their capability to understand and generate human language, can aid in making transport information more accessible. They can translate information, offer voice-activated services, or generate text from data, assisting those with vision impairments or literacy challenges. Ensemble Models, which combine predictions from multiple machine learning models to optimize outcomes, can improve mobility by predicting traffic patterns or optimizing transport routes. This results in more efficient, and accurate, transport planning and operations.
The most speculative of these A’s is Activity Pub, a decentralised social networking protocol best known for Mastodon. While ActivityPub's initial design was oriented around social networking, the protocol's capacity for fostering decentralized and interoperable communication lends itself to a much broader array of applications. In a world increasingly characterised by “smart cities”, digital objects, the Internet of Things (IoT), and seamless digital integration, envision a transport network where ActivityPub serves as the underlying common communication protocol. Individual vehicles, public buses, trains, bike-sharing platforms, traffic signals, road sensors, and even passengers' devices could form an interconnected two-way web of information exchange, fostering an ecosystem of "networked mobility." Many of these platforms are already digital, but the data (like traffic signal data) is effectively held behind government and corporate paywalls. To be clear, this is not at the level of control, for which AVs are necessary (and Connected Vehicles are highly risky), but at a more strategic path planning level, where near real-time information is useful but not necessary.
Users (and AI-enabled applications) could access near real-time traffic updates, routes, and transport options from a variety of sources integrated into a single feed, not relying simply on a private maps vendor, improving their ability to plan and execute travel. The decentralized nature of ActivityPub would make this transport information network resilient and robust. Unlike centralized systems that have a single point of failure, a network based on ActivityPub would ensure continuous data flow even if some nodes malfunction or go offline. This free-flow of information also enhances mobility by enabling individuals to make informed decisions on the most efficient modes and routes of travel.
These emerging technologies present a promising future where everyone, irrespective of their physical abilities or geographical location, can navigate the world with greater ease and efficiency. While challenges regarding infrastructure, technology adoption, and data privacy still need to be addressed, the potential benefits these technologies bring to transport accessibility and mobility are undeniable.
That seemingly innocuous post cost me more subscribers than anything else I have written.
One per cent of the funding that all of this will have attracted in the forthcoming years made available to basic improvements - such as safe street design, expansion of cycling and other forms of soft mobility, and introduction of level boarding on all public transport (including all vehicles and infrastructure) - would deliver wonders never seen before.
I think we should be thinking about this... more often.
Alliteration always allows comprehensive cognition...
The missing bit in the utopia here, I think, is that all of these large data models are very expensive to operate, both for the planet (data centres are heating up the planet such that the cloud might evaporate), and of course financially to pay for the construction and operation of those many data centres. So for a distributed model to work it would have to be monetised. The second bit is data sovereignty, which unfortunately, goes back to the first point. To manage the cost of their own data, most data owners then try and sell the data rather than open source it. Even government agencies are facing these constraints. Finally, the ability to thread all the information together, as opposed to something like Google maps which infers much from user data tracking, is complex and conflictual. LLMs can retag all data from all sources so that (unlike object-based databases) no common key is required, and much can be extracted according to the user needs. But to do that retagging requires that you have the data in your own system, otherwise the planet will burn with the computing effort required to keep refreshed. In all of these things the trap is to be excited by the technology, while the opportunity is about solving issues that affect the ability of people to thrive in their daily lives. So, claps for focussing on accessibility. But without a driver a lot of people can't access a vehicle, so there will still be a need for assisted transport, particularly with the very large very old cohort on its way. But the upside, no problem with connecting mobile assistants with automated vehicles in a service offering! Keep alliterating and having fun with language, something that confuzzles AI no end and contributes to its hallucinogenic behaviours!