Investors & Investment Opportunities

Multi-modal mapping & semantic terrain models for transport and gaming

Don’t just map from A > B – but factor in all the possible routes and changes, be it a Tube / Metro station and how this correlates to known journey times and schedules.

Whether its faster to jump on an e-scooter, or how long it might take to navigate with a pram, allowing you to map and define a journey via user preferences – such as reduced mobility needs, or a preference for walking and then generate a map of that journey that can accommodate changes.

Dynamic Pedestrian Navigation based on changing environments

Pedestrian mapping and modelling can begin simplistically – such as a map of the routes and locations that can be reached in a given period (e.g. 10mins) then augmented to factor in weighting for crime, time of day, preference for illuminated paths, or to avoid crowded areas at peak times.

Even better in areas such as festivals – where there a lot of changing dynamics and the setup isn’t fixed – we can model and map the best way to get to the main stage – or find the working toilets!

Attractors & Repulsors and tangible simulation of real-world input

A key feature of Polaron is the ability to weight maps and the decisions within this. Attractors and Repellers can be thought of as has a push or a pull effect – allowing intagible and tangible effects to be modelled – and where data is available – included in the model.

For instance closing or reducing the access to certain stations at peak time could be modelled as a repeller or negative effect – whilst elevator access or step free access can be weighted – since not all facilities are made equal – allowing a user to assess the impact of changes. You can even copy and paste these from known sites – e.g.

What if we were to retrofit an older station with new accessibility – and model how that might effect use and flow for those with reduced mobility, prams, or maintenance access.

Programmable Cellular Automatons permitting massive scale rule based simulations.

A cellular automaton is one of the means we use to propagate change and simulate within the environment. By exposing this to users via both a user interface and in APIs and configuration – we’re able to allow local simulation tuning and calibration (e.g. cultural behaviours or new rules) which permits refinement and iteration of the underlying simulation model.

Cellular Automatons have a range of different functions in Polaron – from observation, to opposition or providing friction – e.g. traffic or crowds, which can be fed values from real world sensors to simulate not just the observed area but apply accrued intelligence to inform other simulations. We have a machine learning pathway in the works to make this super simple to train and draw insight from.

Interchanges & Capacity modelling including logistics interchanges and modelling

As users move from one type of transport to another there are changes in the capacity (since not all users will transfer to another mode of transport, which allows for a correlation of behaviours between different transport modalities and the ability to see how one effects the other – leading to prediction and capacity modelling as well as real world changes to optimise and facilitate changes in customer behaviour.

The complexity of this scales as more variation is added – however the ability to segment these associations and allow the volumetric database and machine learning to work out the connection makes assessing the use and optimisation simple. Which leads to a realistic model you can use to plan further change and refinement – as well as a means to quantify & qualify different solutions. There is rarely a one-size all or magic bullet – in effect offering A/B testing for the real world.

Future transport networks & what if scenario modelling including C02 modelling

Once a simulation is generated – it is easy to then begin to experiment with and play out how changes might effect those systems. EV cars are interesting as they have a different set of behaviours associated to them and different attractors (e.g. Charging stations).

Perhaps even more important is autonomous vehicles – where the vehicle in effect becomes a moving hotel room or office – which alters behaviours even more – and where expected patterns may differ – since autonomous vehicles in hired vehicle context may tend to “swarm” an area to maximise profits and fairs – which may then lead to a large demand in charging demand post surge – something that is already seen with Pre & Post rush hour for EV Black Cabs and Taxis – as they ensure their charge capacity can meet the expected passenger demand.

Polaron’s simulation capabilities are uniquely suited to this sort of applied theoretical modelling since potential demands and issues can be extrapolated from modified current behaviours.

Volumetric database and spatial simulation augmenting VR / AR interfaces

Behind the visuals and interfaces lies a cutting edge volumetric database – that we believe is the fastest of it’s kind – which permits real-time sensor data to be displayed (e.g. Lidar / Stereo Cameras / etc) this speed is a key feature of Polaron – since the same speed permits huge datasets to be ingested, processed, updated and modelled on. This scale allows us to process the entire OSM DB of the UK in under 2 minutes – which then stores the entire UK @1m2 resolution.

We can then load and instantiate these tiles in milliseconds and refresh or add additional data to these in real-time. Such performance and database size requires lightning fast data sanitisation and error checking – which we accomplish with a proprietary error checking tool – since parsing data isn’t sufficiently robust and prone to errors and failures. We can consequently place tolerances on this too – permitting us to use imperfect data or data the contains uncertainty – and even encode that into the volumetric data map – allowing partial or imperfect data to be used and quantified.

Semantic segmentation & Profiling via a Voxelised Spatial environment

Polaron features a number of Ai & ML tools – that allow for the semantic segmentation of a scene or map – e.g. to find roads, shapes or patterns. This allows you to ingest volumetric data and tag it (or inherit annotated data from other tools) allowing you to build up an object recognition database and apply properties to these – which augments the simulation.

E.g. Detect all the stairways, benches, or platforms within a scene and model how they are used and utilised. Additionally we can import most objects into Polaron and allow users to apply values to them – including phone based scans, Lidar maps and even Volumetric Neural Models (such as NeRFies)

API interoperability and part of a system of systems – from flat data to simulation & insight

art of the challenge of Public Transport lies in the need to tie into other systems and networks – it’s therefore part of our larger efforts to produce an API portal with a simple and easy to use user interface as well as supporting mobile scanning – to permit interoperability with other systems and engage a larger audience of staff, sensors and even the public to add and help map the environment – be that the basic sensors on a phone – through to emissions monitoring and even pulling in thermal data from LEO to build a picture of the environment, consider passenger safety and provide a mechanism for feedback. Ultimately this is the basis for driving preferable customer behaviours and producing the intelligence to alter customer and commercial behaviours to drive better operational practices, plan changes and evaluate impacts and help to decarbonise the network and meet NetZero targets for both safety and combatting climate change.