The 8th edition of the Wherecamp took place in Berlin again. It's my first time at this conference focused on mapping and location. Many interesting talks give an overview on current trends and topics. Here are my notes and highlights.
In the summer the first phones where released with a dual GPS chip for much better accuracy. This allows for a new generation of geo applications: using technology to draw the map. Like Web 2.0 introduced creation in the web, this allows people to create instead of just one way consuming.
From ABC to XYZ: location data for non-location developers
HERE released a set of tools named XYZ in October to help developers get started with location data. Their product shall be very free, but there will be a commercial version.
The current released tools are just the tip of the iceberg, there's more to come.
Data created from any device pops up immediately, it's thought to be data driven maps. When there is data, more complex applications can be build.
Machine Learning for Transport Modelling
Huge changes happen in the transport market. Car sharing, ride hailing and so on emerged in the last few years.
Floating phone data is used for a origin-destination-matrix on a postal code level. With that data the movement and mobility demand can be estimated. There is a high error rate for the first approach.
The models shall be used to calculate the share of a trip with different mobility providers, e.g. going to work by bus and metro.
enviroCar – a citizen science platform for sustainable traffic
An Android app collects car sensor data which is stored and can be analyzed. On a dashboard a user can compare themselves to other drivers.
CO2 emissions can be seen on a map that combines the data from all users. Hotspots can be identified easily. Some where expected like at the end of the highway, but others were unexpected: a small street appeared because of bumpers that forces drivers to s slow down and accelerate again.
GeoAI: feature extraction and classification
The earth observation age is here. And it's changing the way we live.
GeoAI improves the process of the geospatial engine. Buildings, water, cars, roads and so on can be recognized and extracted from the satellite images. The accuracy is high: 95.82%.
Some areas need special treatment. Canals are no rivers, can be identified because they're pretty straight.
Localization Service using sparse visual information based on recent AR platforms
Indoor navigation implemented with Google Tango which worked pretty guide. Some of the technology is now in ARkit.
The localization shall work by recognizing visual clues with AR and providing the guessed location to the user. Similar to geogussr.
Navigation for blind or visually impaired
Accurate positioning is needed to provide instructions. The positioning is done with beacons and Wifi. The instructions are read out then.
The biggest project is the new airport in Istanbul which is still under construction. There are 5.000 beacons and 2.000 routers installed to provide the location.
How future citizens of smart cities take benefit from Geofencing technology.
Air pollution is an urgent problem for most cities. Geofencing is used to notify users when they enter a critical area so they can avoid that area.
Other use cases: location based ads, automonous driving in certain areas, drone flights in allowed areas, …