Niantic announced a partnership with Coco Robotics to use data collected by Pokémon Go players to train autonomous delivery robots, with more than 30 billion images voluntarily uploaded since 2016.
The game’s developer, through Niantic Spatial, reuses augmented reality scans taken at PokéStops and gyms to improve visual navigation systems. Esses data allows robots to identify locations with centimeter accuracy in complex urban environments.
The technology reduces reliance on GPS, which often fails in areas with tall buildings or interference. The partnership represents the first major commercial test of Niantic Spatial’s geospatial model.
Players contributed by scanning real environments during gameplay, often incentivized by rewards like extra items. The records include photos, videos and metadata such as angles, times and weather conditions, forming a vast database of visual references.
Partnership accelerates autonomous navigation
Coco Robotics integrates Visual Positioning System (VPS) from Niantic Spatial into its robot fleet for last-mile deliveries. Esses equipment, approximately the size of suitcases, operates on sidewalks and bike paths in selected cities.
The robots reach speeds of up to 21 km/h and transport items such as meals and shopping. Visual precision helps you avoid obstacles and maintain safe routes on busy streets.
Niantic Spatial highlights that the challenge of positioning virtual objects in the real world, as in the game, coincides with the need for robots to move around safely. Diretores technicians state that the technical problem is essentially the same.
Data accumulated since the game’s launch
Since the launch of Pokémon Go in 2016, millions of users have submitted billions of opt-in scans. The focus is on points of interest, such as monuments and murals, captured from multiple angles and times.
This material constructed detailed three-dimensional models of urban areas. The database overcomes limitations of GPS in “urban canyons”, where satellite signals are blocked by buildings.
Niantic bolstered collection with specific in-app missions that reward players for additional scans. Assim, the bank grew organically throughout the decade.
Limitations of GPS in dense cities
In metropolitan regions, the GPS signal loses effectiveness due to physical obstructions and reflections. Robôs dependent solely on this technology face positioning errors that compromise deliveries.
Niantic Spatial’s visual system uses images of the surroundings for accurate triangulation. Isso allows real-time decisions, such as route adjustments in the face of pedestrians or vehicles.
The approach improves reliability in real delivery scenarios, where centimetric precision makes a difference to safety and efficiency.
Transparency in information collection
Niantic states that privacy policies have always indicated the possibility of using data for purposes other than gaming. Usuários chose to participate in scans by activating the function in the application.
Not all players associated these contributions with robot training. The company maintains that the submissions were voluntary and linked to gameplay benefits.
The debate about crowdsourced data reuse gained momentum with the announcement of the partnership. Especialistas note that the model demonstrates the long-term value of mass interactions in AR apps.
Future applications of geospatial technology
Niantic Spatial plans to expand the use of Large Geospatial Model to other sectors that require dynamic mapping. The partnership with Coco Robotics serves as a proof of concept for autonomous robotics.
Coco robots already deliver to locations in Estados Unidos, China and parts of Europa. VPS integration aims to increase coverage and speed of operations.
Experts indicate that continuous data from robots can feed back into the model, creating a cycle of constant improvement in understanding the urban environment.
