Pokémon Go Data Powers Next-Generation Delivery Robots

Data collected from Pokémon Go players is now being utilized to train delivery robots, enabling more efficient and accurate navigation in complex urban environments. The unexpected application of augmented reality gaming data highlights the growing convergence of gaming, artificial intelligence, and robotics.

From Pikachu to Packages: The Unexpected Synergy

Niantic Spatial, the company behind the augmented reality game Pokémon Go, has partnered with Coco Robotics to leverage its vast database of real-world scans for the development of autonomous delivery robots. Coco Robotics currently operates a fleet of approximately 1,000 suitcase-sized delivery vehicles in cities across the US and Europe [1]. The core challenge? Navigating the “urban canyon” – city streets where GPS signals are often unreliable due to building interference.

How Pokémon Go Data Improves Robot Navigation

Pokémon Go players, over the years, have contributed over 30 billion images and scans of real-world locations whereas playing the game [3]. These scans, focused on PokéStops and Gyms, provide a rich dataset of visual landmarks with precise location data. Niantic Spatial’s Visual Positioning System (VPS) uses this data to create a detailed “world model” that allows robots to pinpoint their location with centimeter-level accuracy, even where GPS fails [2].

Consciously Collected Data

It’s important to note that this data was not collected surreptitiously. Niantic explicitly requested and received consciously recorded and uploaded video scans of specific locations identified within the game [1]. Players scanned these locations as part of gameplay, contributing to the mapping system.

The Evolution of Niantic Spatial

Originally the creator of Pokémon Go, Niantic has rebranded as Niantic Spatial, recognizing the broader applications of its augmented reality technology. Brian McClendon, CTO at Niantic Spatial, explained the shift, noting that the initial focus on AR glasses has given way to a growing demand for real-world data in robotics [2]. He highlighted the surprising parallel between accurately simulating a Pokémon’s movement and enabling a delivery robot to navigate safely.

Implications for the Future

This collaboration demonstrates the potential for repurposing data collected from gaming and augmented reality applications. As autonomous robots develop into more prevalent in urban environments, the need for accurate and reliable navigation systems will only increase. The success of this partnership could pave the way for similar applications of crowdsourced data in other areas, such as autonomous vehicles and smart city infrastructure. The unexpected benefit of a popular game is now contributing to the efficiency of last-mile delivery services.